Klas Eric David Soderquist

INSIDE THE TIER MODEL: PRODUCT DEVELOPMENT ORGANIZATION AND STRATEGIES IN AUTOMOTIVE EXPERT SUPPLIER FIRMS A thesis submitted in partial fulfilment of the requirements for the degree of Doctor of Business Administration By Klas Eric David Söderquist Brunel University Henley Management College Groupe Ecole Supérieure de Commerce de Grenoble November 1997 To my family - Parents and Wife - I thank them for their support and encouragement. 2 ABSTRACT Over the last five to ten years much attention has been paid to the implementation of Japaneseinspired lean production in Western manufacturing industry, in particular in the automobile industry. Current and past research into the changing role of suppliers in this context has identified increasing product development responsibility as one of the major challenges in the automobile component sector. However, the debate has mostly been focused on the carmakers' outsourcing strategies and the transformation of the largest players in the component sector, i.e. the system suppliers. The present research analyses a specific group of suppliers, namely medium-sized expert suppliers. These are companies present both in the first and second tier, with an explicit strategy of supplying components which provide a value added function, largely dependent on their own R&D efforts. The research focuses on product development organization and strategies from an operational perspective. The objective is to specify the lean product development context inside the tier model and conceptualize the product development process in order to pinpoint the essential factors of a successful product development strategy and organization in medium sized expert supplier firms. The research is guided by a qualitative inductive research methodology that aims at developing ideas grounded in field observations. The modelling of the product development process follows the paradigm model by Strauss & Corbin (1990). Data were gathered through semistructured interviews with General Managers and Product Development Managers in eight supplier companies of the type identified above, and through two in-depth case studies of the product development process in two of these companies. Direct observation, interviews and documentary analysis were used in the case studies. Concerning the lean development context it was found that one and the same supplier has to manage a continuum of customer relationships where each one has its specific raison d'être from generating capital to promoting product innovation. An important tendency is still increased participation in the design of components which confirms the special interest that product development represents for these companies. The product development process is conceptualized in a four level model of operational design -representing individual, group, project, and systemic work- and in four distinctive phenomena related to operational design: means of guidance, design support structures, learning, and core capabilities. The content and coherency of guiding visions and performance measurements constituting the means of guidance, will determine how individuals will perform their work tasks and how individual and collective mental models of development work will evolve. Design support structures in terms of product specifications, information systems, and technology scanning have an important impact on the efficiency of development work. Learning emerges as a transversal issue in the entire development process and is analysed through three learning situations: intra-company intra-functional learning, intra-company interfunctional learning, and inter-company intra-functional learning. Finally, the core capability concept is given a more explicit meaning through the proposition of a model of the emergence of capabilities. The interrelationships between the identified phenomena are illustrated through the present research. Learning is identified as the link between guiding visions and the building of core capabilities. Design artefacts and support structures are important means for realizing strategic objectives and promoting learning. 3 ACKNOWLEDGEMENTS I am grateful to a large number of people for their help, support, and interest that made it possible to realize this research project. I would like to begin with the managers, engineers and technicians in the studied companies, especially the General Managers in two case study companies who allowed me to access their product development departments and share strategic, tactical, and operational issues through a direct observation of operational product development work. My supervisors Professor Jean-Jacques Chanaron at Groupe ESC Grenoble and Professor David Birchall at Henley Management College continuously supported and guided me from the very beginning to the very end of the research process. I owe them warm thanks. I am particularly grateful to them for having associated me with several of their challenging ongoing research projects. Special thanks to Professor Jean-Paul Leonardi, Dean of Groupe ESC Grenoble, whose continuous search for innovation in pedagogy and research made it possible to realize this truly international doctoral research, and to Professor Thierry Grange, Associated Dean for Pedagogy and International Affairs, my very first contact at Groupe ESC Grenoble, whose enthusiasm convinced me to start off on this venture. Thanks to both of them for having supported the research financially. I also wish to thank colleagues and friends at the faculty and administration of Groupe ESC Grenoble, Professor Dominique Jolly, Head of Research, Bernard Chapelet, Raffi Duymedjian and Robert Volsy for stimulating conceptual discussions and constructive critique, Véronique Rungette for documentary research, Dominique Brouty for her never ending patient support in daily administrative activities, and Chloë Thomas for proof-reading. Finally, I am grateful to faculty and staff at Henley Management College, in particular Dr David Price, Director of Studies - Doctoral Programme. 4 TABLE OF CONTENTS ABSTRACT ACKNOWLEDGEMENTS TABLE OF CONTENTS LIST OF FIGURES LIST OF TABLES AND EXHIBITS 1. RESEARCH PROBLEM...................................................................................................... 13 1.1 INTRODUCING THE RESEARCH PROBLEM..................................................................... 13 1.2 LEAN PRODUCTION IN THE SUPPLIER FIRM.................................................................. 16 1.3 LEAN DEVELOPMENT: FOCUS ON PRODUCT DEVELOPMENT ORGANIZATION AND STRATEGIES..................................................................................................................... 18 1.4 BENEFITS FROM INTEGRATION...................................................................................... 1.4.1 The Carmaker's Perspective............................................................................................ 1.4.2 The Supplier's Perspective.............................................................................................. 24 24 25 1.5 DEFINING THE PROBLEM.................................................................................................. 1.5.1 Initial Assumptions........................................................................................................ 1.5.2 Overall Research Questions............................................................................................ 27 27 27 1.6 RESEARCH PROCESS AND PERSPECTIVES..................................................................... 28 1.7 SUMMARY AND ANTICIPATED CONTRIBUTIONS OF THE RESEARCH...................... 30 2 THE EUROPEAN, FRENCH, AND REGIONAL AUTOMOTIVE AND AUTOMOTIVE COMPONENT INDUSTRY IN FIGURES................................................................................ 33 2.1 THE EUROPEAN CAR MARKET AND TENDENCIES TOWARDS VERTICAL DISINTEGRATION..................................................................................................................... 33 2.2 THE AUTOMOTIVE COMPONENT SECTOR IN FRANCE................................................ 34 2.3 THE AUTOMOTIVE COMPONENT SECTOR IN THE RHÔNE-ALPES REGION.............. 35 2.4 SUMMARY........................................................................................................................... 40 3. A CHANGING PRODUCTION ORGANIZATION............................................................. 41 3.1 INDUSTRIAL MODELS AND PRODUCTION MODELS - SOME DEFINITIONS.............. 41 5 3.2 A FRAMEWORK FOR THE DESIGN PROCESS.................................................................. 44 3.3 COMPARING TWO MODELS OF PRODUCTION ORGANIZATION: MASS VS. LEAN PRODUCTION............................................................................................................................ 3.3.1 At the Manufacturing Level: Back to Basics................................................................... 3.3.2 A Complete Model, but Confusion about Transfer.......................................................... 3.3.2.1 The Keiretsu System............................................................................................... 3.3.2.2 The Lean Enterprise............................................................................................... 3.3.3 A Hybrid Perception of the Evolution of Lean................................................................. 3.3.4 Beyond Diffusion - Anticipating a 'Truly Lean' Model.................................................... 3.3.5 Classifying the Perceptions of Lean................................................................................ 46 53 54 55 58 60 62 64 3.4 THE CONTRIBUTION FROM REFLECTIVE PRODUCTION............................................. 3.4.1 Confusion about Productivity.......................................................................................... 3.4.2 Reflective Production - Still no Global Model................................................................. 65 67 69 3.5 A CHANGING PRODUCTION ORGANIZATION - SUMMARY AND CONCLUSIONS.......................................................................................................................... 70 4. SUPPLYING IN A NEW CONTEXT.................................................................................... 73 4.1 IDENTIFYING THE GLOBAL CHANGES IN BUYER-SUPPLIER RELATIONSHIPS........ 4.1.1 Partnership Defined........................................................................................................ 4.1.2 Tier Structures and Sourcing Strategies.......................................................................... 73 76 79 4.2 COST REDUCTION - TRANSACTION COST THEORY..................................................... 83 4.3 STRATEGIC COLLABORATION - THEORIES OF COMPETITIVE STRATEGY.............. 4.3.1 Core Competencies or Core Capabilities......................................................................... 4.3.2. Technology Synergy...................................................................................................... 4.3.3 Synchronized Management............................................................................................ 4.3.4 Conclusions about the Strategic Perspective.................................................................... 85 87 91 93 95 4.4 OPERATIONAL COORDINATION IN THE PRODUCT DEVELOPMENT PROCESS THEORIES FOR ORGANIZATIONAL COORDINATION AND INTEGRATION..................... 4.4.1 Project Management....................................................................................................... 4.4.2 Stage Overlapping and Intensive Communication.......................................................... 4.4.2.1 Tacit Knowledge.................................................................................................... 4.4.2.2 The Complexity of the Design Task........................................................................ 4.4.3 Different Theoretical Concepts for Facilitating Integration and Coordination................. 4.4.3.1 Common Cognitive Ground.................................................................................... 4.4.3.2 The Creative Negotiation and the Whetting Dialogue............................................ 4.4.3.3 Structural Support.................................................................................................. 4.4.4 Conclusions About the Operational Coordination Perspective......................................... 96 97 100 101 102 104 105 106 107 108 4.5 SUPPLYING IN A NEW CONTEXT - SUMMARY AND CONCLUSIONS.......................... 109 4.6 REVISITING THE OVERALL RESEARCH QUESTIONS.................................................... 111 6 5. RESEARCH METHODOLOGY........................................................................................... 113 5.1. METHODOLOGICAL CONSIDERATIONS........................................................................ 5.1.1 Qualitative vs. Quantitative Research: An Attempt to Distinguish the Two..................... 113 114 5.2 FIVE STEPS IN THE ELABORATION OF A RESEARCH METHODOLOGY..................... 5.2.1 Real World Context........................................................................................................ 5.2.2 The Research Problem: Philosophical Standpoint, Objective and Overall Questions....... 5.2.3 Unit of Analysis............................................................................................................. 5.2.4 Research Strategies........................................................................................................ 5.2.5 Relation to Theory.......................................................................................................... 5.2.6 Methodological Choice - A Summary............................................................................. 116 116 117 122 123 123 126 5.3 RESEARCH DESIGN............................................................................................................ 5.3.1 The Companies in which Data were Collected................................................................ 5.3.2 Organization of the Empirical Research and Data Collection.......................................... 5.3.3 Data Analysis................................................................................................................. 5.3.3.1 Data Opening........................................................................................................ 5.3.3.2 Data Reduction / Data Condenzation..................................................................... 5.3.3.3 Iterative Testing of Questions/Propositions and the Emergence of the Conceptual Framework........................................................................................................................ 5.3.4 Reliability and Validity.................................................................................................. 127 128 129 135 135 136 5.4 RESEARCH METHODOLOGY - SUMMARY...................................................................... 140 6 THE LEAN PRODUCT DEVELOPMENT CONTEXT........................................................ 143 6.1 THE PLACE AND ROLE OF AN EXPERT SUPPLIER IN THE SUPPLY CHAIN............... 6.1.1 Direct Arm's-Length Supply........................................................................................... 6.1.2 Indirect Expert Supply.................................................................................................... 6.1.3 Direct Expert Supply...................................................................................................... 6.1.4 Place and Role in the Supply Chain - Summary.............................................................. 143 146 148 149 150 6.2 DRIVING FORCES FOR CHANGE...................................................................................... 6.2.1 Just-in -Time - JIT......................................................................................................... 6.2.2 Quality Management...................................................................................................... 6.2.3 Price Pressure................................................................................................................. 153 154 155 157 6.3 CHANGE AND DEVELOPMENT TENDENCIES IN THE PRODUCT DEVELOPMENT PROCESS.................................................................................................................................... 160 6.4 USE AND PERCEPTION OF LEAN TECHNIQUES............................................................. 6.4.1 The Design Organization............................................................................................... 6.4.2 The Production Arrangement......................................................................................... 6.4.3 Performance Measurements and Career Paths................................................................ 6.4.4 Problems with Lean Techniques..................................................................................... 163 163 164 165 166 7 138 138 6.5 THE LEAN PRODUCT DEVELOPMENT CONTEXT - SUMMARY AND CONCLUSIONS.......................................................................................................................... 169 7. MODELLING OPERATIONAL DESIGN WORK.............................................................. 173 7.1 THE CORE CATEGORY - OPERATIONAL DESIGN.......................................................... 173 7.2 OPERATIONAL DESIGN - INDIVIDUAL WORK............................................................... 175 7.3 OPERATIONAL DESIGN - GROUP WORK......................................................................... 7.3.1 Informal Group Work..................................................................................................... 7.3.2 Formal Group Work....................................................................................................... 181 181 182 7.4 OPERATIONAL DESIGN WORK - PROJECT WORK......................................................... 186 7.5 OPERATIONAL DESIGN WORK - SYSTEMIC WORK...................................................... 190 7.6 THE PRODUCT DEVELOPMENT PROCESS - SUMMARY AND IDENTIFICATION OF RELATED CATEGORIES........................................................................................................... 7.6.1 Summarizing the Four Step Analysis of the Product Development Process..................... 7.6.2 Categories...................................................................................................................... 7.6.2.1 Means of Guidance................................................................................................ 7.6.2.2 Design Support Structures...................................................................................... 7.6.2.3 Learning................................................................................................................ 7.6.2.4 Core Capabilities................................................................................................... 196 196 198 198 198 199 200 8 MEANS OF GUIDANCE........................................................................................................ 201 8.1 GUIDING VISIONS FOR PRODUCT DEVELOPMENT- ORIGINS AND CONSTRUCTION....................................................................................................................... 8.1.1 Guiding Vision: Reduction of Global Component Costs.................................................. 8.1.2 Guiding Vision: Prioritizing as a Function of Customer Satisfaction.............................. 202 204 206 8.2 THE DEPLOYMENT OF GUIDING VISIONS FOR PRODUCT DEVELOPMENT.............. 207 8.3 GUIDING VISIONS - CONCEPTUAL LEADERSHIP.......................................................... 209 8.4 PERFORMANCE INDICATORS........................................................................................... 210 8.5 MEANS OF GUIDANCE - SUMMARY AND CONCLUSIONS............................................ 213 8 9 DESIGN SUPPORT STRUCTURES...................................................................................... 215 9.1 SPECIFICATIONS................................................................................................................. 216 9.2 COMPUTER-AIDED SYSTEMS FOR INFORMATION TRANSMISSION......................... 221 9.3 TECHNOLOGY SCANNING................................................................................................ 225 9.4 SUPPORT STRUCTURES IN OPERATIONAL DESIGN - IN SEARCH OF AN ORGANIZATIONAL MEMORY................................................................................................. 229 9.5 DESIGN SUPPORT STRUCTURES - SUMMARY AND CONCLUSIONS........................... 231 10 LEARNING IN PRODUCT DEVELOPMENT.................................................................... 235 10.1 INDIVIDUAL AND ORGANIZATIONAL LEARNING: TWO CONCEPTS....................... 10.1.1 Individual and Organizational Learning - Conclusions to the Literature Analysis......... 10.1.2 Materializing Learning: Towards a Framework for Analysis........................................ 238 241 242 10.2 THREE DIFFERENT LEARNING SITUATIONS - AN ANALYSIS................................... 10.2.1 Intra-company Intra-functional Learning...................................................................... 10.2.2 Intra-company Inter-functional Learning...................................................................... 10.2.3 Inter-company Intra-functional Learning...................................................................... 244 246 249 250 10.3 LEARNING IN PRODUCT DEVELOPMENT - SUMMARY AND CONCLUSIONS.......... 10.3.1 Summarizing the Research Findings on Learning......................................................... 10.3.1.1 Transfer Mechanisms........................................................................................... 10.3.1.2 What Makes the Difference between Single and Double-Loop Learning............... 10.3.1.3 Factors Blocking Learning................................................................................... 10.3.2 Benchmarking the Research Findings on Learning....................................................... 256 257 257 258 259 260 11 CORE CAPABILITIES ......................................................................................................... 263 11.1 CORE CAPABILITIES - AN EMPIRICAL FRAMEW ORK................................................. 264 11.2 HOW CORE CAPABILITIES EMERGE - AN INTERPRETATION.................................... 267 11.3 CORE CAPABILITIES - SUMMARY AND CONCLUSIONS............................................. 272 9 12 SUMMARY, PRESENTATION OF THE INTEGRATED MODEL, AND GENERAL CONCLUSIONS......................................................................................................................... 275 12.1 THE RESEARCH PURPOSE............................................................................................... 275 12.2 RESULTS : THE LEAN PRODUCT DEVELOPMENT CONTEXT..................................... 12.2.1 The Place and Role of an Expert Supplier in the Supply Chain..................................... 12.2.2 Driving Forces for Change........................................................................................... 12.2.3 Change and Development Tendencies in the Product Development Process.................. 12.2.4 Use and Perception of Lean Techniques........................................................................ 12.2.5 The Lean Product Development Context - Concluding Remarks................................... 277 277 278 279 280 281 12.3 RESULTS: THE PRODUCT DEVELOPMENT PROCESS.................................................. 12.3.1 Causal Condition - Means of Guidance......................................................................... 12.3.2 The Core Category - Operational Design...................................................................... 12.3.3 Context - Design Support Structures............................................................................. 12.3.4 Intervening Condition - Learning in Product Development........................................... 12.3.5 Consequence - Core Capabilities.................................................................................. 12.3.6 An Integrated Model for the Product Development Process........................................... 282 283 284 287 290 292 294 12.4 LIMITATIONS OF THE RESEARCH AND AREAS FOR FUTURE RESEARCH............... 297 POSTSCRIPT............................................................................................................................. 299 BIBLIOGRAPHY...................................................................................................................... 300 APPENDICES............................................................................................................................ Appendix 1 Project Resumé.................................................................................................... Appendix 2 Interview Topic Guide.......................................................................................... Appendix 3 Case Study Topic Guide....................................................................................... Appendix 4 Documents Used in the Case Studies.................................................................... Appendix 5 Presentation of the Case Study Companies and List of the Interview Companies.. 310 311 315 323 327 329 10 LIST OF FIGURES Figure 1. The automotive supply chain. After Brocquet, (1995).................................................... 14 Figure 2. Integrated Component Development.............................................................................. 20 Figure 3. The components of industrial models. Source: Boyer & Freyssenet (1995)..................... 42 Figure 4. The components of industrial models, extended version. After Boyer & Freyssenet (1995).......................................................................................................................................... 43 Figure 5. The product development framework............................................................................. 45 Figure 6. The product development framework, supplier case....................................................... 45 Figure 7. The Lean Enterprise. Illustration of the concept developed by Womack & Jones (1994). 59 Figure 8. Perspectives on Lean Production.................................................................................... 64 Figure 9. The Japanese supplier system........................................................................................ 80 Figure 10. Definition of the research problem............................................................................... 122 Figure 11. Deductive vs. inductive research. After Pras & Tarondeau (1979)................................ 124 Figure 12. A Process Model for Methodological Choice................................................................ 127 Figure 13. Organization of field activities..................................................................................... 131 Figure 14. The paradigm model.................................................................................................... 137 Figure 15. Different customer relationships facing an expert supplier........................................... 145 Figure 16. Direct arm's-length supply........................................................................................... 146 Figure 17. Indirect expert supply.................................................................................................. 148 Figure 18. Direct expert supply..................................................................................................... 149 Figure 19. The 'Golden Triangle' of Quality, Cost, and Lead-time................................................ 153 Figure 20. A proposed model for operational design work............................................................ 174 Figure 21. Individual learning, after Kim (1993, p. 40)................................................................ 179 Figure 22. Actors and learning situations in the product development process.............................. 236 Figure 23. Inter-company intra-functional learning in product development................................. 255 Figure 24. How core capabilities emerge - a model integrating research findings with the theoretical frameworks of Bowen et al (1994) and Arrègle (1995)................................................ 269 Figure 25. An integrated model for the product development process............................................ 295 11 LIST OF TABLES AND EXHIBITS Table 1. Condensed list of variables for analysing buyer-supplier relationships............................. 21 Table 2. Selection criteria in the SARA data base, (Source: CRESAL-CNRS, 1996)..................... 35 Table 3. Global Results, turnover and employees, (Source: De Banville et al, 1997)..................... 36 Table 4. Distribution of the number of employees, (Source: CRESAL-CNRS, 1996)..................... 36 Table 5. Activity sectors eliminated from the data base................................................................. 37 Table 6. Component suppliers per product category and employee distribution............................. 38 Table 7. Turnover and number of employees in specialized supplier firms in the French RhôneAlpes Region (68 companies)....................................................................................................... 39 Table 8. Specialized suppliers per product category and employee distribution (number of companies)................................................................................................................................... 40 Table 9. Mass vs. Lean Production: Characteristics. Summary of characteristics from Berggren (1990, 1992), Bouchut (1990), Chanaron & Lung (1995), Clark & Fujimoto (1991), Kawamura (1994), Krafcik (1988), Monden (1983), Womack et al (1990)..................................................... 50 Table 10. Application-adaptation evaluation form, source: Kawamura (1994, p. 27)..................... 60 Table 11. Outline of research topics and data collection methods.................................................. 129 Table 12. Messages from the pilot studies..................................................................................... 132 Table 13. Characteristics of three different customer relationships................................................ 151 Table 14. Five action strategies in product development management for responding to change.. 162 Table 15. Seven learning orientations and ten facilitating factors (After Nevis et al, 1991, p. 77). 243 Table 16. Summary of research findings on learning.................................................................... 257 Table 17. Specification and analysis of seven learning situations (after Nevis et al, 1995)............ 261 Exhibit 1. Perceptions of Lean Production. Summary of the central messages in Abo et al (1994), Krafcik (1988), Lamming (1993), Womack, Jones & Roos (1990).................................. 52 12 1. RESEARCH PROBLEM This first section of the thesis has four objectives: to introduce the research topic and the core problem; to delimit and define the research problem through a discussion and presentation of basic assumptions and research questions; to discuss the research process and the research perspectives; and finally to discuss the anticipated contributions of the present research. 1.1 INTRODUCING THE RESEARCH PROBLEM As a point of departure the research set out to examine what very broadly can be called the operating aspects of lean production implementation in medium-sized specialized component supplier firms in the French automobile industry. Based on initial knowledge in the field (theoretical studies and readings, as well as working experience and field observations), the research problem took shape to become one of exploring how supplier firms are experiencing changes, focused on the more active part that they are supposed to take in the design and development process, and more precisely one of describing how the organizational structure is adapted to support this change. An additional objective was to assess to what extent different practice in the product development process seems satisfactory for reaching the objectives that are set. During the 1950s and 60s the French automobile supply industry was organized into two large and relatively stable groups; subcontractors -an externalized production capacity with no design department-, and suppliers 'on catalogue' -with limited design capacity, and no possibility of offering customer-tailored solutions- (Laigle, 1994). Concerning medium-sized suppliers, this binary typology, from which new relationships have emerged, is particularly relevant (larger suppliers have traditionally had important innovation and design capacity, and tend to remain in direct contact with the carmakers). Among this group of suppliers, the former subcontractors tend to descend to the second or even third tier, or to be purchased and integrated as production capacity by large firms. The former suppliers 'on catalogue' tend to make great efforts in developing their capabilities (product development, just-in-time delivery, quality systems, and management of their own supplier base, to mention only the most significant) in order to become partnership suppliers (Chanaron, et al, 1993; Laigle, 1994). Also smaller firms that are rich in technology and have a long experience of direct supply will continue to be a natural part of the responsibility sharing in lean product development - so long as they are ready to add new competencies (Lamming, 1993). Figure 1 illustrates the generally accepted hierarchy, with its related vocabulary, in the French automotive supply chain. It takes into account the transformation described above. 13 Car Manufacturers System Suppliers Specialized Suppliers Ordering Sub-contractors Components Figure 1. The automotive supply chain. After Brocquet, (1995). This is a generic definition that will be further analysed and criticized both in the literature review and in the results section. However, at this moment it is very useful for defining some of the terminology that will be used throughout the research. The different kind of suppliers can be described in the following way (Brocquet, 1995; Laigle, 1994; Lamming, 1993): • System suppliers provide carmakers with pre-assembled subsystems, for example dashboards, exhaust systems, seats or front ends, consisting of a large number of externally purchased parts with different technical content. Besides R&D, these companies are specialized in manufacture and assembly. They tend more and more to locate assembly units close to or even in direct continuous flow connection to car assembly sites. In the new production organization, they take on a large responsibility for managing subcontractors. System suppliers are large (often several thousand employees), divisionalized companies, present in several markets. Here are found the former large suppliers on catalogue. • Specialized suppliers are medium-sized companies (normally between 50 and 500 employees) that supply system suppliers and in some cases carmakers with technical components. These companies concentrate on product development and manufacturing, they are mainly local but tend to set up subsidiaries or form alliances in new markets. Specialized suppliers develop and manufacture for example engine components, heating and cooling components, fastening devices, plastic parts, and electronic components. Here one finds the former medium-sized suppliers on catalogue. 14 • Subcontractors are small-sized companies (up to 50 employees, in exceptional cases more) specialized in the manufacturing and execution of finished blueprints. Their main customers are specialized suppliers and system suppliers. They supply parts with low technical content such as bolts, nuts, and simple screw cut parts. The present research concerns one specific group of specialized suppliers namely those that continue to maintain direct contact with the carmakers in the product development process. These are often referred to in the industry as expert suppliers1 . At an overall level, expert suppliers can be defined as specialized suppliers that possess specific capabilities (product technology, process technology, human capital) that are of crucial importance for carmakers and that the latter either cannot or do not want to integrate. In the auto industry, expert suppliers represent a specific interest when the industry is restructured. One might assume that they reap substantial benefits from smaller and less formal organizations, and from greater reactivity compared to larger firms. However, one might also assume that they have great difficulties in enlarging competencies and developing the capabilities required for the new integrated product development process, and in managing a complicated set of situations and activities that follow. The particular situation, the advantages and the problems of this group give a first reason for the choice of research subject. Moreover, the existing body of automotive management literature tends to be focused on system suppliers - that by definition preserve a first tier definition. It must immediately be made clear that the expert supplier status is a dynamic one. In this situation it is necessary for a supplier to evolve permanently. It will be argued that the direct contact with the carmakers can be easily lost. Conversely, new entrants can come into this group. The research is organized around two main topics: lean production in the supplier firm, and the organization of product development and the product development strategies in the lean context. 1.2 LEAN PRODUCTION IN THE SUPPLIER FIRM Lean production has been identified as a particular way of organising work, making use of technology, managing relationships between customers and suppliers, streamlining the product development process, dealing with customers, and so on. Lean production is 'lean' because "it 1 This notion was revealed during pilot interviews with managers in both carmaker and supplier firms. Lamming (1993, p. 218) indicates that smaller firms that are rich in technology and have a long experience of direct supply will be a natural part of sharing innovation, cost and quality responsibility. He refers to these as experts in one or two product types. 15 uses less of everything compared with mass production - half the human effort in the factory, half the manufacturing space, half the investment in tools, half the engineering hours to develop a new product in half the time. Also, it requires keeping far less than half the needed inventory on site, results in many fewer defects, and produces a greater and ever growing variety of products" (Womack, Jones & Roos, 1990, p. 13). Even if the report of Womack et al has been criticized for overestimating the generality and novelty of the lean model, and for arguing too much for a 'one best way' (Berggren, 1992; Boyer & Freyssenet, 1995), it is the reference that has conceptualized lean production from practice in the Japanese auto industry and introduced it to a broad audience. This concept has provoked turbulent and rapid changes in the way of organising production in both the European and American manufacturing industry (Boyer & Freyssenet, 1995; Womack & Jones, 1994). Core features in lean production, summarized from Clark & Fujimoto (1991); Bouchut (1990); Kawamura (1994); Krafcik (1988); Monden (1983); and Womack et al (1990) are: • Increased concentration on core activities and outsourcing of component design and manufacturing; • Integrating rather than coordinating operational work, both within the firm between different functions, and as a consequence of outsourcing, across company boundaries in the production chain; • Increased cooperation between different actors, i.e. between different professional groups and across hierarchical levels in flatter organizations; • Emphasis on transversal processes and project management; • Quality assurance and total quality management; • Flexibility in the production tool as well as in personal competencies; • Low inventory levels and just-in-time delivery; • Time as a key success factor (reduced development lead time and reduced delivery lead time); • Focus on employee participation in continuous improvement activities. It is, however, relatively difficult to know to what extent these methods and concepts are used in different companies, how they are used, and if they are applicable to different kinds of supplier firms. The first aim of the present research project is to establish the context of lean component development, i.e. answer questions such as: What are the most significant changes that the medium-sized expert suppliers have gone through and are going through? What are the drivers for change and to what new practice and methods have they led? What are the main priorities for the near future in terms of 16 product development strategy and organization? Focus will be on the outcome, i.e. on established new ways of working and on anticipated changes more than on the change process itself. Throughout the research it will be argued that lean production is a production model that represents a fundamentally different way of organising production within and, above all, in the combination between three elements of production organization: product development, supply management and manufacturing/assembly. As the research deals with component development, it will concentrate on change and interaction in and between product development and supply management. However, it is also recognized that at the manufacturing level, highly standardized tasks and short repetitive work cycles remain in lean production. At this level, in the work organization, lean production is more a "refinement or perfection of traditional mass production" (Ellegård et al, 1994, p. 114) than an alternative model. Up until now, research on lean production has been focused on proposing and describing concepts and ideas, and has dealt less with questions and answers concerning their application. Important authors such as Abo et al, (1994), Clark & Fujimoto, (1991), and Womack et al (1990) conduct their analysis in a strategic management perspective, setting up target models for managing product development, supplier relations, and industrial relations. However, at the strategic level, one neither learns how to achieve the goals that are set up, nor what will be the impact of different actions in relation to existing practice in companies. Moreover, most references, including the ones mentioned above, study the emergence of lean production mainly from the automakers' perspective. The ambition in the present research project is to take a dynamic perspective on the emergence of a new industrial model, to describe and analyse the developments and outcomes in an operational perspective, and, as has already been identified, to concentrate on a supplier perspective. The research will try to go beyond a description of problems with diffusion and implementation of lean production and focus on questions such as: how are the methods and concepts associated with lean production perceived by operating people (design engineers and technicians)? To what extent do lean principles act as driving forces for change? How do new concepts interact with already existing solutions? In the first phase of the research, the application, evolution, and perception of changes related to lean production will be identified through data collection under four themes: • • • The place and the role of the firm in the supply chain, The drivers for change, The change and development tendencies in the product development process, 17 • The use and perception of lean techniques. Two results are expected from this first phase. The theoretical result concerns the establishment of a context for current car component expert supply. To what extent are lean principles present in supplier firms, and to what extent are firms influenced by such principles? The objective is not to provide a general answer to this question, but to compare the actual situation in a limited number of cases to the existing theory. It is a question of analytical generalization (Yin, 1989); "the investigator strives to generalize a particular set of results to some broader theory." (p. 44). The managerial result concerns how the expert suppliers can manage lean design in order to develop a favourable position in the supply chain and turn the change process into a competitive advantage. After the introduction to the context of lean component supply, the main focus of the research will be made on the second phase, focusing on the organization and processes in the product development process that emerge in expert supplier firms under the influence of lean. Lamming (1993) argues that R&D practice provides the most important link between lean customers and suppliers, since it is one of the clearest manifestations of collaboration on the part of both partners. 1.3 LEAN DEVELOPMENT: FOCUS ORGANIZATION AND STRATEGIES ON PRODUCT DEVELOPMENT One feature in the lean production model -which is central for the present research- is integration between different actors in the product development process. This concerns both different functions inside a firm and inter-firm integration between companies in a production chain. Integration also characterizes the theoretical model of the lean enterprise proposed by Womack & Jones (1994). They emphasize both intra-firm and inter-firm integration: day-today activities in engineering, purchasing, and manufacturing should be performed in interfunctional project teams, where the members are under the authority of a transversal process management function. Traditional specialist functions should develop new roles, acting like support systems, systematically summarising current knowledge, searching for new knowledge, and teaching all this to their members. Furthermore, they should develop guidelines for best practice in their domain, and they should, in collaboration with their counterparts in companies up and down the production chain, develop rules for governing common problem solving and establish behavioural codes. Clark & Fujimoto (1991) and Karlson (1994) treat thoroughly the aspects of integration and coordination, emphasising in turn overlapping stages in the product development process, the 18 need for intensive communication, and the need for different actors to work together on common problems to overcome differences in professional languages. Their research into integrated product development is focused on internal integration (within the assembler's organization) mostly between product and process engineering. In this context, customer-supplier relationships are changing from a traditional model characterized by very restricted data and information exchange, the one-sided role of R&D (either assembler or supplier), and sourcing decisions based on wide enquiries, lowest bid and price, to a partnership model characterized by two-way and long-term data and information exchange, shared R&D, and sourcing decisions based on performance history, long-term collaboration and cost (De Banville & Chanaron, 1985; Lamming, 1993). Focusing on increased supplier involvement in the product development process means that to the important integrative activities between product engineering and process engineering is added integration with purchasing, and that to each of these functions, and to the links between them, are added links with the supplier. This is defined as integrated component development, see figure 2. The present project is concerned with coordination and integration within the design process in supplier firms, i.e. between specialists who perform the activities that go into realising the component: design engineers and technicians, project managers, directors of development departments, process engineering staff, quality assurance managers, and purchasing personnel2 . Both internal and external coordination and integration will be analysed on the basis of the different aspects outlined above. It is not without reason that the supplier oval is empty in figure 2. Little has in fact been said about how lean component development and the coordination and integration processes are reflected in supplier firms from their perspective. 2 These actors have been identified in the empirical research. 19 Carmaker Supplier The Customer - Supplier Interface Product Engineering Purchasing Process Engineering Focus of the research project Figure 2. Integrated Component Development. The aim of this research is to address the issue of integrated component development from the medium-sized supplier's perspective. The following questions are central for the research: what organizational structures are developed to pave the way for effective integration? What are the tools and methods used to support this, and which are the ones used in the interface with the client (imposed or negotiated)? How are parallel operations organized? How is "intensive communication" developed and managed? And how are positive attitudes to integration developed? As in the case of lean production, research in the area of buyer-supplier relationships has until now been focused on the development of general frameworks and on the conceptualization of a large number of factors that are changing in these relations with the development of lean production. Both qualitative research (e.g. Chanaron et. al, 1993; Laigle, 1994; and Lamming, 1993) and quantitative surveys (e.g. Cusumano & Takeishi, 1991; Helper, 1991 and 1994) adopt very broad frameworks for identifying models of supplier - assembler relationships. The objective might be to indicate strategies for going lean, or to compare Japanese, American, and Japanese transplant auto-plants by mapping differences concerning supplier relations and management. Table 1. summarizes the variables used in the above mentioned references. - The nature of competition in the components supply market. 20 - The basis upon which sourcing decisions (selection of suppliers) are made. - The role played by data and information transfer and exchange. - Ordering and delivery practices - including inventory management. - The role of product and process R&D in the relationship. - Quality assurance; defect rates. - Productivity implications of different strategies; productivity measurement. - Number of suppliers per part. - Length of contracts. - Pricing practices. - Supplier suggestions. - The level of pressure in the relationship. Table 1. Condensed list of variables for analysing buyer-supplier relationships. For both managers and scholars, this kind of research has identified a new area of highest strategic importance and academic interest. Practitioners have been provided with frameworks that analyse the differences between traditional arm's-length price-bargaining supplier relations, and the partnership model - relating the latter to higher performance (i.e. shorter development lead time, better product quality, and increased productivity3 ) in the lean production model. However, little has yet been said about the transformation of these strategies into working practice, i.e., what does for example joint R&D mean for a design engineer in the supplier company4 ? What organizational procedures must be in place, and what skills does he/she have to develop to work jointly with colleagues from the assembler company or another supplier?5 Focusing on the development process of components, this research aims at complementing more macro-oriented research. The objective is to develop an understanding of integration by looking into the organization at an operational level, emphasising questions such as: How, in practice, is integrated component development being processed at an engineering level? How do operating people perceive and influence the organization and its development? 3 Clark & Fujimoto (1991) define productivity as the level of resources required to take the project from concept to commercial product (p. 69). Productivity is measured in engineering hours which are the sum of hours spent on concept generation, product planning and product engineering. For details on how development lead time, product quality, and productivity can be measured see Clark & Fujimoto (1991) pp. 374 - 381. 4 What, for instance, does a supplier design engineer of rear lighting have to know about how rear bumpers are assembled to avoid late changes and redesign work? 5 It is assumed that in medium-sized supplier firms, problems with internal coordination are less important due to the limited size of the companies. There is for example little risk that different functions are located in different sites or divided into divisions. 21 Concerning the ongoing change in relationships between automakers and suppliers, Helper's (1991 and 1994) quantitative, US-based research addresses several interesting issues, for example the importance of information exchange and commitment in a partnership supplier relations. These two features are linked to each other, because rich information flow "both requires and engenders a high degree of commitment to the relationship" (Helper, 1991, p. 16). Three reasons for this are argued: (1) extensive communication systems are costly to establish and maintain with more than one (or a few, my comment) supplier(s) for the same component or subsystem, (2) exchanging proprietary information requires trust, and (3) by working together over time, customers and suppliers can gain substantial benefits from knowledge of each other's products and processes. Through an analysis of information provided to customers (breakdown of process steps, cost of each process step, SPC charts, and production scheduling), formal commitment (number of suppliers, and length of contracts), and informal commitment (perception of customer fairness, and customer reaction if a competing supplier comes up with a superior component at a similar price), Helper (1991) concludes that concerning information flow the situation is moving towards one of partnership, but that, on the other hand, suppliers do not think that customers have become more trustworthy. Suppliers do not receive much assistance in reducing costs or adopting new techniques, and they are not convinced about the efficacy of Just-in-Time. In other words, the level of integrated problem solving and/or the outcomes of such efforts are still far from being at the level of expectation. These trends are confirmed by surveys in 1983, 1989, and 1993. As late as in 1993 (Helper, 1994), supplier relations were at a cross roads where a compromize between pressure for short time cost reduction (leading to traditional short-term price bidding), and the goal of long-term competitiveness (promoted by a partnership strategy) was found. Helper (1991) states that changes must occur in all functions of the firm in order to achieve the benefits of a partnership strategy: purchasing needs to emphasize other criteria than price, production needs to ensure stable schedules, and engineering needs to establish mechanisms for incorporating suppliers' design suggestions. These remarks refer to the automaker's organization, but little is said about the suppliers' possibilities to influence the situation. This research will explore the structures and processes of supplier-customer integration from the suppliers' perspective and analyse, on the one hand, what suppliers can do in their own organization, and, on the other hand, if and how they can influence the behaviour of their customers to develop a true partnership based on an equal's contribution. The main question in this context is whether common characteristics for good performance can be found, and the objective is to provide a mapping of the 22 determinants for successful development of design capabilities and related organizational design within the framework of integrated component development. To summarize, this brief introduction to the literature indicates that both integrated product development, and a partnership strategy in supplier relations are based on parallel operations and rich information flow, working through tools and methods on the one hand and commitment based on trust and knowledge on the other. In the present project, the empirical research will map these components through case-based interviews and direct observation in search of frameworks describing the interrelationship between new customer relationships and the product development process -including the aspect of the result, i.e. the product technology- in medium-sized expert supplier firms. It will examine if and how joint operations between suppliers and assemblers are formally established and organized, what tools and methods are used to facilitate interaction, and if new organization can improve also less tangible determinants for efficient integration such as attitudes, different professional languages etc. The questions related to product development organization and strategies will be addressed through data collection under the following themes: The evolution of product technology and product functions; • The organizational structures and work processes; • The coordination activities and communication structure. • Focusing on the operating characteristics of component development, the main result expected from this phase of the research is to determine management implications of increased integration in the component development process. It is also to propose a conceptual framework for the situation that medium-sized expert suppliers are facing after the turbulent changes in their role. In keeping with a grounded theory approach this will be done through an inductive line of action (the framework evolves during the research itself) comparing empirical findings with literature proving itself to be relevant to the gathered data (Strauss & Corbin, 1990). Before undertaking research focused on determinants for successful integration, it is highly relevant to ask why integration between different functions and actors in the product development process seems to be something that managers ought to develop in their organizations today. This is necessary so as not to fall into the trap of ends-means confusion. Integration is a means to an end, not an end in itself. 1.4 BENEFITS FROM INTEGRATION 23 Firstly, during previous studies and pilot field interviews, integrating different functions, and even different activities within the same function emerged as a central problem - both to manage and to live with at an operational level. For example, in one medium-sized company (outside the car industry), that was observed for almost two years (outside the context of the present research), a consequently managed programme for cross-functional coordination emphasising employee commitment to new strategies, and internal customer-supplier relations resulted in a 20% reduction of the average throughput time in the second year. The key factors of success were: cross functional learning through team work, and the development of procedures to consequently feed back information and improvement propositions from the operational level to the management level and back again (Birchall et al, 1994). 1.4.1 The Carmaker's Perspective In a major European car manufacturing company, a design engineer, interviewed in the pilot phase of the project, explained the problems caused by sequenced work and poor communication between body engineering and tool making. Here again the development of a more integrated form of information handling was a process of identifying internal customer needs. For example, to order their castings, the tool making department only needed blueprint outlines from body engineering, as they defined castings as either small, medium or large. Body engineering, not being aware of this, did not inform tooling about the basic dimensions, number of panels, number of press lines, etc. at an early stage, but waited until complete, fully dimensioned drawings were ready. This meant a time delay for casting ordering of four to five weeks. Such dysfunction came into light as the company reorganized its development process to cope with high cost (vehicle cost per man rate), poor quality (results from new car buyer surveys), and long development lead time (engineering hours). The reorganization demanded cross fertilization of functions which was perceived as a very difficult objective to accomplish. The fact that earlier involvement and increased design responsibility for suppliers in product development improves development lead time, productivity, and product quality has been advanced in several research reports (Clark, 1989, Clark & Fujimoto, 1991, Cusumano & Takeishi, 1991, Dyer & Ouchi, 1993, Richardson, 1993). For example, Clark & Fujimoto (1991, p. 225-228) found a moderately positive correlation for both engineering lead time and development productivity in relation to simultaneity, i.e. in relation to stage overlapping and intensive communication in product development. Also total product quality showed a similar pattern. They conclude that integrated problem solving "not only allows parallel operations resulting in improved lead time, but also reduces errors and waste that consume engineering hours and jeopardize product reliability" (Clark & Fujimoto, 1991, p. 228). 24 When measuring different performance parameters, it is of course difficult to say in what proportion a higher level of supplier engineering input contributes to better performance compared to other steps and actions such as stage overlapping and improved internal integration in the assembler's organization. Concerning product development lead time, Clark (1989) argues that simply outsourcing work to suppliers will not reduce lead time unless the parts concerned are on the critical path in the development project (and the supplier involvement permits parallel development of the component), and unless suppliers bring additional capability to the project (the supplier performs the tasks cheaper and better than internal development). Moreover, the gains from supplier involvement can be offset due to additional coordination time and difficulties in managing the supplier relationships. Thus, the possible benefits depend on successful integration of working cycles. This, once again, justifies the present research project that aims to study the organizational aspects of integrated problem solving. 1.4.2 The Supplier's Perspective As it has just been argued, partnership supplier relationships, with a high level of supplier engineering and integration of work processes, can be favourable for assemblers. More difficult to grasp are the consequences of the outsourcing tendencies for suppliers, above all for smaller ones that cannot develop large scale system supply. Helper (1994), and Sako, Lamming & Helper (1994) have conducted one of the most extensive surveys of first-tier automotive suppliers in the context of partnership6 . Firstly, their results show that a transfer towards the partnership model is really happening. For example, in 1993/94 68% of 250 surveyed continental Western European suppliers provided their customers with detailed breakdowns of their production process compared to 42% in 1989-90. If used properly, such information should help vehicle manufacturers ensure that their component designs are compatible with suppliers' processes, thus improving productivity and quality. Moreover, in 1993/94 53% of the continental Western European suppliers thought that their customers would help them to match a competitor's effort if the latter offered a lower price for a product of equal quality. This figure was 33% in 1990. Concerning partnership and performance, the data from Sako et al (1994) show that partnership suppliers tend to perform better in all regions: they receive more quality rewards, they enjoy a higher growth in market share, they control their internal costs better, and they defend their profit margins better. However, supplier margins fell in all regions between 1989/90 and 1993/94, except for Japanese suppliers with partnership relations. Thus, 6 In all, 1.400 suppliers in four regions (USA, UK, Continental Europe, and Japan) were surveyed. 25 customers often obtain price reductions by reducing supplier margins rather than supplier costs. Moreover, just-in-time often comes at the suppliers' expense: over half the US suppliers, nearly half of those in Europe, and a third of those in Japan agree with the statement "JIT only transfers inventory responsibility from customers to suppliers". This means that JIT delivery isn't matched by JIT production. Depending on the type of component, on the nature of the production organization, on a certain lack in competencies for flexible production, or a combination of these, many suppliers produce relatively long series of components often resulting in week-long inventories before final delivery. The main problem, thus, is to match JIT delivery with JIT production. Again partnership suppliers perform better than non-partnership ones. Of the European suppliers, the former group have a production inventory that is two days shorter, and a delivery inventory that is one day shorter than the latter. If these results show that partnership and cooperation are beneficial also for suppliers, the problem is that relatively few suppliers have such relations with their customers and that many still suffer from the restructuring of the business. In the UK, for example, a 1994 DTI/SMM&T7 report on relationships between vehicle manufacturers and their suppliers concluded that the national industry as a whole suffered from poor communication, a general lack of trust between individuals and between companies, and a resultant loss of opportunity for generating mutual benefits from collaboration. Vehicle manufacturers were found to be preoccupied with rationalising their own supply base at the expense of contributing to the general development of the UK components industry. French data (De Banville et al, 1997) show similar tendencies. Through these observations, one comes back to the issues of coordination, integration, and communication raised above. Questions concerning specific problems for suppliers are also related to these issues, for example, questions of trust and the possibility of an open discussion on the sharing of 'pain and gain'. Before proceeding to a review of existing literature in the fields of lean production and supply management, the frame for the technical aspects of the research - the research process with its assumptions, research questions, methodology and perspectives will be set. 1.5 DEFINING THE PROBLEM 1.5.1 Initial Assumptions The research set out with the following assumptions, grounded in a brief literature survey concerning product development and supply management, and initial field observations: 7 The UK Department of Trade and Industry. 26 1. In the car industry, a general change in the way of organising production, inspired by managerial techniques developed in Japan -further conceptualized as lean production- is taking place and is even beginning to stabilize. 2. Lean production is not simply a transposition of Japanese practice (and if it was, such a transposition would be extremely difficult), but it is an emerging new model where managerial techniques developed in Japan, concerning various parts of the production organization, are acting as eye-openers and drivers for change, and interact with existing practice. 3. This means firstly that component suppliers are gaining increased importance as design and development resources. This increased importance concerns a wider range of suppliers than simply system suppliers, above all medium-sized ones that are rich in technology and have a long experience in direct supply - so called expert suppliers. 4. It means, secondly, that integration becomes a core element in this new form of industrial organization; the production chain will be managed more and more as an integrated industrial process even though separate actors (e.g. suppliers) remain independent. 1.5.2 Overall Research Questions As an immediate consequence of the initial literature review and assumptions, six overall research questions emerge and are formulated in the following way: 1. How does the emergence of new industrial principles take place in expert supplier firms? 2. What is the place and the role of expert suppliers in the automotive production chain? 3. What lean production techniques are used, and are managers satisfied with their application? 4. How in practice is integrated component development realized? 5. How are work processes designed to support integration? 6. Are there common characteristics between companies forming a 'high performance' organizational structure for lean component development? 1.6 RESEARCH PROCESS AND PERSPECTIVES The overall research questions presented above are to a large extent either 'what' questions of an exploratory kind or 'how' questions. The objective of the research is to explore and describe, to focus on meanings, to try to understand what is happening, to look at the 27 interconnections between different situations studied, and to develop ideas mainly from an inductive data analysis process. This is what generally characterizes qualitative research based on a phenomenological position (Easterby-Smith et al, 1991). According to Easterby-Smith et al (1991), and Yin (1989), a relevant research strategy for this kind of problem, in which the boundaries between phenomenon and context are not clearly evident, is the use of case studies - small samples investigated in depth over time -, and the use of multiple data collection methods to establish different views of a phenomenon. In order to develop these questions further before the field study, a literature review will be undertaken in chapter three and four. In line with a qualitative approach based on induction from data, the objective for the literature analysis is to further justify the importance of the research and to inform and guide the research; such a research design is chosen in order to improve the clarity and focus of the study (Miles & Huberman, 1994). Moreover, in order to establish the context of lean product development in the studied firms, a literature-based definition of this concept is necessary. After the literature review, the overall research questions will therefore be revisited. Definition of the research object or the unit of analysis is another essential step in any research project. As a supplier perspective is taken, the unit of analysis is defined to be the product development process in each studied supplier company. A piece of research must also be representative of something8 , i.e. one must consider questions like: Should a certain industry be studied? Are there national differences? How does the different size of the companies affect the results? What about the products that they manufacture? And what is the firm's place in the production chain? As will be argued in chapter five, the definition of a research problem is influenced from at least two sides: from the philosophical position held by the researcher, and from the real world context. The latter, it will be argued, is the managerial situation in a company, or in any other organization where managerial problems occur, where the research will be carried out, and where the researcher has some orienting knowledge - from working experience and/or observations and readings. Hence, the questions raised above were answered in the following way: • The car industry was chosen as the field for observation because the development of lean practice has reached such a level that the consequences in, and impacts on, operational functions are seen as beginning to stabilize in this sector. This means an opportunity to study organizational adaptation to identified changes. 8 See for example Karlson's (1994) discussion of method and scope. 28 • Concerning size, medium-sized suppliers were chosen because relatively little has been said in relation to lean product development from their perspective. Moreover, the development of their design capacity is essential for regional and small scale competence development, for specialization in complex products that the industrialized countries must focus on and, more closely related to this project, for the contribution to new problem solving methods in component development. • In relation to the product and the place in the production chain, only companies that are direct suppliers to carmakers, that supply at least three manufacturers, that historically have had their own design department, even if very small (i.e. that are former suppliers on catalogue), and that have developed their capabilities for integrated component development as a result of an explicit strategy for R&D and organizational development, were included in the sample. Concerning the size of the companies, a number of qualitative criteria are discussed in the literature besides the quantitative measurement such as number of employees and/or turnover. Examples are owner structure, means of financing, organizational structure (including the role of the managing director), innovativeness and strategic flexibility (APRODI, 1985; Bamberger, 1978; and Derrouch, 1983). When the problem is to define a sample in a research project, it seems clear that in addition to some basic definition, such as the number of employees, the other criteria become dependent on the objective and research questions. The present research is focused on the product development process under the influence of lean production. In the selection of companies it was therefore important that they were organized as independent companies, designing, developing, and manufacturing their products, and that all functions (general management, finance, R&D, quality assurance, human resources, production, and marketing & sales) were managed independently. As a consequence, each company in the sample manufactures one specific group of components, either sharing a core technology or fulfilling a core function. They are not organized in product divisions, as the system suppliers often are. Concerning owner structure and means of financing, the main requirement was an operational independence concerning the product development process in the studied unit. Even if some companies were financially tied to a larger group, the fulfilment of this requirement could qualify them for being part of the sample. For one of the case study companies that belonged to a group within the automotive industry, the company's strategy depended on a few general group guidelines that will be mentioned in the case description. However, these were so general that they did not have any influence at all on the company besides the ones exerted by the general movement in the sector. 29 Concerning organizational structure and strategic flexibility, finally, the studied companies are characterized by fast information channelling and overlapping functional responsibilities that (at least theoretically) allow for a high reactivity to change. The empirical research took place in three phases. Firstly, interviews were conducted with the general managers in four supplier firms of the kind identified above (and in three of them also with the product development managers). These interviews focused on the formal organization of the company and of the design process in particular, on the place and role of the supplier in the production chain, and on changes and problems related to lean product development. Secondly, in-depth case studies were conducted in two of these supplier firms. Thirdly, follow-up interviews focusing on the issues revealed in the case studies were conducted with product development managers in four additional supplier firms. Further details on the research design will be given in chapter five. 1.7 SUMMARY AND ANTICIPATED CONTRIBUTIONS OF THE RESEARCH In the automobile industry, the increasing competitive pressure coming from Japanese carmakers has pushed both scholars and practitioners to seek a better understanding and knowledge about industrial relations, work organization and managerial practice in the Japanese industry. Intensive and longitudinal studies of the world auto industry has resulted in the formulation of the lean production concept (Womack et al, 1990) featuring distinctive differences compared to traditional mass production in product development management (Clark & Fujimoto, 1991), supplier relationships (Lamming, 1993), and production and operations management (Krafcik, 1989; Monden, 1983). The present research will focus on the interface between two of these domains; product development and supply management. The first problem deals with the context of lean product development. It will be of central importance to map the main drivers for change and the new priorities and working methods that have resulted. Parallel to this, an assessment of the perception of the new product development model will be attempted. Then, the organization and strategies in integrated component development viewed in the perspective of expert suppliers will be the focus of the research. The literature clearly indicates that there is a tendency for closer cooperation between buyers and suppliers, i.e. an integration of the design activities in different firms in a production chain. Some go as far as talking about partnership in buyer-supplier relationships. It has been demonstrated that an integrated approach to product design can be favourable both for carmakers and suppliers. The literature 30 indicates improvements concerning cost, quality, market position, and technical performance of the developed products as a result of this strategy. The research aims to develop an understanding of integration by looking at the organization at an operational level; how is integrated product development being processed and how do operating people perceive and influence the organization and its development? Special attention will be paid to the structures and processes of customer-supplier integration from the supplier's perspective. Concerning customer relationships, the concept of 'an equal's contribution' introduces questions concerning the possibilities of developing an optimal structure that takes both the priorities of the suppliers and the customers into consideration. The nature of the research questions and the complexity of the problem, i.e. the intention to take a holistic perspective on a managerial problem in a specific category of firms, calls for a qualitative research methodology and a research strategy based on case studies. The research is focused on an in-depth analysis and description of specialized supplier firms within the framework of lean product development and supply management. The following outcomes can be expected from this study: • Concerning industrial relationships, this group of suppliers is difficult to classify in the current model of industrial organization. Through the analysis, some hypotheses about how to refine this model should result. • Relatively little is still known about the state-of-the-art in lean production implementation at an operational level. This research will focus on the general impression and penetration of lean techniques in the studied companies in order to give a case example of how lean production is practised at firm level. • The focus of the research will be the internal organization of the supplier firm. The core question is what measures are taken and what strategies are developed in order to respond to the trend of increased integration in the design activities and a greater design responsibility taken on by suppliers - features that are related to the lean production model. The objective is to develop a model for organizational design and competence development in the product development process in supplier firms. The model will look for common characteristics and determinants for efficiently managing product development and improving the design capabilities in these firms. Priority will be given to the exploratory dimension of research, which means that interviews and direct observation will be only approximately guided by a framework of questions. However, in order to improve the relevance and timeliness of the research, a literature review, aimed at analysing the most significant approaches and theories concerning lean production 31 and changing buyer-supplier relationships and their impact on product development is undertaken. Finally, as management requires both thought and action, (Easterby-Smith et al, 1991) this piece of management research is intended to provide some tangible and useful conclusions for practitioners. 32 2 THE EUROPEAN, FRENCH, AND REGIONAL AUTOMOTIVE AND AUTOMOTIVE COMPONENT INDUSTRY IN FIGURES This chapter will describe the structure and importance of the European and in particular the French automotive components sector in order to set the economic context for the study. After a general description involving some key figures, the situation in the Rhône-Alpes region, where the research has been conducted, will be discussed in more depth. The chapter also explains the process of defining the sample of companies from which interview and case study companies were selected. 2.1 THE EUROPEAN CAR MARKET AND TENDENCIES TOWARDS VERTICAL DISINTEGRATION The automobile industry is considered in all industrialized and newly industrialized countries as a key sector in the manufacturing industry - the product is complex, and its manufacturing process involves a very large number of actors contributing with specialized technologies (Chanaron & Lung, 1995). Safety restrictions, pollution problems and customer pressure for better value for money push the automobile manufacturers to continuously search for better solutions in order to reduce costs and improve productivity and quality. In 1994, Western Europe represented 27% of the world wide automobile production (cars and small utility vehicles up to 5 tons) and 27,6% of the total number of registrations1 . Concerning France, the number of registrations has been oscillating around 1,8 million vehicles per year (cars only) for the last ten years. This represents around five percent of the worldwide registrations and about 16% of the European1 . Concerning production, the two French manufacturers, PSA (Peugeot and Citroën) and Renault represented 7,9% of the world automobile production in 1993-941 . As discussed in chapter one, a major tendency during the last ten years has been a vertical disintegration in the automobile sector leading to an increased importance of external supply in the total cost of a car. Relevant general figures on this evolution are difficult to obtain. Different sources use different measurements that are difficult to compare and cross-check. For example, the Boston Consulting Group (BCG, 1991) estimated the rate of vertical integration at 56% (including parts divisions) in the European car manufacturing sector and at 36% in the Japanese. These figures relate to 1991 and were calculated as value added on turnover. Ellison et al (1995) use another measure namely the rate of purchased parts that are 1 Source: Comité des Constructeurs Français d'Automobiles (French Automotive Manufacturers Association), quoted in Broquet (1995). 33 engineered in-house vs. by suppliers. Using this measure, they conclude that in the late 1980s, European carmakers relied on supplier engineering for roughly 35% of all purchased parts while the Japanese did so in 70%. If the trend moves to an application of Japanese concepts in the European industry, an increasing supplier engineering rate and a decreasing vertical integration rate would be expected to appear. However, the opinions are not unanimous on this point. The Boston Consulting Group indicates that the vertical integration rate diminished by at least five points between 1991 and 1993 (Brocquet, 1995), while Ellison et al (1995) indicate a stable situation in Europe. They estimate the rate of supplier engineering at around 35-36% in 1993. Concerning Europe it must be underlined that the situation is very heterogeneous. Firstly, the degree of vertical integration is highly different between companies. According to BCG (1991), Volkswagen, Renault and Mercedes-Benz are at the top with a vertical integration rate above 50% compared to Ford, General Motors and Fiat who do not exceed 40%. Secondly, some manufacturers prefer to remain at a high level of detail-controlled parts (where the carmaker is responsible for detailed design). Interview data in the present research indicate that the American subsidiaries, Ford and Opel/Vauxhall, tend to keep more of arm's-length supplier relationships (demanding execution of finalized drawings), while the French manufacturers seem to have opted for a more integrated approach to component design (suppliers are given free rein to develop their own functional solutions within a framework of cost, quality, and basic engineering data). Thirdly, outsourcing strategies were applied earlier by some manufacturers than by others, and with different degrees of determination (De Banville & Chanaron, 1991). In this context, what does the French automotive component sector look like? The following section will try to set the frame of this particular industry. 2.2 THE AUTOMOTIVE COMPONENT SECTOR IN FRANCE According to a study of the European automotive components industry conducted by The Boston Consulting Group (BCG, 1993), this industry accounted for around 560 billion French Francs2 and employed 940,000 people in 1992. The French industry accounted for 110 billion Francs (20%) and 144,000 people (15%). It was the second largest market after Germany. Concerning component demand, France was also second after Germany accounting for 18% of the total demand. 2 Of which the original equipment market represents 77% and the replacement market 23%. 34 The big players, i.e. the divisions of large foreign supplier groups and the French system suppliers, represent around three quarters of the total turnover in the sector (Brocquet, 1995). He identifies the following foreign implants: Rockwell, Dana, ACG, Bendix, ITT, TRW, Teneco (American); Lucas, TN, GKN, BBA-PLC, Laird (British); Magneti Marelli (Italian); and Bosch, Teves, Mannesmann, Siemens, Behr, Freudenberg (German). Some of the largest and most internationalized French system suppliers are: Valeo (clutch systems, motor and compartment temperature regulation, and compartment security equipment), Bertrand Faure (seats), Ecia (plastics, exhaust systems, seats), Sommer-Allibert (interiors, dashboards), and Plastic Omnium (dashboards, bumpers, fuel tanks). Deranlot (1993) has made an estimation of the number of suppliers that are direct suppliers to the carmakers. If one focuses on the component suppliers and eliminates system suppliers and the sectors tyres, screws and bolts, glass, metal treatment, and foundry, his data base contains around 420 direct suppliers on the French national level. This figure will be compared to the situation in the Rhône-Alpes Region described in the following section. 2.3 THE AUTOMOTIVE COMPONENT SECTOR IN THE RHÔNE-ALPES REGION This section is based on the SARA data base (CRESAL-CNRS, 1996; De Banville et al, 1997)3 which was the most recent and exhaustive data base on the automotive activity in the Rhône-Alpes Region at the moment of the study. It comprises 876 units implanted in the region corresponding to the criteria displayed in table 2. Total Turnover Superior or equal to 5 MF4 in 1994 Turnover in the Automobile Sector Superior or equal to 1 MF in 1994 or Superior or equal to 10% of total turnover Number of Employees Superior or equal to 10 in 1994 Table 2. Selection criteria in the SARA data base, (Source: CRESAL-CNRS, 1996). The Rhône-Alpes Region comprises eight departments (corresponding to the British counties): Rhône, Loire, Isère, Ain, Savoie, Haute-Savoie, Drôme and Ardèche. The total population 3 The data base was ready in the spring of 1996, but only presented publicly in January 1997. The present chapter makes use both of a preliminary report (CRESAL-CNRS, 1996) and the final report (De Banville et al, 1997). 4 MF: Million French Francs. 35 was 5,5 million in 1994. The most important agglomerations are: Lyon, Saint-Etienne, Grenoble, Bourg-en-Bresse, Chambéry, and Valence. The SARA database contains industrial companies, i.e. design and development companies (without manufacturing), manufacturers of tools, machinery and special equipment, component and parts suppliers (system suppliers, specialized suppliers and subcontractors) including tyres, batteries, painting and glass. Distribution and consumer sales of cars and components, and after-sales and repair activities are not included. The overall results concerning turnover and employees are presented in table 3. Item Number of Units Total Number of Employees 1994 Item 1994 876 Global Turnover (MF) 63 201 78 997 Turnover Mean (MF) 72 23 Employees Mean 90 Turnover Median (MF) Employees Median 38 Turnover per Employee (F) 800 052 Table 3. Global Results, turnover and employees, (Source: De Banville et al, 1997). The distribution of the entire sample according to size in terms of the number of employees and turnover is illustrated in table 4. As can be seen, a large majority (65%) are small-sized companies with fewer than 50 employees. Concerning turnover, it was less than 50 MF in 75% of the companies in 1994. The turnover limit under which 65% (the same percentage as that employing less than 50 people) of the companies were situated was 35 MF. Turnover (MF) Percent of Companies Number of Employees Percent of Companies < 20 45% 10-20 22% 20-50 30% 20-50 43% 50-100 11% 50-100 18% 100-200 7% 100-200 8% 200-500 5% 200-500 6% > 500 2% > 500 3% Table 4. Distribution of the number of employees, (Source: CRESAL-CNRS, 1996). In the French component industry, the Rhône-Alpes region is the second most important after the Paris region (Ile-de-France) both in terms of number of companies and turnover (FIEV, 1994). As indicated above, 65% of the companies in Rhône-Alpes have fewer than 50 employees. This figure is around 50% at the national level (FIEV, 1994) which indicates a 36 larger domination of SMEs in Rhône-Alpes. Generally speaking, however, the Rhône-Alpes Region is representative of the French situation as a whole. In order to have an idea of the situation concerning the specific group of suppliers that are of central interest to this study, specific analyses were undertaken in the data base. Firstly, certain categories of companies that are not suppliers of technical vehicle components were eliminated from the data base, see table 5 (final assemblers, raw material suppliers, tooling and machine equipment manufacturers, and suppliers of commodities were eliminated). Moreover, suppliers that had less than 20% of their turnover in the car industry were eliminated. Interview data indicated that below this figure, supplies were not considered to be aligned with the car industry5 . Manufacturers of automobiles and trucks(including Mould manufacturers (tooling) specialized firms assembling different utility vehicles and city cars) Manufacturers of car bodies Manufacturers of packing materials Raw material manufacturers / extractors Manufacturers of textiles First transformers of raw materials Manufacturers of glass Companies specialized in treatment of metals Manufacturers of tyres Production equipment and machine manufacturers Manufacturers of screws and bolts Tool manufacturers Table 5. Activity sectors eliminated from the data base. After the first elimination, 485 companies considered as component suppliers remained in the data base. After the second, 250 aligned component suppliers remained. These 250 remaining companies refer to a reduced population because information on the tier position was given only in 281 of the 485 companies identified previously. However, of the 204 companies where this information was missing, only 28 employed more than 50 people. This indicates that the bias in the identification of the aligned component suppliers is limited even though detailed information was missing in 204 cases. The distribution per activity sector and the distribution of the number of employees in the 250 identified suppliers are exposed in table 6. Activity Manufacturing of rubber articles other than tyres Total Number of Companies 10-50 employees 50-500 employees >500 employees 9 2 7 0 5 Interviews with two former Purchasing Directors in PSA and Renault. 37 Manufacturing of technical plastic parts 48 15 31 2 Stamping, cutting and deep-drawing 17 5 12 0 Screw-cutting 47 25 21 1 General mechanics 56 46 10 0 Manufacturing of springs and small metal articles 4 2 1 1 Manufacturing of hydraulic and pneumatic transmissions and mechanical parts for transmissions 8 2 3 3 Manufacturing of bearings 3 2 0 1 Manufacturing of electrical motors, electrical equipment and cables. 22 4 16 2 Manufacturing of equipment for automobiles 36 15 16 5 Total number of companies 250 118 117 15 Table 6. Component suppliers per product category and employee distribution. In order to have an idea about the number of expert suppliers that are present in the region and develop a list from which potential companies to study could be selected, the following operations were made in the database - leading to an identification of 68 companies (see table 7): • The 14 identified system suppliers were first eliminated6 , • Then, all those suppliers that had indicated that they had no first tier business in the car industry were eliminated, • Third were eliminated the companies that indicated a turnover in the car industry inferior to 50%, • Finally, those that were 100% domestic were eliminated. The expert supplier population on the national level, corresponding to the one identified above, comprised 420 companies (Deranlot, 1993). As the Rhône-Alpes region accounts for 13% of the members of the French Vehicle Equipment Industry Federation, the proportion of 68 companies out of 420, which equals 16%, once again indicates that the region is representative for the French situation as a whole. Turnover Number of Companies Percent of Companies Number of Employees Number of Companies Percent of Companies 6 The system suppliers present in the region were: Valeo, Plastic Omnium, Rockwell, Mannesmann, Bosch, ZF, Sagem, MGI, Coutier, Sylea/Labina, lT&N, Lamberet, Koyo, Burelle and L'Electricfil. 38 < 20 MF 8 12,0% 10-20 5 7,0% 20-50 MF 14 20,5% 20-50 12 18,0% 50-100 MF 13 19,0% 50-100 16 23,5% 100-200 MF 16 23,5% 100-200 12 18,0% 200-500 MF 13 19,0% 200-500 14 20,5% > 500 MF 4 6,0% > 500 9 13,0% Table 7. Turnover and number of employees in specialized supplier firms in the French RhôneAlpes Region (68 companies). The average percentage of first tier supply for the companies identified above was 47%. The maximum value was 100% (6 companies) and the minimum value was 6%. The average percentage of second tier supply for these companies was 26%. The maximum value was 75% and the minimum value was 0% (13 companies)7 . This clearly demonstrates that there is a large group of companies outside the system supplier group that play an important role in the first tier and that a majority of them (81%) are first and second tier suppliers at the same time. Thus, they correspond to the definition of the medium-sized expert suppliers presented in chapter one. The distribution per activity sector and the distribution in number of employees of these companies is illustrated in table 8. It is interesting to observe that 25% of the companies identified as expert suppliers are small sized companies. This modifies the generally held idea (c.f. e.g. Brocquet, 1995) that only very large firms remain in the first tier. The smallest companies were in stamping, screw cutting, general mechanics and automobile equipment. Their turnover in the first tier was generally inferior to 20%, which might indicate that they are on their way out. Eight companies remained in high to very high levels of first tier supply (50-100%). These are highly specialized in high technology and supply specific components mostly for the luxury segment. When the smallest companies are eliminated, for reasons of a too simple organization to accurately illustrate internal problems of integration, it leaves us with a sample of 51 companies from which interview and case study companies were selected. Activity Total Number of Companies 10-50 employees 50-500 employees >500 employees Manufacturing of rubber articles other than tyres 3 0 3 0 Manufacturing of technical plastic parts 5 0 5 0 Stamping, cutting and deep-drawing 6 2 4 0 7 The 13-6=7 companies that had no business in the second tier but less than 100% in the first tier had their remaining business outside the auto industry. 39 Screw-cutting 14 3 10 1 General mechanics 5 2 3 0 Manufacturing of springs and small metal articles 1 0 0 1 Manufacturing of hydraulic and pneumatic transmissions and mechanical parts for transmissions 2 0 1 1 Manufacturing of bearings 1 0 0 1 Manufacturing of electrical motors, electrical equipment and cables. 8 2 5 1 Manufacturing of equipment for automobiles 23 8 11 4 Total 68 17 42 9 Table 8. Specialized suppliers per product category and employee distribution (number of companies). 2.4 SUMMARY The Rhône-Alpes region is the second most important region in the French automotive component industry. The structure of the industry, in terms of company size and turnover is representative of France. The statistics indicate that there is a clear pyramidal structure in this industry. In the sample of 250 aligned component suppliers (table 6), 5% were system suppliers, around 30% were expert suppliers and the rest, i.e. around 65% were small-sized second and third tier suppliers. This indicates that the outsourcing and tier structuring logics have been applied in the French automotive industry. The statistics analysed in this chapter confirm the existence of so-called expert suppliers as defined in chapter one, i.e. medium-sized suppliers with a diversified customer portfolio and an important activity in the first tier. As discussed in chapter one, these represent a specific interest when the industry is restructured. 40 3. A CHANGING PRODUCTION ORGANIZATION After the introduction of the research topic and the presentation of the French car component industry, this section will widen the perspective considerably. First the notion of industrial models and production models is discussed, and a framework for studying the design process is developed. Then, lean production is presented and analysed in relation to traditional mass production. Finally, the human centred model of reflective production is reviewed. The objective of the chapter is threefold. Firstly, it identifies the most important items of lean production described in relevant literature. This is important in order to set the context for lean component development. Secondly, it clarifies different perspectives that can be adopted in studying production organization - in particular the evolution of production models. This is necessary in order to define and justify the position of the present research. Thirdly it identifies a number of general questions related to lean production that inspire some particular orientations for the empirical research. 3.1 INDUSTRIAL DEFINITIONS MODELS AND PRODUCTION MODELS - SOME The objective of this section is to examine some current definitions of both 'industrial models' and 'production models', making it possible, later on, to qualify (or not) lean production as a new model. Important work in this field has been conducted by Boyer & Freyssenet (1995) and the discussion here is based on their work. Hirata (1993) distinguishes three normally accepted definitions of the term model: 1. A model should consist of a set of practices which should also make it 'appear' in the daily management of companies. Presented in this perspective, a model stands for an ideal to which to aspire. The main problem in this first category is that the target model might be represented as a superior practice without a more in-depth reflection. 2. A model corresponds to the 'packaging' or 'stylization' of a number of de facto existing practices. The objective is no longer a theoretical construct or a normative programme , i.e. conceptualisation, but to represent, i.e. summarize and synthesize, a large diversity of practices inevitably through a simplification. In this category, there is a problem of identifying relevant traits and of according them different importance. 3. A model is a vehicle or a method for pinpointing the coherence and the relevance of a theoretical construction. It is aimed at representing the core of behaviours and dynamics observed in the system that is studied. This line of action makes it possible to define all 73 theoretically logical changes that would occur if the model was applied to the firm and the construction of ideal model types. Here there might be a problem of validity as it is difficult to dissociate the theoretical work from the researcher's own values, preferences, and assumptions. Taking into account the weaknesses of each one of these categories, Boyer & Freyssenet (1995) propose a definition by consolidating the explicit or de facto character of the first category, the descriptive one in the second, and the force of abstraction in the third: • The definition consists firstly of recognising a new industrial model only if it "pursues objectives distinct from those of the precedent, implements original technical, organizational, managerial and social apparatuses or at the very least a recombination of the old ones, and finally if it generally delivers results that are in general superior to those of the preceding model" (p. 92). • It consists secondly of rendering coherent/appropriate a certain number of elements (principles, strategies, practices, institutions, apparatuses, rules...) so that there can be a "relative and temporary viability/efficiency/stability under given conditions; that is to say so that there can be an industrial model" (p. 84). These elements are initially defined as illustrated in figure 3. Structure of the firm * degree of centralization * relations with suppliers and competitors Link with the market Organization of production * marketing studies * nature and mode of division of labour * sales network * involvement of clients * type and degree of automation Wage relationship Access to finance * internal/external mobility * link with financial system * system of professional relations * relations with shareholders Figure 3. The components of industrial models. Source: Boyer & Freyssenet (1995). The definition from Boyer & Freyssenet (1995) continues by considering aspects of uncertainties, reductions of uncertainties, conditions, forms of development, periods, and 74 phases. Taking together the above specifications and aspects, they define industrial models as: "the process of making internally coherent and externally appropriate the technical, organizational, managerial and social apparatuses and practices of companies and institutions with the intention of reducing the uncertainties of the market and work, in the forms they take locally and historically" (p. 113). This should lead to a relative predictability of social and economic development rooted in social and productive principles, and that take place under certain conditions of feasibility and viability. The principles should be applicable to, and the conditions fulfilled by, several macro-economic and social configurations. Among other things, this definition underlines that an industrial model must go beyond both a company and a specific geographical setting. The refined and extended components of an industrial model are illustrated in figure 4. Context Market types (products, equipment) Work situation (labour force, unions) Power structure Profit strategies Social and productive principles Organisation Administration Wage relation Apparatuses and practices to design Dynamics to buy to make Contradictions Results to sell Irreversibilities Figure 4. The components of industrial models, extended version. After Boyer & Freyssenet (1995). The discussion of industrial models stays, by definition, relatively abstract and general. The present research aims at describing real developments and outcomes in medium-sized component suppliers under the influence of lean production, emphasising the impact on structure and organization in the product development process. In order to do so, it is necessary to operate with a more detailed framework focused on the production dimension. The items directly pertaining to this dimension are 'to design', 'to buy', and 'to make', in figure 4. As shown in chapter 1, the product development process is not only about designing a 75 product, but also about integrating with suppliers and manufacturing. Taking this part of the industrial model as the point of departure, a production model can be defined. Boyer & Freyssenet (1995) define the three apparatuses or practices pertaining to the production dimension in the following way (p. 114): • 'To design' contains the design organization, product/process integration, and links with customers and suppliers. • 'To buy' contains supplier relations and supplier networks and location. • 'To make' contains production organization (workshop level), technical choices, and work organization. This part of the industrial model is close the approach of Womack, Jones & Roos (1990). In their comparative study of lean and mass production in the car industry, they investigate five elements in car production: 'running the factory', 'designing the product', 'coordinating the supply chain', 'dealing with customers' and 'managing the lean enterprise'. The first three of these elements are directly mirrored in the typology identified above (as are, beyond the focus of this research, the items 'to sell' and 'dealing with customers'). They are therefore retained as a definition of the production organization for the purposes of the present research. Thus, a production model is a configuration consisting of coherent and original principles, strategies, practices, institutions, apparatuses, and rules in product and process design, supply management, and manufacturing that delivers superior results compared to an earlier and different configuration. 3.2 A FRAMEWORK FOR THE DESIGN PROCESS If one crosses the three elements designing the product / to design; coordinating (or managing) the supply chain / to buy; and running the factory / to make (or manufacture), with the four major stages of product development from Clark & Fujimoto (1991) -concept generation, product planning, product engineering, process engineering, and production process-, a framework for studying the part of the production organization that is the product development process can be identified, see figure 5. GLO BAL PROCESSES Designing the Product FUNCTIONS Concept Generation 76 Managing the Supply Chain Manufacturing the Product Product Planning Product Engineering Process Engineering Production Process Figure 5. The product development framework. The vertical elements represent transversal processes in production organization, the horizontal elements represent phases, where specific activities take place. As just stated, the latter correspond to Clark & Fujimoto's (1991) model of product development (p. 27). This model outlines a generic sequence of activities for manufactured and assembled products, and similar models are referred to by the authors. They underline that the development stages correspond to the description of the innovative process in technology management literature: idea generation, problem solving and product planning. In other words, the above framework represents a relevant model underpinning the research that is undertaken in the present project. The framework in figure 5 reflects a final assembler's perspective, which is natural because both Womack, Jones & Roos (1990) and Clark & Fujimoto (1991) principally study assemblers. If the unit of analysis is a supplier, another transversal process comes in to the production organization, namely integrating with customers. The reason why customers come in to the production organization in the supplier case but not in the assembler one is that in business to business relations, the customer defines product specifications, price limits, delivery terms and times, service terms, payment terms and order quantities (Kotler, 1988, p. 212). Hence, the customer directly influences all of the five product development steps in the supplier's production organization. For a supplier, the framework would then look like figure 6. GLOBAL PROCESSES FUNCTIONS Integrating with customers Designing the Product Managing the Supply Chain Manufacturing the Product Concept Generation Product Planning Product Engineering Process Engineering Production Process Figure 6. The product development framework, supplier case. In the present research, the product development framework is used at a primary level to identify changes in product development and supplier integration, first in the literature review, 77 then in the empirical research. The steps in product development (horizontal element in the framework) can be described as follows (Clark and Fujimoto 1991): • In concept generation, designers and product planners define the character of the product from a customer's perspective. The product concept is thus more than a set of dimensions or a list of specifications. • In product planning, the concept is translated into specifics for detailed design, including major specifications, technical choices and cost targets. The central problem is to create a plan that balances competing objectives and requirements. • In product engineering product plans are transformed into real parts and components, integrating product performance targets and constraints. Detailed design of components, stored in blueprints or CAD-systems, are then converted into prototypes. Prototypes are tested and changes are made in an iterative process before official approval. • In process engineering the manufacturing tool that will realize the product is developed. Material flows, plant lay out, work organization and tasks are defined. Design of tools, dies and other equipment is conducted, as well as programming and/or information loading of machines. • In the production process, final products are made and assembled for the end customer. 3.3 COMPARING TWO MODELS OF PRODUCTION ORGANIZATION: MASS VS. LEAN PRODUCTION After the previous definitions and the development of frameworks for studying production organization, lean production will be positioned in relation to traditional mass production and some different perceptions of lean will be discussed in greater depth. Emphasis is placed on clarifying origins, development paths, and terminology. Then, the human-centred model or reflective production is analysed. This concept has emerged as a controversial alternative, first to mass production, then to lean production. An analysis of its contribution is therefore indispensable in a discussion of production models. The chapter identifies more detailed research questions within the frame of the research problem. The gradual emergence of a fundamentally different way of organizing production has been thoroughly treated in literature since the beginning of the 1980s: through operationally oriented manuals (Shingo 1981), in-depth studies of 'the new production paradigm' (Monden 1983), and full range comparative studies of mass versus lean production (Womack et al 1990), this new production model has emerged as 'best practice' and the dominant model for the American and European manufacturing industry. 78 However, several things have to be clarified when talking about a new paradigm in production organization. First there is the question of terminology. Talking about the Japanese production model can be misleading as there is a wide range of manufacturing performance in Japan, North America and Europe - none of these regions can be considered a monolithic entity in productivity and quality performance (Krafcik, 1988). However, and as Krafcik (1988) also recognizes, in can't be denied that the majority of operating principles, management practice, and organizational forms in the new model have been observed and exemplified in Japan, above all at Toyota. A common name for this system is therefore the Toyota Production System or 'Toyotism'. Further, if a founder of it should be nominated it would be T. Ohno of Toyota.1 Using a company name for a production model is nevertheless misleading if, for example, the model is considered to be under development, i.e. permanently integrating different problem solving methods during its diffusion, c.f. the definition of an industrial model from Boyer & Freyssenet (1995). They point out that, over time, each firm goes through significant changes of its strategy, deep crises and/or strong growth, and changes in geographical penetration (market and manufacturing investments in new regions, or withdrawals). Hence, 'Fordism' or 'Toyotism' can only refer to the specific trajectory of respectively these very firms. Strictly speaking, 'Fordism' would exclude for example General Motors, Renault or Fiat (firms that have adopted Ford's principles but that have modified them and added new features to them), and 'Toyotism' would exclude Honda and Nissan (firms which have developed production systems similar to Toyota's). In the beginning of mass production, there was Fordism as there was nothing else of that kind, but the pioneering continuous flow system developed by Ford -steel and rubber came in at the beginning of the factory, and ready-to-drive cars came out at the other end 2 (Krafcik, 1988)has to be completed with a number of managerial principles before it is possible to set up a model of traditional mass production. Womack et al (1990) and Berggren (1992) underline the importance of the development at General Motors during the 1920s under its new president Alfred Sloan for the establishment of a traditional mass production model: • Creation of decentralized divisions managed 'by numbers' from a small corporate headquarters. This was also the managerial technique used for setting up foreign subsidiaries; • Development of a multi-model product range, from cheap to expensive, in order to satisfy a broad market; 1 T. Ohno was a manager of assembly and machine shops from 1945 to 1953. He went on to head the Motomachi factory, the main plant in Toyota City, and the Kamigo factory during the 1960s. He retired from Toyota as Executive Vice President in 1978. (Source: Cusumano 1985, p. 268-269). 2 This system was fully in place at Ford's Highland Park unit by 1914. 79 • Development of stable sources for outside funding. Ford was entirely internally financed during this period (until 1956); • Creation of new professions of financial managers and marketing specialists to complement the engineering professions; • Merger of the need to standardize, to cut costs, and the market demand for diversity; mechanical items were standardized while external appearance altered annually. These ideas were a managerial revolution in the manufacturing industry, and combined with Ford's manufacturing system (technology) and Taylor's scientific management (organization) they founded the traditional mass production system. It is important to emphasize, as does Womack et al (1990), that the horizontal division of labour was total not only at the workshop level but also between the emerging specialist functions that were developed at GM. The mass production system, born in the United States, was then transferred and adapted to different socio-economic and technological environments where it was developed until the end of the 1970s (Berggren, 1992; Camuffo & Volpato, 1993; Freyssenet, 1993; Loubet, 1993). In spite of different problems and different emphasis in different firms, several basic characteristics of the traditional mass production model remained: the narrow span of control in operations, the high work standardization defined by managers, and the need for standardized high-volume production to achieve productivity. When looking back at the beginning of the nineteen seventies, hybrid but quite similarly developed mass production systems were found world-wide in the car industry. A parallel reasoning to the one above can be applied to the Toyota production system and the emergence of lean. When it was first presented in the West at the beginning of the 1980s it was so developed, and traditional mass production so far from pure Fordism that the Toyota model was considered as a new paradigm in production organization. There was nothing else like this model and it was therefore called Toyotism. However, since that date, Toyotism has also been transferred, developed, and adapted - 'hybridized'3 (c.f. Abo, 1994) to "take account of new operating environment" (Lamming, 1993, p. 179). As the model is spread and transformed, the name Toyotism seems, as with the case of Fordism, inappropriate. Krafcik (1988) proposes lean production which is retained here as Womack et al (1990) show that the distinctive characteristics, embedded in the name, are global as far as the conception of a production model, as outlined above, is concerned (i.e. touches global organization, product development, supply management, and manufacturing). These two references, to which one 3 A term used for describing that lean production undergoes transformation during the diffusion process. This phenomenon is most visible in Japanese overseas transplants, which will be further discussed in the analysis of Abo et al below. 80 has to return to trace the original definition of lean production, have developed the concept from studies of the car industry within the IMVP4 . To summarize, for the production model based on scientific management, a strict horizontal and vertical division of labour and functional departments, and a highly specialized production tool adapted to long series, the name traditional mass production will be used. For the production model based on continuous improvement, team working, a flat organization, and a flexible production tool keeping inventory and lead-times short also with a wide variety of products, the notion lean production will be used. How then, is the lean production model presented in literature? Table 9 provides a summary of how relevant references define the characteristics of lean production as opposed to those of traditional mass production. Generally speaking, traditional mass production is representative of the French situation from the mid 1950s until the mid 1970s (Freyssenet, 1993; Loubet, 1993). Thereafter, a period of turbulent changes took place in the American and Western European car industry as a whole. Initially, the pressure for price reductions increased considerably on suppliers. However, this resulted more in squeezed supplier margins than in new approaches to cost savings through improved productivity, and a lot of suppliers left the industry at this moment (Lamming, 1993). He further points out that traditional approaches to cost reductions, on the one hand, and quality improvements on the other, could not balance the trade-off between better quality and lower cost. Therefore, from the mid 1980s some radically new ways of organising production, spanning all the way from the organization of product development to inventory management and delivery practice, have been developed in France as well as in the entire Western automobile industry. The emerging alternative -lean production as conceptualized in table 9- is a result of a combination of extensive improvement efforts and trial-and-error procedures in single companies, intensive benchmarking between competitors, studies and consultant analysis of Japanese practice, learning alliances (Kogut, 1988; Hamel, 1991) between competitors (such as GM-Toyota and Rover-Honda - before the take-over of Rover by BMW in 1994), and finally the dissemination of practice in Japanese transplants in Europe and the United States. THEORY/CONCEPT CHARACTERISTICS Global Strategy MASS PRODUCTION LEAN PRODUCTION − Task-oriented Scientific Management. Product, and shareholder wealth in focus. − Organization-oriented micromanagement. Customer, and organization in focus. − Economy of scale, volume as a key − Economy of scope, time to market success factor. as key success factor. 4 International Motor Vehicle Program of the Massachusetts Institute of Technology, Cambridge, USA. The program started in 1986 and has since then conducted a variety of research projects in the car industry. 81 Design Organization − Numerous hierarchical levels. Mechanistic through highly specialized and strictly separated functions. − Sequenced hand-overs between departments and functions where all responsibility is taken over by the receiver. Supplier Relationships − Subcontracting to suppliers. − Arm's-lengths price bargaining, short contracts. Production Arrangement Production Tool Inventory Operations Quality Control Labour Performance Measurements and Career Paths LEADING FIRM − Highly specialized production tool. − Considerable inventory at all stages. − High work standardisation - by managers. − Narrow span of worker control. − Quality control separated from the manufacturing process. − Labour as a variable cost. One worker equals one task. − Flat organization. Organic through inter-functional cooperation promoting design for manufacturing. − Continual progress of project through specialist functions under the supervision of a project management team. Project manager responsible for project from concept generation to market introduction. − Outsourcing to suppliers. − Partnership approach to long-term supplier relations. Increased strategic importance of suppliers. − Flexible production tool. − Zero stock. Just-in-Time pull system. − High work standardisation - by teams. − Moderate span of worker control. − Continuous quality control. − Piece work at shop floor. − Hierarchical advancement within function. − Labour as a fix cost. Each worker trained for a variety of tasks. − Continuous improvement (Kaizen) at all levels, propositions, quality circles. − Advancement related to interfunctional skill acquisition. − Ford from the nineteen tens. − Toyota from the nineteen fifties. Table 9. Mass vs. Lean Production: Characteristics. Summary of characteristics from Berggren (1990, 1992), Bouchut (1990), Chanaron & Lung (1995), Clark & Fujimoto (1991), Kawamura (1994), Krafcik (1988), Monden (1983), Womack et al (1990). Table 9, generated from the literature survey, proposes items used to develop questions about the application of lean techniques in the empirical research. When elaborating it, attention has been given to two particular problems identified by Sako (1992, p. 12): • Firstly, lean and mass production have been identified each in their own right. If one extreme of a dimension is defined only as residual of the other, only one dimension will be operationalized. • Secondly, particular care is taken so that both mass and lean production are equally made up of empirically observable elements making it possible to verify or refute each concept empirically. The table provides a benchmark for lean production. Comparing it to the proposed definition of a production model (chapter 3.1), at least the criteria of pursuing distinct objectives, 82 implementing original principles, and delivering superior results are fulfilled. The criteria of coherency between the different principles in lean production has also been highlighted (Womack et al, 1990; Lamming, 1993), but whether this coherency will be stable over time remains to be seen. Thus, lean production is close to qualifying as a new production model. However, in order to gain an in-depth understanding of lean it is necessary to analyse a selected part of the literature in more detail. Four references will be analysed in the following sections. Krafcik (1988) was the first author to use the name lean production and to give it an initial definition. Womack, Jones & Roos (1990) developed the global concept of lean production. Lamming (1993) focused on supplier relationships where he analysed the changes in the automobile components industry and developed the concept of lean further. Abo et al (1994), finally, addressed the important issue of transfer of the Japanese inspired concept to the West. The perception of lean production in this literature is presented in exhibit 1. Before taking a closer look at these perceptions, some general comments on research perspectives seem necessary. All these references represent a strategic level of analysis (i.e. more focused on propositions and descriptions of concepts and ideas than with questions and answers concerning their applications) and use mainly quantitative indicators for measuring performance (Womack et al, 1990; Krafcik, 1988) or the degree of application of Japanese style management (Abo et al, 1994). From a managerial perspective, their contribution is essentially to make managers aware of new practice in production organization and to set up check lists for strategic objectives such as reducing development lead time, implementing JIT, increasing team working, reducing the number of suppliers, developing design capacity if you are a supplier, and so on. Little is said about how to develop and integrate lean practice at firm level, or the possible problems and effects such an implantation might have on an operational level. Several case studies (Camuffo & Volpato, 1993, Freyssenet, 1993, Loubet, 1993, Gianfaldoni, 1994), and business press articles focused on specific companies5 show that a general shift in the way of organising production, towards the lean model, is taking place. But they also show that the new models that emerge in this sense are integrating existing practice with managerial techniques developed in Japan, where these later act as eye-openers. Thus, there is a clear need to develop understanding of lean implementation also on an operational level; the objective of the present research. 5 For example L'Expansion no 492, 1995 on Chrysler, Capital, January 1995 on Citroën, or Usine Nouvelle, no 2439, 1994 on the PSA-Fiat minivan. 83 Referenc e Krafcik (1988) Perception of Lean Production Up until the early 1980s there were clear differences between the production system of Toyota and most Western producers. Recently these differences have begun to diminish, as many Western producers have returned to partially Western roots by adapting Toyota's interpretation of pure Fordism. Differences and similarities are analysed through six production system characteristics: work standardisation, span of control, inventories, buffers, repair areas, and teamwork. In summary, span of worker control is different in lean while work standardisation and inventory level are similar but differently managed. The term lean is introduced in opposition to the term buffered. A buffered production system has high inventory levels buffering against unexpected quality problems; assembly lines have built-in buffers to keep production moving if equipment breaks down; legions of utility workers are kept on the payroll to buffer unexpected periods of high absenteeism; repair areas are huge to buffer against poor assembly line quality, and so on. Lean plants, stated as best exemplified by Toyota, show the opposite conditions to these: inventory levels are kept at an absolute minimum so that quality problems can be quickly detected and solved; bufferless assembly lines assure continuous flow production; utility workers are no longer necessary; repair areas are tiny as a result of the belief that quality should be achieved within the process. Lean operations can be considered high-risk/high-return ventures. Much of the risk can be neutralized given an experienced, well-trained workforce, responsive suppliers, and good product designs. Womack, Jones and Roos (1990) The use of the name lean production refers back to Krafcik. The research, focused on production organization in the world automobile industry, presents lean production, original from Toyota, through five elements: running the factory, designing the car, coordinating the supply chain, dealing with customers, and managing the lean enterprise. To understand lean production, each step in the process, not only final assembly, must be regarded. The study takes a global perspective. Lean production exists: lean design, lean supply, lean manufacturing, and lean sales management are regarded as fully developed at Toyota and Honda, and adopted by some Western manufacturers (c.f. Ford, mainly due to learning from the alliance with Mazda). Each element is described through a comparison with classical mass production and lean production. Lean production is argued to be superior to traditional mass production in all elements. This is supported by performance indicators such as product development lead time, assembly hours, assembly defects, and inventory. The world auto industry needs to adopt the lean model. Its diffusion is discussed, but not its transferability. Introduction of the notion 'lean enterprise' which is the mechanism of coordination needed to bring all elements into harmony also on a global scale. Exhibit 1. Perceptions of Lean Production. Summary of the central messages in Abo et al (1994), Krafcik (1988), Lamming (1993), Womack, Jones and Roos (1990). Referenc e Perception of Lean Production 84 Abo et al (1994) The research is focused on the transfer of the Japanese production system to the United States. Five groups of elements for Japanese Style Management are discussed: work organization and administration, production control, parts procurement, group consciousness, and labour relations. Japanese Style Management represents specific characteristics as opposed to the ones of traditional mass production (exemplified with American practice) in each of the five groups. The first three groups represent the core constituent elements, while the two remaining groups constitute a supplementary framework for the effective application of the core elements. The notion lean is not frequently used. Only in the foreword do we learn that "The lean production model, now taken as the major paradigm in international competitiveness debates, emerged in a very particular Japanese context" (p. xvi). Lamming (1993) The research is focused on supply management in the car industry. The Japanese model for assembler-supplier relations is described as developed in a specific context: in Japan between 1940 and 1980. It would be insufficient on a global level. Best practice firms would need to modify and extend this model, to take account of new operating environment. Lean supply is a strategic model for assembler-supplier relationships: a target for continuous improvement programs. It is practicable - the evolution in the industry all point to it. In most fields, however, there is some way to go before it becomes a reality. In Japanese practice (that globally serves as an example also for lean supply), the assemblers and their suppliers have developed as senior and junior partners. Collaboration in these circumstances lacks the richness of an equal's contribution to be truly lean. Exhibit 1, continued. Perceptions of Lean Production. Summary of the central messages in Abo et al (1994), Krafcik (1988), Lamming (1993), Womack, Jones and Roos (1990). 3.3.1 At the Manufacturing Level: Back to Basics Krafcik (1988) presents lean production as a production policy developed at Toyota, on the basis of Ford's original principles, not omnipresent in Japan, but on the other hand also present in some U.S. and European plants (especially Ford). This representation of lean has to be seen in the light of the unit of analysis - the assembly plant, and the chosen criteria for comparison. In fact, work standardisation -but tasks defined by teams-, continuous flow production with small buffers -but managed through just-in-time-, and small repair areas -but due to quality assurance rather than experience from repeating-, are all examples of "original Fordism with a Japanese flavour" (p. 42). However, there is no discussion of supply management or product development organization either separately or in combination. The importance of early involvement of suppliers in product design, joint component development, supplier engineered parts, collaboration in the context of Japanese Keiretzu6 , or stage overlapping and simultaneous engineering in the product development process are not analysed. It is only mentioned that these features would minimize the risks with lean manufacturing. This might be explained by the fact that Krafcik's (1988) research is one of the first reports that goes beyond a description of the Japanese 6 Supplier groupings with cross equity between the participating companies (suppliers and assemblers). 85 manufacturing system, by trying to conceptualize a comparison between new production management, developed at Toyota, and traditional 'buffered' manufacturing; the perspective must be reduced, and the manufacturing level, at the time the most obvious and documented part, was chosen. Krafcik's (1988) presentation of lean production is, thus, limited in its scope; to manufacturing at workshop level. It sees the presence of this lean manufacturing more as depending on corporate culture and production management policies (company-specific) than on geographical location. Whether the lean model is likely to undergo further development, as Western-based producers are transforming their management policies toward it, is not clearly stated. Supply management and design for manufacturing are seen as support systems to lean manufacturing and not as an integral part of a more global model. 3.3.2 A Complete Model, but Confusion about Transfer After Krafcik's (1988) presentation of lean production, Womack, Jones & Roos (1990) provide a conceptualisation of the complete lean production system - from generating the concept of a new product to distribution and sales. This recognition of lean production as a complete system, integrating all aspects of production organization in the sense defined in chapter 3.2, is an important result of their research, but causes at the same time a problem. The authors argue that lean production should be adopted everywhere and as soon as possible, but if lean production is a system where one part can't function without the other, an important question in relation to the transfer of the model must be how to develop all of its components in another context. However, Womack et al (1990) do not explicitly discuss transferability or 'hybridisation' of the model. This problem is related to the fact that little emphasis is placed on diversity within production models. In fact, there is diversity within traditional mass production, for example concerning the degree of automation, or the degree of supplier engineering (De Banville & Chanaron, 1991). Moreover, the human centred experiments in Germany and Sweden have played an important role in the redefinition of assembly work (Berggren, 1992). This has also influenced the work organization in the French and Italian car industry. In Italy, for example, after strikes and fairly violent conflicts, flexible job assignments, limits to work saturation, and social control over production cycles were negotiated at the beginning of the nineteen seventies (Camuffo & Volpato, 1993). However, the main measures taken in order to reduce dependency on workers and union influence were heavy investments in labour-saving manufacturing technologies, leading to the diversity in degree of automation already identified. There is also diversity within existing lean practice. For example Toyota has a large product range from 86 small city cars, via family cars to luxury, and sports models, not forgetting utility vehicles and 4WD, whilst Honda focuses on the upper niches of the medium segment and also on luxury and sports models. Moreover, Toyota has focused on production and operations management whilst Honda emphasizes engineering, above all engine development (Mair, 1994; Shimizu, 1993). Finally, and maybe above all, diversity is still to be explored in a 'truly lean' model, to borrow the term from Lamming (1993), that might emerge when lean principles are refined and integrated with other solutions. Taking all this into account, Womack et al (1990) seem to overestimate the generality and the novelty of the model they present (Boyer & Freyssenet, 1995). Many items and results from the research of Womack, Jones & Roos (1990) have already been analysed, c.f. chapter 3.2 and table 3.1. Here, some additional aspects will be addressed: the financial support system keiretsu; the notion of the lean enterprise and the ideas of how to manage inter-firm value streams; and, related to this, the specific Japanese career management. 3.3.2.1 The Keiretsu System Beginning with the keiretsu system, this seems to be an important and controversial factor for the lean production model and its diffusion. First the description from Womack et al (1990): "Each keiretsu consists of perhaps twenty major companies, one in each industrial sector. Unlike the zaibatsu, there is no holding company at the top of the organization. Nor are the companies legally united. Rather, they are held together by cross-locking equity structures each company owns a proportion of every other company's equity in a circular pattern - and a sense of reciprocal obligation" (p. 194)7 . Among the key companies in each group are a bank, an insurance company, and a trading company. The keiretsus developed for two main reasons: a mistrust of the arm's-length stock-market (Japanese industrials could not imagine a system where there was no reciprocal obligation); and concern about foreign take-overs as the economy took off and many companies became profitable after World War II (Womack et al, 1990). The Japanese auto manufacturers, Toyota, Mazda, Mitsubishi, and Nissan, are all affiliated to a keiretsu. Also certain suppliers belong to the keiretsu. Primarily, however, so-called cooperative suppliers (Gong, 1993) to Japanese carmakers belong to a grouping organized by the carmaker in order to coordinate the supply of the most important systems and components, so-called kyoryoku-kai (De 7 A zaibatsu is a family-owned holding company that controlled industrial empires consisting of one large company in each of the sectors steel, shipbuilding, construction, insurance, and finance. This system characterised the first era of Japanese industrialization from 1870. 87 Banville & Chanaron, 1990; Hines, 1995)8 . In these groups, general meetings and sub-group meetings are held regularly with the objective of diffusing technical, commercial and managerial information in the specific production system corresponding to each group (Lecler, 1993). Toyota's kyoryoku-kai consists of around 170 suppliers, about half of them have equity relationships with Toyota, thus belong to the keiretsu (Gong, 1993). The other half are either family owned companies that commercially depend heavily on Toyota, or more independent suppliers also belonging to other assemblers' kyoryoku-kai. Obviously, for certain suppliers, membership of a kyoryoku-kai is not exclusive. This complicates even more the interpretation of Japanese buyer-supplier relationships; according to Gong (1993) certain cooperative suppliers are neither tied to the assembler financially, nor do they exclusively cooperate with one assembler. A hypothetical conclusion from this literature analysis might be that the keiretsu system is and has certainly been of the utmost importance in the development of collaborative relations between buyers and suppliers in Japan, and one must therefore be aware of its mechanisms. However, some suppliers seem to manage close relations without the financial incentive and, moreover, they do so with more than one supplier. In order to proceed in the analysis, a closer look at the keiretsu system seems important. Womack et al (1990) find the keiretsu system the most dynamic and efficient system of industrial finance yet devised. Let's now take a critical look at this 'mechanism of the lean enterprise' in the perspective diffusion of lean production and the consequences for medium-sized suppliers. Four points are particularly important: 1. Most scholars seem to agree that the keiretsu system is unique to Japan, and impossible to transfer to the West - at least in the foreseeable future (Womack et al 1990, Lamming, 1993, Lecler, 1993). 2. It must be more natural, and hence easier, to develop advanced support systems for collaboration such as open books in cost accounting, common accounting and quality assurance systems, or integrated design procedures, when belonging to the same grouping in the way that is the case with the keiretsus. 3. An important risk reducing factor is built into the keiretsus (besides financial risk reduction); there is little risk that a specific quality approach or cost analysis method is spread to competitors outside each keiretsu. As the keiretsu system is practised, it minimizes the risk, for suppliers, of losing business to competitors when sharing critical information such as cost break-downs, new technological solutions, etc. with their partnership customer. 8 The term kyoryoku-kai was also used by Sako during a conference at the GERPISA Second International Colloquium, June, 1994. Other researchers use the terms industrial keiretsus (Lecler, 1993) or cooperative supplier groupings (Gong 1993) for describing the same phenomenon. 88 4. There is another side to the coin of competitive risk reduction, namely the quasi obligation for the suppliers to collaborate through the dominating position that the manufacturers keep in the system (Lamming, 1993). If the sense of reciprocal obligation falters, the more pragmatic factor of 'hostage equity', held by the keiretsu, can be used to, for example, prevent a supplier from changing product or market strategy, or more radically to replace a too autonomous company management. These points show that the keiretsu system is a decisive factor behind the efficient collaboration between Japanese assemblers and suppliers in several ways. But if it is impossible to transfer, how will these close relationships, proven to be so important for reducing development lead time, component costs, and for improving quality, develop in the West? In this situation, without the financial incentive, the partnership approach has to be justified, as Lamming (1993) argues, only by the advantages of the collaboration itself (i.e. shorter development lead-time, cost reductions, innovativeness, and quality improvement). The fact that the supplier groups exist in parallel with the keiretsus in Japan also means that this structural support for collaboration is present. The problem with not emphasizing the quasi obligation for suppliers to collaborate due to the financial ties in the keiretsu system or the industrial ones in the kyoryoku-kai, is that formal procedures and organizational structures that, in the West, must replace these systems to support the new ways of working, as well as the learning effort that increased collaboration demands, might be underestimated. When financial ties exist, the relationship actually starts with a formal and common structural mechanism upon which the rest of the necessary principles and methods can be built. To attain the advantages of partnership takes time and effort, because without this basic component, even greater attention must be paid to development and organization of common structures and collaboration. These structures will govern the entire relationship between a supplier and its customers, integrating, of course, the problem solving cycles in product development. From the above one concludes that organizational links, coordination processes, and communication structures seem to be the breeding ground for closer buyer-supplier relationships. This gives rise to several questions for the field research: • What links exist between a supplier company and its customers and own suppliers? • How are different tasks coordinated with focus on the product development process? • What communication structures, both formal and informal, exist? • What are the support systems, tools and methods that are developed in companies, and how are they managed? 89 3.3.2.2 The Lean Enterprise In a later article, Womack & Jones (1994) develop the notion of the lean enterprise. This time their analysis emphasizes the organizational aspects more than the financial ones in the synchronized management, hence responding somewhat to the criticism presented above. If the reference from 1990 provided an explanation of the keiretsu system and facilitated an analysis of its importance for the lean production model (if it is important for supply management it is also important for product development performance), the reference from 1994 puts less emphasis on this aspect. In fact it recognizes, between the lines, that the keiretsu won't be the best solution when transferring the lean model to the West: "The lean enterprise is also very different from the vertical keiretsu of Japan, whose members cement their relationships by taking equity stakes in each other. Unlike keiretsu members, participants in the lean enterprise must be free to leave if collaborators fail to improve their performance or refuse to reveal their situation" (p 103). The basic elements in the lean enterprise listed by Womack & Jones (1994) are: a re-focus on a limited number of activities within each firm; strong collaborative ties with clear agreements on target costing, levels of process performance, the rate of continuous improvement and cost reduction; consistent accounting systems; and formulae for splitting pain and gain. In addition, a particular career management within flat organizations focused on skill development more than on hierarchical advancement, and rotation of mid and senior managers between the company's operations, suppliers and foreign operations are central elements of this new industrial structure (Womack & Jones, 1994). In the lean enterprise, highly specialized companies, using state-of-the-art technology in their special core activity, form value streams where each participant adds a piece of value throughout the production chain - from raw material to distribution and sales of a finished product, see figure 79 . Once common principles have been agreed upon within the value stream, the companies must practise mutual verification with audits conducted jointly and in both directions supplier - customer - supplier. 9 The idea of the value stream is of course based on Porter's (1985) value chain concept. However, while Porter emphasizes the optimization of discrete stages in the value chain, the value stream gives renaissance to the importance of the links between different activities within one firm and between firms. Thus it focuses directly on the changing roles of all actors in a production chain. 90 Inter-firm Value Stream Intra-firm Value Streams Raw-material/Sub-contractors Components/Sub-assemblies Assembly The Specialist Functions. They become 'schools', systematically summarising current knowledge, searching for new knowledge, and teaching it to their members. Moreover, functions should develop guidelines for best practice in their specialized field, define long term partners, and rules for governing common problem solving processes with their counterparts up and down the value stream. The process-management function. A new function which defines the rules for managing cross-functional teams and the continuous-flow production, including quality assurance. It teaches team leaders in product development and production how to apply these rules, and it constantly searches for better approaches. Figure 7. The Lean Enterprise. Illustration of the concept developed by Womack & Jones (1994). In relation to the value stream concept, two things are of special interest for this research: • Firstly, the customer-supplier relation becomes crucial as highly vertically integrated companies tend to disappear in favour of specialized assemblers, suppliers and subcontractors10 forming value streams. This calls for studying the processes that represent the interface between customers and suppliers; • Secondly, the organizational structures and systems, the work processes, and the human qualifications that must develop to support this kind of enterprise are new and relatively unexplored. This calls for in-depth case studies that explore what organizational processes are being developed, that describe how these structures are organized and work, and that assess their performance in relation to initial targets. The lean enterprise is an extended part of the lean production model and it is presented as a new model for industrial organization applicable once single firms have applied lean techniques to their specific activities. It is therefore important to have the lean enterprise in mind, as a component of the theory of lean, when observing it in the field. In fact, the lean enterprise is in some way an answer to the question about links between suppliers and customers in a production chain. There is, however, an important need to discover empirically in more detail 10 It is interesting to reflect upon what might happen with the much debated system supply in this context. Will a system supplier focus more and more on assembly as the core activity, and become a sub-assembler to the car manufacturer while ceasing an increasing part of its design activity? 91 what is done in order to build these coordination links in and around the specific group of suppliers that is studied. 3.3.3 A Hybrid Perception of the Evolution of Lean Abo et al (1994) take up one part of the thread left by Womack et al (1990), i.e. transfer of Japanese Style Management to the West. Why only one part of the thread? Because they study how Japanese automotive and electronic firms with manufacturing operations in the United States apply and adapt Japanese management, and not how Western companies integrate lean principles. Thus, the focus is on social and cultural factors: can the human factors that underlie the Japanese production system survive the transfer to a foreign environment; and will the radically different social and cultural environment in the U.S. make such a transfer very difficult? (p. V). Their basic assumptions are that Japanese assembly industries show superior performance due to an emphasis on human factors on the shop floor and certain characteristics of administrative operations, and that in order to apply their specific methods, an adaptation to various local environmental conditions must be undertaken. What is particularly interesting with the work of Abo et al is that they provide a very detailed presentation of Japanese style management including an in-depth description of its different items (see table 10). I Work Organization and 8 Quality control V Labour Relations Administration 9 Maintenance 17 Hiring policy 1 Job classification 10 Operations management 18 Job security 2 Wage system III Procurement 19 Labour unions 3 Job rotation 11 Local content 20 Grievance procedures 4 Education and training 12 Suppliers VI Parent-Subsidiary Relations 5 Promotion 13 Procurement method 21 Ratio of Japanese expatriates 6 Firstline supervisors (or team leaders) IV Group Consciousness 22 Delegation of authority II Production control 14 Small group activities 23 Managerial position of Americans 7 Equipment 15 Information sharing VII Community Relations 16 Sense of unity 24 Donations and volunteer activity Table 10. Application-adaptation evaluation form, source: Kawamura (1994, p. 27). From the framework in table 10 one can see that, as with Lamming (1993) and Krafcik (1988), lean production is analysed in components and not as a global production model. As stated above, the focus is on human factors and on the production process. Supplier relationships are treated only briefly, and product development organization is hardly mentioned at all. However, the framework in table 10 provides several interesting and 92 complementary inputs to table 9 and was used as partial input in the present research for constructing interview questions concerning the use and perception of lean techniques (see chapter five for further details on the interview topic guide). The work of Abo et al (1994) consists of using the above 23-item 6-group (group seven was omitted by the authors) framework for evaluating, on a five point scale, the degree of application - adaptation of Japanese style management in 41 U.S. transplant plants. The score five was awarded to items that revealed the maximum degree of application and the score one to items considered close to the American system. Although the research methodology can be criticized (application of quantitative data presentation techniques, whereas most of the collected data were of a qualitative nature) several interesting results were obtained. Generally speaking, it was found that Japanese transplants operate with Japanese and American elements together. A typology of ten different types of Japanese transplants reveal, however, some significant differences11 : • Japanese-led firms show the highest degree of application in human resource management - which is the area considered by the researchers as core in the Japanese model. However, the structural stability of such companies can be questioned due to the high dependency of Japanese expatriate managers; will the promising results persist when management responsibilities are gradually transferred to local staff? • American-led firms fail in implementing human resource principles, but are quite successful on the 'material' side, i.e. in the degree of local supply and in cost cutting. Some firms try a 'revisionist' application, i.e. they positively accept American methods while trying to adopt the logic of the parent company's methods. If such an effort can be proven to be as efficient (in productivity, quality and cost) as the Japanese solution for a specific item, the authors qualify this option as the most ideal form of transfer. Abo et al (1994) also discuss productivity in order to assess the relationship performance application-adaptation. The performance measurement used is the volume produced per worker per hour (this again marks the manufacturing perspective). The results, although influenced by some bias12 show that the 'parent' - transplant difference was about 20-30% in favour of the former. Due to the fact that transplants generally manufacture fewer models, employ more quality control personnel, have an important quantity of Japanese expatriates, and do more off-the-job training and more supplier development, the advantage held by • 11 The analysis here is limited to results concerning the auto industry. The discussion of the electronics industry referring to maquiladoras was of little relevance to the present project. 12 Mainly related to difficulties in controlling for differences in production models and production processes that were not necessarily the same either between transplants, or between transplants and parent companies. 93 'parent' companies might be even larger. Independent of the application - adaptation rate the volume per worker per hour rate was better in 'parent' companies. Even though important lessons can be drawn from the work of Abo et al (1994), the problem is very different from the one Western manufacturers are confronted with when trying to implement new practices into an existing structure. In the case of transplants, where owners, management, and even the site are different, there are fewer problems of resistance to incremental change; everything is different instantly and people expect it to be so. In contrast to what can be expected from the title of their research 'Hybrid Factory', the research perspective remains static with respect to the evolution of production models. This is because the Japanese model is considered as a fixed best practice. For the authors the adaptation is not seen as a possibility to generally refine the lean model in the 'truly' lean sense developed by Lamming (1993) that will be further discussed in the following section. Hybridisation is not presented as a process that contributes to reinforcing or developing lean production as a production model, but only a process that integrates different cultural elements. To the question whether they consider that the adaptation will be useful also in Europe or ultimately in Japan, the answer must be no. Therefore their perspective is dynamic only under the condition of a changing geographical location and the research perspective itself is hybrid with respect to a dynamic or static perspective on the evolution of production models. 3.3.4 Beyond Diffusion - Anticipating a 'Truly Lean' Model Lamming (1993) is one of the most complete references on supplier relationships in lean production. In this chapter, the analysis is strictly limited to what is said about production models. Chapter four will come back to the works of Lamming when reviewing the current conceptualisation of buyer-supplier relationships. Going beyond a description of Japanese buyer-supplier relationships and beyond a comparison between different phases in supply management, Lamming (1993) proposes a theory development after having criticized the actual conceptualisation of lean. By re-focusing on a particular component of the production system, supply management, he develops a more nuanced picture of how to study the emergence of a new model. Such an approach can be grounded in field observations, as in Lamming's case, or more intuitively formulated as an assumption as is the case in the present research. 94 He uses the name lean to describe a context dependent best practice: lean supply as a strategic model for buyer-supplier relationships. In opposition to Womack et al (1990) and Abo et al (1994), Lamming (1993) emphasizes the differences between the Japanese model of buyer-supplier relationships and the practice that is currently under development - 'lean supply'. This lean supply has a potential to become a new best practice by overcoming problems in the Japanese practice such as the weak position of suppliers in relation to the assemblers. This weak position appears in several ways. For example, rather than simply following a senior partner (the carmaker), the lean supplier has to decide for itself how to balance a portfolio of customers and relationships in order to exploit its resources and assets in a complementary manner in several directions at the same time. Another example analysed by Lamming (1993) where lean supply goes beyond a Japanese model is in product innovation. In lean supply the supplier must not only contribute to product technology via the medium of collaborative effort with the assembler, but also develop technologies independently of the assemblers' requirements. Hence, Lamming (1993) proposes a theoretical best practice model which has not yet been put into practice. He stands out from the above references, not by the level of study which is strategic, but by the more dynamic perspective on the diffusion of lean. His perspective is the testing and the further development of the new concept. This perspective, where the basic elements of lean (c.f. table 3.1) are a priori present in practice, but where each of them integrate with existing solutions which gives room for new concepts to emerge is close to the standpoint developed in the present research. As explained in chapter one, a basic assumption is that an establishment of lean production as a new paradigm in production organization can be done only with such a dynamic perspective. To be a real alternative, the lean model has to refer to modification and extension of practice developed in the specific context13 of Japan between 1940 and 1980, taking account of "a new operating environment" (Lamming 1993, p. 179). 3.3.5 Classifying the Perceptions of Lean After the above analysis it might be interesting to attempt a classification of the perceptions of lean production. Firstly, the different perspectives reflect the different definitions of an industrial model from Hirata (1993) discussed in chapter 3.1. Womack, Jones & Roos (1990) are close to the first group; lean production is an ideal to which to aspire. Abo et al (1994) develop and use their application-adaptation framework to summarize a large diversity of practices through simplification; this mirrors the second category. The third category, finally, is represented by 13 Briefly, this specific context consisted of the Japanese government's trade protection and export orientation that created a steady and stable increase in demand. Governmental efforts were also made with the objective of nourishing an increasingly talented pyramid of suppliers Abo et al (1994). 95 Lamming's (1993) work; he discusses the logical implications of a theoretically ideal model type. There are also several other factors that distinguish the above references. Two of the most obvious are: • The scope that is covered, i.e. is lean production viewed as a global system (design and supply management and manufacturing) or analysed in components (design or supply management or manufacturing) • The perspective on development of lean production; is there a dynamic or a static perspective taken on the evolution of the model? A classification along these two dimensions is proposed in figure 8. Lean as an existing (Japanese) model. Lean as a global system Lean as context dependent best practice under development Womack et al (1990) Lamming (1994) Lean examined in components Abo et al (1994) Krafcik (1988) Figure 8. Perspectives on Lean Production. Comparing with the definition of a production model from chapter 3.1, it is obvious that only Womack et al (1990) attempt to propose a complete model - a production model in their case. The problem with taking a global perspective, and above all with attempting to develop an industrial model, as do Boyer & Freyssenet (1995), is that different components must be reduced into analytically manageable units, and that the relationships and interconnections between items become extremely complex. These problems have been illustrated in the critical discussion in this chapter both of the work of Boyer & Freyssenet (1995), leading to the specification of the production organization framework, and the one of Womack et al (1990) leading to the formulation of research questions. Conversely, there are also obvious problems with analysing new practices in components only. The problem with the references placed in the lower half in figure 8, and as will be argued also with the criticism of lean production coming from researchers in the human-centred paradigm, 96 is that they talk about 'new model' or 'not new model' only from the perspective of their core problem (design or supply management or manufacturing). If the interconnections to other steps and stages in the production process are ignored, it is difficult to see if the new practice pertains to a new production or industrial model. The ambition in the present research project is to recognize lean production as a global production model, i.e. covering design, supply management and manufacturing. Furthermore, the literature analysis has led to the assumption that lean production is not a ready made recipe, but integrates and experiments with a diversity of practices within the frame of coherent and differentiating principles. To summarize, the research perspective in the present project is situated in the upper right-hand square in figure 8. This means, firstly, that the product development process in medium-sized expert supplier firms will be studied in a transversal perspective according to the supplier product development framework in figure 5. Secondly, it means that the research, after a benchmark of the most significant hitherto identified items of lean production, is looking for the conceptualisation of current and actual practice beyond existing models and theories. 3.4 THE CONTRIBUTION FROM REFLECTIVE PRODUCTION Besides mass and lean production, a third concept - the human-centred model or reflective production - has gained increasing attention in the international debate, above all due to the active experimentation in the Swedish auto industry (Ellegård, 1989; Berggren 1990; Ellegård et al 1992). The controversial closure of the Volvo Uddevalla 14 and Kalmar plants has started a new debate about the significance of these experiments for the development of new industrial models (Berggren, 1994:2; Charron & Freyssenet, 1994; Freyssenet, 1994; Sandberg, 1994) Reflective production is also discussed as an alternative to lean production (Berggren, 1992 and 1994:1; Adler & Cole, 1993 and 1994, Ellegård et al, 1994). Reflective production has its roots in socio-technical theory, and emerged at the end of the 1960s as an alternative to the Taylorist work organization and as an attempt to overcome problems with employee turnover, absenteeism, recruitment, strikes and other social conflicts (Berggren 1990). In Germany, the industrial sociologists Kern & Schumann have made 14 In January 1995 Volvo and the British firm TWR (Tom Wilkinshaw Racing) decided to open the Uddevalla plant again to produce a coupé and a cabriolet Volvo 850. 20 000 units will be produced per year according to reflective production principles. A subsidary with Volvo (49%) and TWR (51%) has been created. Reflective production principles will be extended to a body and a painting shop completing the site. Source: GERPISA 1994:1. In January 1997, two models, a coupé and a convertible baptised C70, were presented to the public. In 1997, 7 000 to 8 000 cars of the coupé model are scheduled for production. For 1998 the goal is to produce 20 000 units of the two models. Source: Volvo Press Release reported in Ny Teknik , 10th January, 1997. 97 important contributions to the conceptualisation of a human-centred production model. For example, Kern & Schumann (1986) argue that the beginning of the 1980s saw a different way of rationalising production. In the new production concepts, human work is not seen as an obstacle to productivity or as a disturbing factor; it is seen as a productivity factor to make use of: "in more holistic work patterns lie important possibilities and not dangers" (p. 19). In both Sweden and Germany, socio-technical efforts went further than in most other countries. Labour market conditions (high work force qualification and full employment) and strong union influence together with consensus industrial relations were the main reasons for this. In Sweden, a particular continuity can be observed, especially within Volvo, where the socio-technical interest persisted in the truck division during the 70s, and came back in the car division in the mid 80s (Berggren, 1990). These experiences were combined for the construction of the Uddevalla plant in 1985, a plant that should combine "quality, efficiency, and a concentration on the human being"15 , and that became the leading example of reflective production. Now, what are the characteristics of reflective production; how is this model related to lean and mass production, and how can it contribute to this research? The following paragraphs will try to answer these questions. Reflective production focuses on work organization at the assembly level. The model proposes task enrichment and production-group organization for car assembly. The assembly line is abolished, and teams of around ten multi-skilled workers assemble the car during a work cycle of two hours on average (Adler & Cole, 1993). The teams are supported by a partially automated materials handling process with parallel material flows, that for each car deliver an individual kit of parts and components to each assembly shop enabling the autonomous teams to assemble objects independently (Berggren, 1994:2; Ellegård et al, 1994). Work tasks assigned to teams and individuals comprize a large number of work operations and this means longer work cycles that require but also generate extended competence (Ellegård et al, 1994). That the reflective production model is much more than just task enlargement becomes obvious when considering the important technological innovations for material handling that were developed parallel to the organizational innovations. The name reflective production refers to the learning principles that constitute the basis of the model: so called functional learning in opposition to traditional additive learning (Ellegård et al, 1994). With functional learning, the objective is no longer to memorize a more or less important number of distinct operations, but to learn a complete process of assembly. The memory will be supported by the logic of the product, and the assembly of an engine or a 15 Volvo internal project document 1985, quoted by Ellegård (1989, p. 13). 98 whole car becomes a logical process where each step has its natural and functional place (Freyssenet, 1994). In car assembly for example, this means that instead of codifying components and operations with numbers adapted to machines, adequate and expressive names, which support the human understanding of each component, operation and sequence in the assembly process as a whole are used. Experiments at Volvo Uddevalla have shown that it was possible for one skilled worker to assemble a complete car alone. Compared with the extremely weak possibilities that workers on a traditional assembly line have in understanding and surveying the logic of the product and its assembly processes as a whole, this reveals the important potential for work enrichment that functional learning implies. 3.4.1 Confusion about Productivity The Volvo experience indicates that abolishing the line regime in favour of autonomous work groups improves work content and provides new opportunities to radically improve ergonomics in the assembly work (Ellegård et al, 1992). What is more controversial, however, is how productivity, quality, and learning are affected. Leading IMVP researchers have strongly criticized the model, and have pointed out that lean production delivers the greatest efficiency and quality (Adler & Cole, 1993 and 1994). Some commentators have gone so far as calling the Uddevalla plant a "dead horse"16 . It should immediately be made clear that lean manufacturing organization is seen by both the IMVP- and the reflective production researchers as a modified, but basically Taylorist organization with highly standardized, short, and repetitive work cycles - and of course, the presence of an assembly line. What differs lean from traditional mass production is that workers are involved in the definition and improvement of procedures and tasks, that they have larger possibilities of work rotation and participation in quality assurance and preventive maintenance, and that the hierarchy provides support and expertise instead of control and command (Adler & Cole, 1993; Ellegård et al 1992). The radically different way of organising assembly work in reflective production is of course recognized in the critique, but nevertheless rejected as a serious alternative on the following grounds (Adler & Cole, 1993, comparing the NUMMI factory in the United States with Uddevalla): • Lean production is the more effective model for organizational learning, even if reflective production promises a high potential for individual learning. This is because in lean production all improvements are standardized and formalized in work instructions and therefore made explicit to everyone. Hence, lean production assures a higher growth rate for productivity and manufacturing quality; 16 New York Times, July 7, 1991. "Edges Fray on Volvo's Brave New Humanistic World", quoted by Berggren (1994:1, p. 1) 99 • Assembling a vehicle took about forty hours in Uddevalla compared to twenty-two as an average for Japanese transplants practising lean production in the U.S (figures from 1992); • The long work cycles in reflective production prevent workers from tracking their task performance to a more detailed level, and improving performance continually through microscopic continuous improvement opportunities. Moreover, the lack of standardized work processes means that there are no mechanisms for identifying, testing, or diffusing the improvements that individual workers might make. To summarize this critique, there are few hard data, except for the assembly hours, that support the lean way of assembling cars. It is more a question of anticipating weaker possibilities of productivity and quality improvements in the reflective model. These weaknesses are argued to be due to a lack of built-in strategies for organizational learning, or the laissez-faire attitude to micro-improvements that long work cycles could result in. Berggren (1994:1) argues against this critique: • Firstly the gap in assembly hours can be explained by differences in product manufacturability, and by the fact that NUMMI has adopted an organization developed for years in Japan, while Uddevalla was at the beginning of the learning curve. Moreover, assembly time per car was constantly reduced until the closure decision in November 1992. • Secondly, the discussion of learning opportunities remains a hypothesis: there were no data measuring the actual learning outcomes, and, moreover, learning is a process that must be studied over time. • Thirdly, a strong division of labour persists in the lean model. Kaizen activities in Toyota and Mitsubishi in Japan, for example, are conducted by specialized teams where only highly experienced and selected operators participate. Foremen, supervisors, and engineers play the key role. Besides answering the critique, Berggren (1994:1 and 1994:2) also pinpoints two productivity benefits that the reflective model offers, and that are of a slightly different character from those discussed in the critique: • Parallel team assembly and broad worker competence meant superior flexibility in model change programmes compared to traditional line assembly in other Volvo plants. For example, work-station based team groups were able to return to normal productivity in half the time after a model change, compared to the Gothenburg plant; • A wider perspective on productivity than just improving assembly performance was taken in the production groups by focusing on total cycle time from customer order to delivery. The Uddevalla plant was able to deliver cars to the Swedish market on customer orders 100 only. This customer orientation was directly built on the flexible team assembly production organization. Rather than disturbing a pre-programmed sequence on an assembly line, the introduction of customer order planning meant an additional motivational advantage for the teams. 3.4.2 Reflective Production - Still no Global Model Through the discussion above it becomes obvious that the question of productivity is related to the measurements chosen by different researchers. There is simply no best way, and priorities are certainly very different if one manufactures high-end luxury cars or more standardized products. Reflective production will probably continue to be an important source of inspiration in the work organization debate. For example, increased difficulties in recruiting assembly workers in the Japanese car industry have been reported as due to the neglect of work conditions in the implementation of productivity improvement programmes17 . Surprisingly enough this happened just after Volvo abandoned the human-centred assembly group projects, except for a short series of convertibles and sports models (among other things due to over-capacity problems). The discussion of reflective production has shown that three distinct models of production organization exist at the manufacturing level18 : mass production, lean production, and reflective production. In final assembly, reflective production is radically different from the Taylorist organization in traditional mass production, while, in fact, the lean alternative is more a refinement or perfection of this practice. However, when enlarging the perspective to product development and supply management, there are still few indications that reflective production represents a third practice in these processes19 . This is confirmed in the conceptualisation of reflective production, as well as in the debate on performance and learning opportunities; the assembly work is in focus. The literature review indicates that reflective production is indeed a unique manufacturing organization, but still not a global production model. 17 Reported by representatives of the Japanese Metal Workers' Union at the third European Ecology of Work Conference, The European Foundation for the Improvement of Living and Working Condtions, Dublin, 2-5 November 1993. 18 Manufacturing processes in the case of automobiles include the stamping, welding and painting of bodies, as well as the final assembly. Ellegård et al 1994. 19 One exception is Blomgren & Karlsson (1993), who show that inherited in the Volvo Uddevalla assembly concept (reflective production) was a potential to improve the ease of assembly, absent in a traditional organisation. Due to the holistic organisation of assembly, where each worker is responsible for several steps, they discovered a problem with the fitting of the fuel tank that was fed back to the design department. The problem was not discovered in the traditional assembly plant in Torslanda (Göteborg). This example gives an indication that a different design organisation, where assembly workers directly take part, might possibly emerge from the reflective assembly organisation. To further investigate this extension of reflective production is certainly an interesting topic, but beyond the scope of this research. 101 As this research is concerned with integrated component development, changes in organization and processes will be related to lean vs. mass production. However, it is important to recognize that a limitation to these two paradigms might be reductive if the manufacturing level is studied. 3.5 A CHANGING PRODUCTION ORGANIZATION - SUMMARY AND CONCLUSIONS This chapter has positioned lean production as a distinctive production model compared to traditional mass production. This conclusion was possible to reach through a two step analysis. Firstly, a production model was identified as a configuration consisting of coherent and original principles in product and process design, supply management and manufacturing, that delivers superior results in comparison to earlier and different configurations. Secondly, a systematic comparison of different factors in lean and traditional mass production was undertaken. The global distinctive features of lean production were identified to be a flat and organic organization based on project management, a vertical disintegration coupled with a heavy reliance on supplier engineering, and a flexible production tool with zero stock and quality control integrated into operations. Team working and continuous improvement are features present in all functions and hierarchical levels. Superior performance of the new model has been argued in the literature (Clark & Fujimoto, 1991; Ellison et al, 1995; Womack et al, 1990). However, the argument of a one best way of production organization seems to have been slightly premature (Boyer & Freyssenet, 1995). As indicated in chapter one, lean production can be assumed to undergo further evolution by integrating existing practice and acting as a driving force for the development of new managerial concepts, methods and techniques. It is to this deeper understanding of lean production that the present research aims to contribute. In order to do so a more detailed analysis of the key literature was undertaken. This literature seems to be possible to classify into four groups: •• Lean as a global system and an existing Japanese inspired model; •• Lean as a global system and a context dependent best practice under development; • Lean examined in components as an existing Japanese inspired model; • Lean examined in components as a context-dependent best practice model. The work of Womack et al (1990) clearly belongs to the first group. It considers lean principles even beyond the core elements of industrial production by integrating lean 102 distribution and industrial organization principles in the lean enterprise, the latter also further developed by Womack & Jones (1994). At the same time, their approach is very normative. They argue that lean production -conceptualized from Japanese practice- should be adopted everywhere and as soon as possible. Abo et al (1994) and Krafcik (1988) examine lean production at the manufacturing level, i.e. an analysis in components with respect to the previous definition of a production model. Krafcik (1988) clearly demonstrates that lean production is a model under development. He found that, at least at the manufacturing level, some U.S. companies were performing at equal levels to those of the best Japanese carmakers through the application of similar structures and methods. The perspective of Abo et al (1994) is more difficult to classify as far as the perspective on the development of lean production is concerned. They consider the Japanese model as a fixed best practice where adaptation is seen only as a necessity in order to adapt the basic principles to different cultural contexts. However, the different features of the model are never questioned as a result of this adaptation. Lamming (1993) finally, focuses on supply management, but in a systemic way integrating, above all, the consequences of changing buyer-supplier relationships on product design. He goes beyond a description of lean practice by anticipating the further evolution of lean supply. His work can therefore be classified in the middle between the second and the fourth group identified above. Besides providing a general framework for generating research questions concerning change tendencies, drivers for change and perceptions of lean techniques (c.f. section 1.2 and appendix 2, phase 1) the literature review undertaken in this chapter contributed to focusing the attention on a couple of specific issues. Firstly, a comparison of the argument that lean production should be adopted on a world-wide scale and the fact that the very special financial support system keiretsu -identified as a decisive factor behind the development of close buyer-supplier relationships- would be unrealistic to transfer to the West, leads to the question of what replacing organizational structures and procedures can be developed in order to support new more integrated work processes. This focuses the research on the links between different companies, including communication structures and other support systems. The analysis of the concept of the lean enterprise reinforces these questions. Through an argument based on a core competence strategy, this concept emphasizes the interface structures between different companies in a production chain and the internal process-based organizational structures. Secondly, by analysing Lamming's (1993) perspective on lean production and comparing it to the other identified perspectives, the assumption that lean production should be regarded as a 103 context dependent best practice under evolution was reinforced. Lamming (1993) emphasizes that there are differences between Japanese practice and the practice that is under development currently. The most distinctive differences in the product development and supply management problems are the active balancing of a customer and relationship portfolio on the part of the suppliers, and their independent management of product development and innovation. Particular attention will be paid to these features in the empirical research. The last section of the present chapter was devoted to an analysis of the reflective production concept - developed at Volvo in Sweden. This seemed to be necessary in order to complete the discussion of production models. It was concluded that this concept cannot be considered as a global production model because the focus is limited to the manufacturing level. Its impact on the product development and supply management problem is therefore limited. 104 4. SUPPLYING IN A NEW CONTEXT Generally speaking, new practice in buyer-supplier relationships has been initiated by car manufacturers inspired by Japanese practice. The focus of this research is on the situation of the component supplier once the customer has decided to purchase a component externally; make or buy decision problems will not be treated. The present chapter begins with a review of the changing buyer-supplier relationships, including a more precise definition of the partnership concept and an analysis of the tiered supply chain model. Secondly, different theoretical approaches to buyer-supplier relationships are analysed. This includes a review of transaction cost theory, different strategic management approaches -above all the resourcebased theory-, and theories of organizational structures related to cooperation in the component design process. As in chapter three, the literature review will generate questions to further orient the research. 4.1 IDENTIFYING THE GLOBAL CHANGES IN BUYER-SUPPLIER RELATIONSHIPS New supplier relationships are already a reality in industries where the product contains a large number of complex purchased components, for example in the car industry. Suppliers are gaining increased importance in the whole production process, because such a large proportion of production costs are transferred to suppliers, that their efficiency ultimately determines retail prices, and the quality and competitiveness of the final product (Cusumano, 1985). Heavy reliance on supplier engineering is often argued to be a strong reason for the competitive advantage of the Japanese production system. A review of the literature dealing with the recent evolution of supply management and the changing role of suppliers (Bertodo, 1991; Chanaron et al, 1993; Cusumano, 1985; Cusumano & Takeishi, 1991; Dyer & Ouchi, 1993; Hyun, 1994; Lamming, 1993; McMillan, 1990; Spekman, 1988; Womack et al, 1990) identifies the following tendencies in automobile component supply: 105 • From in-house parts supply to outsourcing, but ensuring strong collaborative ties and vigorous supplier selection procedures, • From a relation marked by mutual suspicion and hiding of information to an interchange of all information, including long-term plans, needed to be able to work together for mutual benefit in a framework of ground-rules for determining prices, ordering and delivery, proprietary rights, material supply, and quality assurance. Increased information sharing is a necessary condition for successful cooperation and integration through synchronized management, • From bidding on short term contracts (even if the relationship might be long) between a large number of suppliers competing on price, to buying from a few long-term members of a supplier group with a proven record of performance, • The smaller number of suppliers is a result of a restructuring of the supply chain at different levels - tiers, to optimize combination of complementary assets. The assembler only interacts with first tier suppliers, which means more consolidation above all in the purchasing function, • An important feature in the tiered supplier structure is system supply. System suppliers concentrate on products with high added value and sub-assemblies. The classical example is car seats where the possible 30 components in a seat bought from different suppliers and assembled on the line are replaced by one system; the seat assembled in the system supplier's plant, • From late call for suppliers to realize finished drawings, to early selection of suppliers at product development outset, not only integrating them into the design process but also giving them total design and development responsibility (development of so-called black box parts), • Assemblers offer assistance to skilled suppliers, also to those not directly involved in original development work, in improving their production organization and administrative routines, mainly working with quality and delivery to integrate them in the ground rule framework (second point above). In short, the development sketched out above can be described as a form of quasi-vertical integration: the relationships are characterized by longevity, closeness and exclusivity (Richardson 1993). As Richardson (1993) underlines, it is often argued that this development is a strong reason for the competitive advantage of the Japanese production system, and thus an important component in the lean production model. Evidently, the dimensions of collaboration and the new responsibilities have to be understood by both parties - customers and suppliers. Spekman (1988, p. 77) argues: "The relationship is bilateral; both parties have the power to shape its nature and future direction over time. Mutual commitment to the future 106 and a balanced power relationship are essential to the process". Clark (1989) makes the same point arguing that besides the degree of involvement, the quality of the relationship (information exchange, confidence, relationship management, etc.) is crucial for successful collaboration. In a historical perspective, the relatively stable supplier structure with two distinct groups, 'executing subcontractors' and 'suppliers on catalogue', that took shape in Europe after the Second World-War, has moved towards a diversity of situations structured around 'integrated design of tailor-made components' (Laigle, 1994). The latter is characterized by an integration of development skills that upgrades suppliers from the subcontracting mode, and by earlier involvement in the development process and adaptation of products to individual customer needs. This makes the supplier an integrated part of the process of designing and developing the final product instead of only selling parts from a catalogue. Central to the changing conditions in product development is the introduction of so-called black box parts, that very quickly have taken a dominating position compared to detailcontrolled parts (supplied by executing subcontractors) and supplier proprietary parts (supplied by suppliers on catalogue), (Clark & Fujimoto 1991; Cusumano & Takeishi, 1991). They define the three groups in the following way: • Supplier proprietary parts are products taken from concept to ready-made product by the supplier and sold to the carmaker through catalogue. In this category, that in the new development context tends to become less important, there is little possibility for the carmaker to influence the design and there is normally no other communication between the supplier and the final assembler than through ordering and delivery. • In detail-controlled parts most of the engineering of components is done in-house in the carmaker firm. Normally, suppliers take responsibility for process engineering and production, but sometimes even the production process is developed in-house and lent to the supplier. There is no interaction with the supplier before detailed engineering is almost finished. This practice represents the largely criticized approach where the supplier is little more than a provider of production capacity, and, as a result, becomes very poor in terms of technological capabilities. This concentration of R&D to the carmaker can also lead to sub-optimising as it is difficult for the carmaker to possess state-of-the-art technology in a wide range of fields. As a result, the core competence of total vehicle coordination and the overall competitiveness run the risk of suffering in comparison to firms relying on suppliers for component engineering. • In the case of black box parts, the assembler generates overall requirements on product functionality and performance, cost targets, and development lead time, then communicates this information to two or three potential suppliers in a logic of parallel sourcing (see chapter 4.1.2), who carry out detailed engineering and testing. This practice 107 relies on partnership and represents early involvement of suppliers and intensive interaction from concept generation and onwards. Assemblers might benefit from higher quality and lower cost due to suppliers' specialized expertise, while they maintain overall control over cost, quality and specifications of technical systems through integrated management. From the above literature review, a number of 'elements' that characterize a partnership relation, the notion generally used when talking about buyer-supplier relationships today, have been identified (e.g. longevity, closeness, exclusivity, trust, equal power structure, and mutual commitment). However, to proceed to an operational framework for assessing an expert supplier's relation to its customers, it seems necessary to provide a more strict definition of a partnership relationship and to examine the underlying theories to this concept. 4.1.1 Partnership Defined The above introduction to the literature has identified change tendencies and underlying concepts in supply management. Component suppliers to the car industry work in strictly coordinated tier structures, more and more as sole sources for a specific component to a specific model, and more and more in close technical collaboration with their customers. A common concept for characterising the relations, is the one of partnership relationships (De Banville & Chanaron, 1985). The objective in this section is to provide a clarification of the partnership concept. Several authors underline that the classification of the repeated transactions between a selling company and a buying company as a 'relationship' is relatively new (Sako, 1992; Lamming 1993; Lecler, 1993). One of the first theories that challenges traditional marketing and purchasing literature by considering the more complex situation in business-to-business transactions is the interaction, or network, approach developed by the International Marketing and Purchasing Group (IMP), (Ford, 1980; Håkansson, 1982, 1987; Johansson & Mattsson, 1987). This concept describes the buyer-supplier transaction as a long-term interaction process characterized by institutionalization, adaptations and relationships. The process is conditioned by the industrial environment (market structure, dynamism, internationalization, stage in the production chain where the transaction takes place, and social system), and the atmosphere in the relationship (power/dependence, cooperation, closeness, expectations), (Håkansson, 1982). In the IMP work, networks are regarded as a mode of organization that can be created by managers to position their firms in a favourable environment; the competitive advantage of a firm lies in its capacity to gain access to, and exploit, valued external resources and expertise through the network (Sako, 1992). As Lamming (1993) argues, the IMP model 108 conceptualizes several factors that later have been seen as relevant in the new buyer-supplier relationships: • Emphasis on both the short-term interaction process - 'episodes', and the long-term 'relationship'. The episodes, i.e. delivery on time, payment on time, information exchange, and social exchange, condition the long-term relationship and develop trust between the parties. The concept of episodes or more precisely the 'taken-for-granted assumptions' that condition the episodes are getting renewed interest when looking for structures that might condition the relationships in a similar way to the Japanese keiretsu. • Recognition of the institutionalization of expectations and behaviour patterns, something that facilitates the management of interaction for mutual gain. This phenomenon exists both in traditional 'simple' but long-term subcontracting relationships and in 'complex' long-term relationships involving integrated design. For example, Laigle (1994) found that between many subcontractors and their long-term customers, no formal contract existed at all. Ward et al (1995) found that Toyota managers explain the rapid and informal development process in their company partly by the fact that suppliers need to communicate less because of their long association. Moreover, partnership suppliers do their own prototype development without waiting to be ordered by the customer - a typical example of IMP's 'taken-for-granted' assumptions about the behaviour of one another. • Recognition of the importance of adaptations that each part must make to propagate the relationship. This rejoins the discussion in chapter three where it was concluded that mutual adjustment, and an equal's contribution are crucial for developing lean supply. As identified above, the IMP model provides a solid foundation for analysing buyer-supplier relationships and it emphasizes several elements that were empirically confirmed later (c.f. e.g. McMillan, 1990; Cusumano & Takeishi, 1991; Sako, 1992; Richardson, 1993). However, to enrich the understanding of partnership buyer-supplier relationships, it is necessary to turn to literature that explicitly analyses the 'model' of this kind of supply - the Japanese car industry. Lecler (1993) has conducted in-depth case studies of buyer-supplier relationships in Japan, with the objective of gaining detailed understanding of the way these processes, that serve as a model for the restructuring of the European automobile supplier system, are managed. She points out that the notion of partnership didn't exist in Japan before arriving from the West as a name for what the Japanese call a "system of cooperating companies" (p. 17). Furthermore, she argues that naming the emerging alternatives in buyer-supplier relationships 'partnership' relations marks a change in the position of large firms vis-à-vis other participants in the production chain. This 'change in the state of mind', as Lecler (1993) calls it, is the result of efficiency shortcomings in the traditional practice compared to the Japanese alternative. Lecler 109 (1993) also calls on IMP work and transaction cost theory by using the term network for describing an industrial organization between markets and hierarchies. Referring to Thompson et al (1991), she argues that if price competition is the essential coordination mechanism in markets, and if administrative control is the one in hierarchies, then trust and cooperation are the central mechanisms in networks between partner firms. Lecler (1993) criticizes the abuse in using the word partnership (partenariat in French) and proposes the following limitations in order to better define the concept: • Partnership is not a synonym for inter-firm relationships, but one of its principal components, • More than a specific form of inter-firm relationships, partnership reflects a change in the state of mind concerning the elaboration of such relationships. From here stems its association with the network notion, • This change in the state of mind necessitates far-reaching organizational restructuring of the involved firms. Moreover, the partnership can take very different forms, aim at solving very different problems, concern specific functions more than others, or be applied to the entire organization of the firm, • In any case, the concept of partnership is legitimate only if it marks a breakpoint compared to the nature of earlier relationships. With this as point of departure, Lecler (1993, p. 29-30) defines a relationship as a partnership only if: • The relation is long-term; • Common objectives are negotiated and not simply imposed by the stronger party; • The means for realising common objectives are clearly stated (by a written or moral contract), in order to define the more or less formal organizational structures that will support the relation; • The two partners gain benefits from the relation, which then evolves on the basis of reciprocal trust. The first three points show clear parallels to Lamming's (1993) discussion of an equal's contribution (see chapter 3.3.1), and to Helper's (1994) definition of partnership relations which take up items such as exchange of details on process steps, long-term (more than three years) relations, and customer assistance to selected suppliers. Concerning the concept of trust, this has been thoroughly analysed by Sako (1992). She argues that what underpins heavy mutual dependence as an acceptable, and even preferred, 110 state of affairs is the existence of 'goodwill trust'. "Goodwill trust is a sure feeling that trading partners possess a moral commitment to maintaining a trading relationship. It might manifest itself in not taking unfair advantage of one's circumstances (for which shared principles of fairness exist) and in offering preferential treatment or help whenever the need arises" (Sako, 1992, p. 10). This diffuse kind of trust, "trust in open commitment" (p. 38) is to be distinguished from more specific types of trust, namely contractual trust (expectations that promises made are kept) and competence trust (confidence in a trading partner's competence to carry out a specific task), both of which are necessary for the smooth working of any trading relationship. "Goodwill trust is [...] contextual and therefore verifiable only in particular settings; a buyer and supplier have to start trading and see if they entertain shared principles of fairness and convergent mutual expectations about informal obligations" (Sako, 1992, p. 39). Sako (1992) specifies that goodwill trust implies the expectation that trading partners are committed to take initiatives to exploit new opportunities over and above what is explicitly promised. An example of the operationalization of this idea is the Toyota case discussed by Ward et al (1995). To summarize the discussion in this chapter, the perception of buyer-supplier connections as relationships has its origins in the network approach developed in industrial purchasing theory. The analysis of this concept shows that emphasis is made on integration, coordination, and communication and the structures that support these features. This justifies a focus on these items for understanding the new buyer-supplier relations. Moreover, from being considered as relationships, buyer-supplier connections develop towards partnerships. There, it is important not to consider just any customer as a partner. Some minimal criteria must be fulfilled (c.f. Lecler, 1992; Helper, 1994), and the empirical research will take these criteria as a point of departure when analysing the relations that studied suppliers entertain with their customers. 4.1.2 Tier Structures and Sourcing Strategies Based on the above literature survey, the new buyer-supplier relationships in automobile component supply can be summarized in three points: • An important reduction of the number of suppliers with whom the carmaker has direct contact, • An increased design responsibility for the remaining suppliers, and A tight coordination and control over the restructured supply chain, involving specific forms of trust. Behind these changes lies the need for cost reductions, quality improvements, and lead-time reductions - both of component development and delivery lead-time, all of which are driven by • 111 the globalization of competition and the emergence of Japanese products (i.e. cars) that showed higher performance in these three criteria20 , and that have shown an extraordinary growth in the world-wide market share of passenger cars21 . On the macro level, the changes are resulting in the development of a new industrial organization; the pyramidal tier-structured supply base. Coupled with this is a new sourcing strategy for the buying part; single or parallel sourcing. Before discussing the theoretical foundation of new buyer-supplier relationships, these factors, which play an important role in the new context, will be briefly analysed. The typical Japanese hierarchical supplier system, defining the tier concept, is illustrated in figure 9. Assembly and inhouse component Carmaker production First tier suppliers Second tier suppliers Third and fourth tier suppliers Figure 9. The Japanese supplier system. This illustration is of course simplified. It only assumes one assembler, while in reality most suppliers, independent of tier position, supply more than one customer. Moreover, it does not consider the complicated real world pattern where one specific supplier is first and second tier 20 For figures see Womack et al (1990). Reduced development lead time becomes visible to the end- customer above all through quicker adaptation to switching trends in the demand. 21 From 3,6% to 25,5% between 1965 and 1989 according to the Motor Vehicle Manufacturers Association of the United States, quoted in Dyer & Ouchi, (1993). 112 supplier at the same time, either in relation to different customers or in relation to different product groups. Lamming (1993, p. 186-188) proposes another basis for the tier classification than the physical supply, namely 'supply' of know-how or development intelligence. He introduces the notions of direct and indirect supplier to illustrate this phenomenon. The following situations are identified: • Direct supplier of components and know-how. This is typically a system supplier, integrating several components into a system, which is delivered to the assembly line ready for fitting into a vehicle. • Indirect suppliers of components (no intelligence links). These supply components to the direct suppliers on a contract. • Indirect/direct supplier. This kind of supplier has a direct relationship (components and know-how) to the assembler and an indirect relationship through other direct suppliers. Depending on how aligned such a supplier is or would like to be with the car industry (thus depending on the product/market strategy of the supplier), it will tend to move either towards more direct supply, taking on more technological responsibility, or be phased out from the direct supplier base by the carmaker. It could also be possible for this kind of supplier to maintain a dual role. • Indirect influential supplier. Suppliers of high-tech products or materials tend to have intelligence exchange with the carmakers even though they supply their products to be incorporated (e.g. microprocessors) or transformed (e.g. composite materials) by another supplier to be of use to the assembler. This typology is very interesting and a large step towards a better understanding of the European situation in the car component sector compared to the tier model. It indicates that it is theoretically irrelevant to talk about suppliers only in terms of tiers; both supply of components and development intelligence must be considered. Moreover, the degree to which a supplier is aligned with the car industry, in terms of current situation and strategy for the future, has to be analysed. The evolution of a supplier's situation in such relational models leaves room for speculation, however. Firstly, how are the integration structures organized between direct suppliers, indirect suppliers, and assemblers? Secondly will suppliers automatically fall into one of these categories and, if not, what are the demands on management for maintaining a dual role? Finally, are there tensions between a struggle on the part of suppliers to position themselves in a specific situation (for example as indirect influential supplier) and the role that the carmakers try to impose on them? These are questions that will be addressed in the empirical research. Lamming (1993, p. 190) argues that the tiers (or levels) are "groupings formed by collaboration for certain purposes. Companies might be first and second tier at the same time, 113 even with the same customer". This point is, however, little discussed either in Lamming's continued analysis or in other research or in specialist or business press22 . It therefore needs further exploration in the field. Concerning changing sourcing strategies, and more precisely the abandonment of multiple sourcing that goes hand in hand with a reorganization of the supplier chain in tiers or direct/indirect levels, McMillan (1990) and Richardson (1993) provide an explanation of the Japanese model based on empirical observations of Asanuma (1985a, b) and Smitka (1991). Three strategies are identified: • Multiple sourcing. This was the practice of traditional mass production in the West until the emergence of lean. For each component, the assembler has typically five to ten suppliers in his supplier base, which are played off against each other through price bargaining. This practice, together with a high degree of internal assembly (no system supply), resulted in supplier bases with thousands of references. For example in 1988, Mercedes had 2,500 direct suppliers, Ford Western Germany had 1,300, VW had 2,600, Opel had 1,150 (De Banville & Chanaron, 1991), and PSA and Renault together had 2,200 in 1985 (Brocquet, 1995). • Sole sourcing 23 . This means one component - one supplier for all models, and has been the general perception of Japanese sourcing strategy. However, some European manufacturers have also practised this form of supply, but the situation was very heterogeneous. For example 1,000 of Opel's 1,150 suppliers were single source in 1988, but only 900 of VW's 2,600 (De Banville & Chanaron, 1991). • Parallel sourcing. This means one component - normally two or three suppliers approved for each. According to Asanuma (1985a, b) and Smitka (1991) as quoted in Richardson (1993), this is the most common type of sourcing strategy in Japan. Parallel sourcing functions in different ways, for example in relation to plants (for the same functional component and the same model, two different suppliers are used in two different assembly sites); or in relation to car models (for the same component, one supplier is used for model 1, and another for model 2). Richardson (1993, p. 342) summarizes the distinctive feature of parallel sourcing: "two or more suppliers with similar capabilities are concurrently sole-source suppliers for very similar components". A 'teetering' of opinions concerning sourcing strategies can be observed among car manufacturers; those that have a high degree of sole sources wish to enlarge and fraction their 22 For example in a document edited for the 1995 business congress Equip' Auto (Brocquet, 1995), the discussion is focused on the pyramidal tier model (figure, 4.1 above). 23 Sole sourcing is the term used by Richardson (1993). Single sourcing is a more common name for the same thing. 114 supplier base, while the opposite position is also observed frequently (De Banville & Chanaron, 1991). If there is a trade-off between the advantages of multiple sourcing (lower price due to intensive competition, diversification of the sources for technological innovation, lower risk for ruptures in supply, etc.) and those of single sourcing (customer tailored solutions, increased technological synergy between interconnected functions, and operational synergy in terms of quality control, open books in accounting, and continuous information exchange), Richardson (1993) concludes that parallel sourcing might be the solution that combines the best of these two worlds. Also Hines (1995) confirms the advantages of parallel sourcing (or network sourcing as he calls it). In the following discussion of different theoretical approaches to supplier-customer relationships, the pros and cons of these different sourcing strategies will be further analysed. Clearly, the customers' sourcing strategy will have an impact on the relationship with a specific supplier. However, the literature is essentially concerned with the choice of strategy from the buyer's perspective. A question in the empirical research will therefore be to explore how suppliers work in relation to sole or parallel sourcing, and what the impact on the management in the supplier firm is in different cases. Being aware of all these changes, how can customer-supplier relationships be analysed, and what are the theories relevant for this analysis? A review of the literature identifies three theoretically anchored driving forces for closer relationships in buyer-supplier relationships: transaction cost theory, strategic collaboration theory, and theories for operational coordination and integration. These will be reviewed below. 4.2 COST REDUCTION - TRANSACTION COST THEORY This perspective focuses on the possibility of reducing different costs related to managing and executing transactions when components are developed and change ownership. Transaction cost theory, as developed by Williamson (1975, 1981, 1985), is a central element in the theoretical model building of inter-firm cooperation. A transaction occurs when a good or a service is transferred across a technologically separable interface; one stage of processing or assembly activity terminates and another begins (Williamson 1981). The transaction has a cost: one has to find a supplier or a customer so that the transaction can take place, an agreement has to be negotiated and established, and the implementation of the agreement has to be ensured and managed (Thiétart & Koenig 1987). High transaction costs occur when assets are highly specific (to support trade between only a few parties), when uncertainty surrounds the transaction, and if such transactions also occur frequently. This is what 115 Williamson calls 'small number bargaining in a situation of bilateral governance'. Firms will choose to transact in a way that minimizes the sum of production and transaction costs. Williamson's work opposes two modes for transaction: to internalize (hierarchy) or to externalize (market). Now, in terms of transaction cost theory, a third mode exists: quasi-vertical integration or hybrid structures located between markets and hierarchies (Borys & Jemison 1989, Garrette 1991). Kogut (1988) states that such collaborative agreements (exemplified by a joint venture) differ from a market transaction (contract), because cooperation is administrated within an organizational hierarchy. They also differ from vertical integration as two (or more) firms claim ownership to the residual value of assets and claim control rights over their use. How this form of structure, to which partnership supplier relationships belong (Lamming, 1993), affect costs is a central element for transaction cost theorists. In the case of hybrid partnership structures, considerable specific investments that increase transaction costs occur from both parties. At the same time, however, uncertainty and hazards are reduced, a factor that leads to reduced transaction cost. In other word it is essential to separate different kinds of transaction costs. Richardson (1993) focuses on different transaction costs that occur in buyer - supplier relationships, and compares cost performance in traditional multiple sourcing vs. lean practice of parallel sourcing. He identifies three types of costs: • Set-up costs including search costs and supplier development costs (e.g. training and technology transfer), • Trading costs including ongoing costs of coordinating exchanges as they occur (e.g. ordering, scheduling delivery, contract enforcement, etc.), and • Competitiveness costs which are lost sales resulting from poor supplier quality, unreliable quality, etc. For a given relationship, set-up costs are fixed costs while trading and competitiveness costs are variable ones - all in relation to short-term production. Richardson (1993) concludes that in terms of these costs, the benefit of a close coordination is reduced trading and competitiveness costs. Even if set-up costs are considerably higher, sole or parallel sourcing could minimize the sum of these cost categories. Moreover, set-up costs are likely to decrease over time as switching between suppliers occurs more rarely. Thus, transaction cost theory provides support for parallel sourcing in relation to sole sourcing under the conditions that the relationships are effective an long lasting: while retaining the benefits of reduced trading and competitiveness costs in sole sourcing, parallel sourcing also reduces the kind of uncertainty, related to opportunistic behaviour and too great a dependence on one supplier, that might occur in sole sourcing. 116 In relation to the core problem in the present research, however, transaction cost theory is situated on a level quite far from the central preoccupation of the present research. Cost is, naturally, the focus in this approach, and in the model used by Richardson (1993), many elements of industry structure, product characteristics, and firm attributes that affect buyersupplier relationships are distilled into the three cost variables discussed above. For the purpose of this research, transaction cost theory provides a solid basis for evaluating the relevance of partnership relations for a given transaction, but it cannot provide explanations to the core problem of the research: what are the organizational structures needed to succeed in integrated component development? 4.3 STRATEGIC STRATEGY COLLABORATION - THEORIES OF COMPETITIVE According to Anthony (1965), there are three different levels of control, planning and decision in the firm: the operational level, the tactical level, and the strategic level. The operational level comprises the use and the management of resources: how to reach objectives and how to organize resources in order to accomplish the everyday work tasks. The tactical level is concerned with acquisition or disposal of resources. At the strategic level, the object for control, planning and decision is the firm's relations to its environment. Strategy is called the "complex and difficult exercise of choosing the general goals and objectives that the firm intends to follow" (Strategor, 1993, p. 1) and the adoption of courses of actions and the allocation of resources necessary for carrying out those goals (Chandler, 1962). It is a consistent approach over time which is intended to give results in the medium and long-term for specific problems. This implies that there is a hierarchy of decision choices so that strategic decisions will result in guidelines laying down policy to steer the more specific decisions taken by operational managers (Thurley & Wood, 1983). The motivation to collaborate for strategic reasons resides in the assumption that firms transact by the mode which maximizes profits through improving their competitive position vis-à-vis rivals (Kogut 1988). Strategor (1993) identifies two levels of strategy; corporate strategy and business strategy. The corporate strategy determines the firm's business activities. It guides the firm as to whether to enter a specific sector or leave another in order to develop a balanced business portfolio. The business strategy, or the competitive strategy, defines what steps the firm has to take to create a favourable position in relation to its competitors in a given sector. Classical strategic management concepts such as SWOT analysis, strategic positioning, 117 Porter's (1980) generic strategies24 , or the cost-volume and differentiation strategies from Blanc, Dussauge and Quélin (1991)25 , refer to these global questions. Examples of strategic decisions are: a change of the product range, diversification or refocus on core competencies, penetration of new markets, alliances and acquisitions, process and facility investments, specific projects such as ISO quality certification, introduction of total quality management, introduction of lean principles such as J-I-T, reduced delivery and development lead time, cross functional teams, and so on. What is important to understand when discussing strategy is that the strategic level is concerned with global objectives for gaining competitive advantage, translated among other things by the identification of key success factors and the development of market niches. The traditional approach in strategy analysis focuses upon the link between strategy and the external environment (Grant, 1991). However, at this level, one never learns how to achieve the goals that are set up, nor what the impact of certain projects in relation to existing processes in a company will be. As underlined in Strategor (1993, p. 3), "if a strategy, to be applied with success, supposes that the structure of the firm will follow it, a given structure influences, in a very important way, the strategy that will be chosen. A new strategy is a product of a structure that exists, but it will also, in its turn, create a different structure". Even though strategy and structure go hand in hand, research founded in theories for competitive advantage is by definition more concerned with propositions and descriptions of concepts and ideas than with questions and answers regarding their application. In this traditional perspective, Porter & Fuller (1986) present four strategic drivers for collaboration based on theories of competitive advantage. The first motive is to achieve scale or learning economies by concentrating an activity within one entity that serves both firms. Increasing volume raises the scale of the activity and/or the rate of learning about the way to perform it. The second class of motives are related to access to knowledge or ability of how to perform an activity where there are asymmetries in the competencies between firms. One partner has already born the cost of developing the ability, enjoys a preferred position in the activity, or has superior resources. The third class of motives is risk reduction. In a collaborative agreement, neither of the partners bear the full risk and cost for the shared activities. The fourth motive listed by Porter & Fuller is competition shaping (market power 24 In coping with the five competitive forces (industry competition, potential entrants, substitutes, suppliers, and buyers) Porter (1980) elaborates three potentially successful generic strategic approaches to outperform other firms in an industry: overall cost leadership, differentiation, and focus. 25 Based on the concept of 'reference offer', which is the mix of characteristics that the customer expects to find in the offer (a product group, for example a car) coming from a certain industry, the cost-volume strategy consists of proposing exactly this mix and seeking advantage only by low price due to low cost (for example the small car segment). The differentiation strategy means to differentiate from this reference offer by either improving it (luxury cars or safety cars) or 'striping' it (for example the three cylinder Japanese cars), balancing these actions by a higher or a lower price. 118 motive), as collaboration can influence with whom a firm competes and hence alter the basic rules of competition. A review of related literature (Contractor & Lorange, 1988; Hamel 1990; Garrette, 1991) shows, however, that these goals apply more to horizontal collaboration between competitors than to vertical collaboration between suppliers and customers. Strategic drivers for a partnership approach to supply management are essentially found in three more specific domains that are also woven into each other: (1) a strategy based on core competencies (Prahalad & Hamel, 1990), (2) a strategy based on technology exchanges or synergy (Contractor & Lorange, 1988), and (3) a strategy of linking complementary contributions of partners in a production chain (Contractor & Lorange, 1988). A closer analysis of these strategies, beginning with the first - focus and development of core competencies or core capabilities will be undertaken in the following sections. 4.3.1 Core Competencies or Core Capabilities Core competencies are defined as "the collective learning in the organization, especially how to coordinate diverse production skills and integrate multiple streams of technologies" (Prahalad & Hamel, 1990, p. 82). In fact, Prahalad & Hamel reactualize ideas from Wernerfelt (1984) who was the first to introduce what later became the generic notion for the theory around capabilities management, namely Resource-Based. After the 1990 paper of Prahalad & Hamel, the concept of core capabilities has had an important impact on strategic management, and the resource-based theory has been further refined and developed (Black & Boal, 1994; Conner, 1991; Grant, 1991; Kogut & Zander 1992; Mahoney & Pandian, 1992; Stalk et al 1992; Sterne 1992). This has led to a modification of vocabulary; core competence has given way to core capabilities. Stalk et al (1992) propose a distinction between the two notions based on their 'visibility' to the customer: "...whereas core competence emphasizes technological and production expertise along specific points in the value chain, capabilities are more broadly based, encompassing the entire value chain. In this respect capabilities are visible to the customer in a way that core competencies rarely are" (p.66). However, this distinction has not been further operationalized, and as the following definition will show, the notion core capabilities is used as a generic name embracing both the above situations. The resource-based theory implies a rupture with the traditional approach to strategic management discussed in the previous section. More than analysing market/product positioning or industry sectors, the resource-based approach turns to organizational capabilities, competencies, resources, routines, and skills. This approach seems to be particularly interesting in the car component sector. Carr (1993), quoted in Black & Boal (1994), found that firms in this industry utilizing a resource-based strategy outperformed 119 competitors following a market-power-based strategy on multiple performance measures. The following is an effort to define the concept. Capabilities are based on resources, in the sense that resources are inputs to capabilities (Grant, 1991). Resources include items of capital equipment, skills of individual employees, patents, brand names, finance, process and product technology, organizational structure and information transmission patterns (Grant, 1991, Sterne 1992). These can be called basic resources because they are omni-present in any manufacturing company. A capability, then, is the capacity for a set of resources to perform some task or activity (Grant, 1991). Bowen et al (1994) conceptualize the resources to a more abstract level: capabilities are composed of four interdependent dimensions: knowledge and skills embodied in a company's employees, managerial systems, physical systems, and values. The first three items are clearly a combination of basic resources while the fourth one is the 'glue' of attitudes, behaviour and norms that tie the other together. Thus, capabilities are a systems-notion to which Arrègle (1995) adds a dynamic dimension; they result from the interaction of technology, collective learning, and organizational process. This definition confirms the important statement made by Grant (1991) that there is no predetermined functional relationship between the resources of a firm and its core capabilities. Already the combination of basic resources into knowledge and skills, and into managerial and physical systems will be unique for different companies. The value dimension and the three dynamic elements of technological development, collective learning, and organizational process will further individualize the capabilities of each firm. If capabilities can be defined as above, what then are core capabilities? Core or strategic capabilities are those that are grounded in the firm's resources and differentiate it from its competitors (Lacity et al 1995). Thus, as Grant (1991) underlines, a critical task is to assess capabilities relative to those of competitors. The key to transforming a set of individual business processes into strategic capabilities is to connect them to real customer needs; a capability is strategic only when it begins and ends with the customer (Stalk et al 1992). In addition to this, core capabilities ought to be durable, clearly owned and controlled by the concerned company, and difficult for competitors to identify, understand, transfer, and replicate in order to preserve the competitive advantage they offer (Grant, 1991). Examples of core capabilities mentioned in the above literature are: total vehicle architecture (Ford Motor Co.), cross-division integration (Hewlett-Packard), silver-halide technology (Kodak), networking (Digital), dealer management (Honda), engine engineering (Honda) integration of computer and telecommunications technology (NEC), and miniaturization (Sony). 120 All of the above mean that core capabilities are collective knowledge and aptitudes embedded in the structure, unique to each enterprise, and operationalized in order to perform specific tasks and activities. In an ideal situation, the evolution of a company's resources and capabilities should be a perpetual process towards ever new distinctive competitive advantage. Arrègle (1995) argues that the dynamic character of capabilities make them create new basic resources that can be combined in new innovative ways. After this specification of the core capability concept, the literature review will be strictly limited to the way in which a core capability strategy is a driver for closer buyer-supplier relationships. From a buyer-supplier relationship perspective it is the combination of a concentration of the firm's own resources on a set of core capabilities and an outsourcing of other activities for which the firm has neither a critical strategic need, nor special competencies that are of central interest. One benefit from this is the full utilization of external suppliers' investments, innovations, and specialized professional skills that would be prohibitively expensive or even impossible to duplicate internally (Quinn & Hilmer, 1994). The core competence strategy fits very well in the car industry. Being a complex product, mobilising several 'new' technologies (micro-electronics, new materials, etc.), the automobile requires more and more in-depth and differentiated competencies pushing the manufacturers to resort to more specialized competencies that suppliers dispose of or are able to develop (De Banville & Chanaron, 1991). Hence, a strategy for focusing on core capabilities is by definition a driving force for outsourcing of secondary activities (the difficulty, of course, being to decide what should be regarded as core vs. secondary26 ). If the discussion is limited to the car industry and the industrial process of manufacturing cars, supplier relations comes in focus in this context27 . Carmakers want to keep the overall control over the manufacturing of vehicles, and orchestrate the overall production process. Moreover, they specialize in basic research and development (new materials, new production technologies, etc.), engine and transmission development28 , and final assembly. On a more detailed level, and as far as design and development are concerned, they retain total control over the concept generation process, i.e. 26 Different tools and methods helping managers in this choice can be found in the literature, one example is the analysis matrix developed by Quinn & Hilmer (1994). 27 Recent examples of the application of a core capability strategy in the car industry are the reorganisations of the Volvo and Daimler-Benz Groups. Rover also went through a similar process during the alliance period with Honda (Mair, 1994). The focus has been on how to get rid of a too dispersed portfolio of activities and refocus on the profession of conceiving and assembling vehicles. 28 Prahalad & Hamel (1990) define Honda's engine and transmission expertise as a core competence, allowing the company to gain a distinctive advantage in its car, motorcycle, lawn mower, and generator businesses. Indeed, engines and power-trains are a kind of automotive world's micro-processor. If ultimately customers look for 'Honda inside' or 'BMW inside' as PC customers today look for Intel, carmakers like them or Porsche and Alfa Romeo, or Peugeot and Mercedes-Benz (diesel engines) might envisage new competitive advantages. 121 where "information on future market needs, technological possibilities, and economic feasibility are merged and translated into a product description that embodies the experiences the product will deliver to the customer" (Clark & Fujimoto, 1991, p. 105). Through the product concept, the carmaker decides what the product will do, what it will be, whom it will serve, and what it will mean for the customer. Generating vehicle concepts becomes a core competence of the carmaker. The next step in the product development process (c.f. section 3.2 and the product development framework) is to translate the concept into more concrete specifications such as cost and performance targets, styling, and component choice. Clark & Fujimoto (1991) label this process product planning. The carmaker's job in this phase is to achieve consistency: between concept and product plans; and between specifications, component choice, styling, and layout. As exemplified by Clark & Fujimoto (1991), planning a new car model is like solving a huge simultaneous equation system. It is in the way of solving this equation system that a core competence strategy comes into the picture. A company applying this strategy will, instead of "doing too much and doing it poorly" (Womack & Jones, 1994, p. 101), confine detailed design and solution of trade-offs in component design to the experts, i.e. the component suppliers. These will be asked to take greater innovation and design responsibility and are given a free rein to come up with solutions from brief product specifications (c.f. the definition of black-box parts in section 4.1). For this reason, a strategy of core competencies in the industrial process of car manufacturing, is leading to a renaissance of the component suppliers in terms of their engineering capacity. This strategy has been beneficial for those suppliers that immediately decided to play the game and that have learnt how to derive advantage from it (Sako et al 1994). Once the core competence strategy has been applied by the final customer, it is by its nature quickly spread downwards. First tier and direct suppliers have been pushed to identify their own core capabilities: an aptitude to design and an aptitude to manufacture a well defined function (Lamming, 1993). This leads them to rationalize and apply similar strategies to their proper organizations, and to become more specific in their demands to their own suppliers, helping these, by outsourcing secondary activities, to define the core competencies of indirect second and third tier suppliers. In fact, it seems that the core competence strategy has a snowball effect which, ideally, revitalizes the entire supply chain. Stalk et al (1992) argue that a central advantage of capability based competition is that capabilities, if sufficiently robust and flexible, can serve many different businesses. Hence, companies will be less dependent on their product portfolio in favour of their capability portfolio. How this shift in strategy has been experienced on the supplier level, how it can be managed, and what the related problems are form important 122 questions for the empirical research. Here one can only point out that it has been relevant for the shift towards supplier relationships of the partnership type. Strictly speaking, a strategy of core competencies does not mean a driver for collaboration and partnership in itself. As the analysis of lean production in chapter three shows, the transfer of activities will only be productive if it is accompanied by processes for coordinating the production chain and for sharing of technological competencies. 4.3.2. Technology Synergy To begin with technology, access and appropriation of precisely defined and selected technological skills are central elements in Prahalad & Hamel's (1990) discussion of core capabilities, and they are equally important in the design of car components. Following only the logic of a core capability strategy, however, traditional arm's length relations to subcontractors (that were responsible only for the execution of in-house designed components) could in fact be transformed into arm's length relations with expert suppliers entirely responsible for the design of off-the-shelf components that the carmaker buys and assembles together into a vehicle. That this is not the case is due to at least three related factors - that also drive partnership relations. The first two have already been discussed. The nature of the product (thousands of interconnected parts that have to be linked together like a huge equation system), and the fact that the traditional core capability of carmakers is 'total vehicle architecture' (Bowen et al 1994) both make it necessary to integrate with suppliers to realise the final vehicle. The third reason is related to the fact that the rapprochement of related technologies, i.e. technology synergy is beneficial both for productivity and product innovation. One concrete example of technology synergy is system supply which may bring together previously unlinked technological competencies (Lamming, 1993). Another example is the creation of technical 'platforms' in common between the assembler and the different suppliers of a specific component or system. Both are recognised as key factors for improving functional solutions in product technology (Clark & Fujimoto, 1991; Midler, 1993; Moisdon & Weil, 1992). The theoretical founding of technology synergy is quite simple. Rothwell (1991) characterises innovation as a process of know-how accumulation based on a complementary mix of inhouse R&D coupled to the results of R&D performed elsewhere. (Twiss, 1986) shows that in mature industries (like the car industry), there is scope for renewed growth from the combination of previously disparate technologies. Von Hippel (1988) qualifies 'informal knowhow trading' as a pattern of informally cooperative inter-firm R&D that is beneficial for innovation on a general level. Chanaron & Lung, (1995) argue that the successful merger between different technologies, for example micro-electronics and fine-mechanics in anti-lock 123 brakes, fuel injections or air bags has changed the focus of innovation in the auto industry from process to product. All this leads to the conclusion that an association of R&D activities within different organizations will promote the process of producing exploitable innovations. That product technology innovations, with few exceptions, have been minor in the auto industry might be due to the fact that Western mass production industries have seen product innovation as a threat, i.e. a destruction of capital. This is because innovation makes existing investment in both marketing and manufacturing organizations obsolete (Hayes & Abernathy 1980). Abernathy & Utterback (1978) argue that new products which require reorientation of corporate goals or production facilities tend to originate outside organizations devoted to a 'specific' production system. This indicates that suppliers are an important source for innovations as they are less tied to a rigid production system than the carmakers. Relating technology synergy to the competitive advantage of the Japanese production system, Abo et al (1994) argue that one specificity of the Japanese model is that technological knowledge in one sector or activity has provided the basis for innovation in another, i.e. knowledge 'spills' from its point of origin. This is due to the tight links between buyers and suppliers. The tighter the links are, the greater the spillover and the more the course of a few industries and technologies can shape an entire economy. Gong (1993) has a similar opinion; one reason for efficiency in a partnership mode is that assemblers and suppliers use the intermediate organization to create a common space for exchange of information and knowledge. Using this common space they "make a network of information and knowledge and thus can fully utilize the synergy effect by way of pooling information and knowledge" (Gong, 1993, p. 8). He refers to this as accumulation of technology. From this one might assume that the innovations seen in the car industry coming from Japanese manufacturers, such as sophisticated electronic engine-management systems, are the fruit of tight collaboration with highly specialized and innovative component suppliers given free rein to design a specified technology. From the literature dealing with the relations between core capabilities and technology synergy, and from the analysis of the concept of technology synergy itself, two conclusions can be made. Firstly, the rapprochement between different technologies is beneficial for innovation, and, thus, is a driving force for collaboration. Secondly, this is a feature of the Japanese inspired lean production model and the result has been a variety of novel product functions. Where literature leaves a great deal of room for speculation, however, is in questions like: What are the concrete processes in the design work, where suppliers are involved, that support technology synergy? What are the relationships between the restructuring of the supply chains and the evolution of product technology? To attempt to answer such questions, the management of product technology will be a specific topic for investigation in the empirical research. 124 4.3.3 Synchronized Management As argued above, the application of a core capability strategy in combination with outsourcing of secondary activities leads to a renaissance of the engineering capability of suppliers. In comparison to vertical integration, this reinforces suppliers' incentives to innovate and continuously improve through the mechanism of market discipline (Dyer & Ouchi, 1993). In comparison to arm's-length price bargaining, it reinforces the exploitation of technological synergy between engineering departments and it can also shrink transaction costs. To complete the picture, however, some 'glue' of integrated management systems must be added in order to "build trust and goal congruence between companies" (Dyer & Ouchi, 1993, p. 58). In an industry like the car industry, with a complicated and long production chain with thousands of actors (raw-material and all tiers included), a focus on core capabilities can be successful only if it goes hand in hand with integrated management and tight control over the whole supply chain. This is clearly shown in a study of 71 first tier automotive component plants in nine countries conducted by Andersen Consulting (1994): "The high performers maintain tight discipline and control over their internal processes. This rigour extends to the whole supply chain, including second tier suppliers and the car assemblers" (p. 2). So, as in the case of technology synergy, synchronized management is a necessary condition for a successful core capability - outsourcing strategy. Some elements concerning synchronized management have already been discussed in chapter three, above all in the discussion of the lean enterprise. The first task here is to relate this to the core capability concept and to develop understanding by reference to additional literature. The inter-firm value stream in the 'lean enterprise', as presented by Womack and Jones (1994) (analysed in section 3.3.2.2), is in fact a strong concept that builds on the strategy of core competencies and the strategy of linking complementary contributions in a production chain. One of the points discussed is a specific form of career management where employees transfer from firm to firm in a production chain. This particular question is also analysed by Cusumano (1985), Dyer & Ouchi (1993) and Gerlach (1987). For example, Cusumano (1985) found that employee transfers are often the first sign of a rapprochement between an automotive customer and a selected supplier; they precede technical assistance, exclusive contracts or other partner-specific investments. Dyer & Ouch (1993) found that almost 30 percent of the top management teams at Nissan's group suppliers were former Nissan employees. They also emphasize the related issue of guest engineers -temporary employee swaps allowing for faceto-face contacts- as an important element for tying customers and suppliers together. 125 However, all the evidence concerning this practice comes from Japan. Whether a similar practice is under development in Europe remains to be seen. Integrated quality assurance systems, another element of the lean enterprise, are a reality in the European car industry today. The assemblers have their specific norms (Q1 of Ford is probably the most well known) that provide detailed guidelines on a wide range of items, for example quality policy and organization, product engineering, process engineering, control and testing equipment, quality assurance of externally purchased parts, quality in running production (ppm), etc.29 . System suppliers are also developing their own standards, and in addition there is the ISO 9000 normalization that is starting to reach a large number of suppliers also among the smaller sub contractors. Generally speaking, all French system suppliers, all specialist suppliers and an ever increasing number of sub contractors are certified by assemblers and/or ISO 9000 (Deranlot, 1993). This normalization and the related auditing procedures are a formal assurance that basic levels in process performance, organizational structure, and continuous improvement are respected. It improves among other things the rigour in work methods, the formalizing of know-how and the control of the production process (SOFRES, 1995). These items were identified earlier as important for facilitating integration. In spite of the obvious importance of normalization procedures and quality systems, little literature exists on how quality management and quality systems influence the relational structure between buyers and suppliers. The main question here, related to the objective of the research, is to see whether quality systems can play a similar role, at least theoretically, as the keiretsu systems in Japan. At the outset of the discussion of strategic drivers for a partnership approach to supply management, the complementary assets strategy was identified from the work of Contractor & Lorange (1988). However, the focus in the complementary assets perspective, reviewed by Lamming (1993), is on the identification of complementary assets or on the decision of how to approach them (through vertical integration, partnership/alliance, or market transaction) but less on how to manage their operational integration. The perspective is therefore more economics-oriented than management-oriented and except for that which has been summarized above concerning career management, quality assurance, and cost management, very little is said about the tangible means for coordination; of what they consist, how they emerge, and how they are managed. The orienting questions grounded in the literature analysis above are summarized: 29 Items in the PSA-Renault AQF standard (Supplier Quality Assurance). 126 • Does a transfer of employees between different companies take place and are guest engineers present in the supplier firm (from the carmaker or from systems suppliers) or in the carmaker or system supplier firm (from the specialized supplier)? If yes, to what degree and what does this mean in terms of improved design engineering for the medium sized expert supplier? • What governance role is played by quality systems and normalization? Beyond these questions, the empirical research will try to identify and then describe all the links (governance structure) that a supplier has to its customers. Emphasis will be made on the aspects of emergence, design, and management of such links. New information technology might also play an important role in operational coordination. Bensaou & Ventkatraman (1995) argue that the recent development of electronic data interchange makes it possible for an assembler to electronically coordinate with suppliers in information intensive activities such as contract negotiation, component design, tool development, production control, etc. Therefore, the expert supplier's use and perception of information technology will be an additional element of observation. 4.3.4 Conclusions about the Strategic Perspective The initial 'back to basics' discussion of strategy, competition and cooperation seems necessary for an understanding of the underlying assumptions in the traditional strategic management literature. A strategic management analysis of collaborative agreements traditionally takes an 'up-stream' perspective with respect to operational reality by relating the phenomenon to models of global competitive advantage creation. The resource-based approach, however, marks a breaking point with this, especially when accompanying concepts of technology synergy and synchronized management are considered. The focus on organizational capabilities, competencies, resources, routines and skills emphasize that managers must consider, to a larger extent than traditionally, the development and institutionalization of individual and collective skills. The resource-based perspective questions the assumption that strategy primarily involves competitive position in the market place (Bettis et al 1992). Capabilities built on skills and competencies, for example a capability to initiate or quickly respond to changes in the market place, can "provide a rational basis on which to build and sustain a competitive advantage over a period of years" (Bettis et al 1992, p. 14). They argue that the competitive position that a firm can maintain is a measure that will indicate whether the means of accomplishing and maintaining it, i.e. the core capabilities, are the right ones. 127 It has been argued that the outsourcing tendency is the very breeding ground for increased design responsibility for suppliers. The definition of an expert supplier given in chapter one indicates that the core capability concept is particularly important for this kind of supplier. An important issue for the empirical research will therefore be to look for traits of core capabilities, to see if they are identified by management, and, if so, to assess the way in which they are deployed, 'nourished' and developed in the organization. Throughout the literature review, outsourcing has been explicitly viewed as coupled with a core capability strategy, i.e. applied as a consequence of an identification of core capabilities. This means that the dangers of an operational, i.e. a defensive, score-keeping oriented view of outsourcing advanced by Bettis et al (1992), should be minimized. Moreover, an analysis of such dangers, for example the loss of essential manufacturing competencies, is beyond the scope of the present research because they concern mainly the carmakers and the system suppliers. 4.4 OPERATIONAL COORDINATION IN THE PRODUCT DEVELOPMENT PROCESS - THEORIES FOR ORGANIZATIONAL COORDINATION AND INTEGRATION This perspective originates from the necessity of product development engineers and technicians in buyer and supplier firms to coordinate their design efforts, and from the necessity to tear down barriers between different functions inside an organization such as purchasing, design, process engineering, and manufacturing (Clark & Fujimoto, 1994; Karlson, 1994). Traditionally, product development activities are principally considered to take place in the concept generation, product planning and product engineering phases, c.f. the product development framework in chapter 3.2. Recently, however, the impact of design decisions, and more precisely component design, on process engineering, and, ultimately, on running production has gained an increasing interest in research and practice (Clark & Fujimoto, 1991; Midler, 1993; Takeuchi & Nonaka, 1986; Wheelwright & Clark, 1992). Different techniques such as value analysis30 and design for manufacturing have been developed to systematically integrate design and manufacturing. To analyse the design process has become urgent as its impact is crucial on such problems as long development lead-time, high level of running changes during product development, unstable quality, and complex supply patterns due to carryover31 and shared parts (Clark and Fujimoto 1991). Parallel to all this, integration 30 Value analysis and value engineering are often used as synonyms. One distinction is that 'value engineering' is often used in relation to the design of new products and 'value analysis' in relation to existing products (Wild 1989, p. 78). 31 Parts used in existing models, carried over to new ones to be used there as well. 128 of component suppliers gives a new dimension to the product development process, as emphasized in section 4.1. Being aware of these problems, the different measures that are taken, or can be taken to improve the situation, will be identified. 4.4.1 Project Management One major change is an organizational one; the traditional product development matrixorganization is replaced by a project organization (Womack et al 1990; Midler, 1993). These authors describe the matrix organization, where people report both to their functional department and to a development program, as inefficient as far as development lead-time, cost, quality and technological performance of the final product are concerned. The main reason, they argue, is that employees depend on their functional manager, who might have other priorities than to lend personnel to the development programme. In a project organization, conversely, all the necessary resources for developing a new product become under the direct control of the project manager for the project life time, i.e. from idea generation to product launch (Clark & Fujimoto, 1991; Midler, 1993; Womack et al, 1990). It is also important that the project members serve on the team from the time of product concept until the running production has started (Smith and Reinertsen 1991), and that the team's character is broad (Whitney, 1988). The composition of the core project group in the case of the Renault Twingo development project, a project organized in accordance with the lean production model and characterized, among other things, by the presence of a 'heavyweight product manager'32 , highlights the complexity of inter-functional relations that have to work to successfully take a new product to the market following this organizational model. The group consisted of managers and specialists representing the following functions: engineering design (including body and chassis engineering specialists), process engineering (including assembly and stamping specialists), product management, quality management, purchasing, planning and logistics, styling, marketing and sales, finance, and communication (Midler, 1993). A major driver behind the transformation of the product development organization to one based on project groups has been the concept of concurrent or simultaneous engineering (Carter, 1992; Durand 1995; Hayes, Wheelwright & Clark, 1988). Simultaneous engineering, in the context of interfunctional product development groups, relies on strong and permanent interaction between different engineering functions during the whole design and development process; the basic idea is that a product's design and the associated manufacturing process for it should evolve simultaneously (Durand, 1995; Smith and Reinertsen 1991). In the context of 32 The heavy-weight product manager is a notion introduced by Clark & Fujimoto (1991) to describe an organization where one manager, often at the same or a higher rank as the heads of the specialist functions, acts as a general manager of a new product from concept generation to market introduction. 129 simultaneous engineering the design for manufacturing (or design for assembly) concept (Smith and Reinertsen, 1991; Whitney, 1988) emphasizes the importance of early involvement of process engineering and manufacturing personnel in product planning and product engineering, in order to avoid late changes in a design that is already evolved and where all product functions are already interdependent. Karlson (1991) shows for example that around 40% of the drawing changes in two big development projects of turbines were due to factors outside the design process. Whitney (1988) refers to examples from the car and car components industry, where he shows that better process knowledge of design personnel resulted in important cost and time savings. In order to minimize running design changes, to increase quality and speed up development time, an early involvement of process development and manufacturing personnel in product planning is necessary but not enough. A broad range of competencies in the design team, like the project group of the kind identified by Midler (1993), is also essential for the integration of functions. Furthermore, the design for manufacturing philosophy makes it possible to reverse the traditional sequence "adaptation of manufacturing to new design" and adapt the design to an innovative manufacturing process (Whitney, 1988). A project organization can take different forms. Bowen et al (1994) identify four "stylized patterns" (p. 127) of project leadership and organization. These are (p. 127-128): • A functional system, where the basic work gets done within, and leadership occurs through, the functional organizations; • A lightweight project team system, where a project manager coordinates and schedules activities through liaison representatives but the basic work and much of the substantive leadership occurs through the functions; • A heavyweight project team system, where the work is done in the functions, but a project leader and a core team of functional leaders take responsibility for all aspects of the project; and • A dedicated project team system, where the people working on the project are pulled out of their functional organizations and dedicated to a team led by a strong and empowered project leader. Furthermore, they emphasize that a project leader must have a clear concept of what the product should be and a vision of the project's mission. His or her work is about integrating the three basic objectives of the development process which they argue are functionality, coherence and fit. Bowen et al (1994) identify these as follows: (p. 136) 130 • Functionality is a basic requirement for market introduction. It concerns meeting the defined standards of functional performance. Functionality is rooted in the performance of specific components and sub-systems. • Coherency is a property of the product as a technical system. It measures the extent to which functional components and sub-systems work well together. It also applies to the interaction between the design of the product and the design of the process. • Fit is a property of the relationship between the customer and the product. It is about the match between a component or sub-system and the final product. The concept of fit is related to the product concept discussed in section 4.3.1; principal components and sub-systems must fit with what the final product will do, what it will be, whom it will serve, and what it will mean for the customer. Successful project management also calls for a structure and process that value and support leadership within an organization. The project manager must be a leader with a mandate; he or she must be respected and dispose of the status that the role demands. In the Renault case, for example, the project manager relies on the formal power he or she obtains through the direct attachment to Senior Management, i.e. the same position in the organizational chart as the Department Directors (Midler, 1993). The above discussion concerns project management in very large firms. All of the quoted references have analysed project organization and leadership within carmaker (and other large) firms. Concerning suppliers, those who are concerned with integrated product development of parts or sub-systems with the carmakers also turn towards a project organization (Laigle, 1994). A project organization is in fact a requirement for passing for example the supplier quality assurance audit undertaken by carmakers and system suppliers. Among many other things, the auditors check that "A project manager is appointed to each project in order to ensure its entire development and industrialization. His or her missions are clearly defined"33 . Thus, project management is something crucial also for suppliers, and in the present research special attention will be paid to the organization of projects and the role of project managers in supplier firms in order to complete the picture of project management in the automobile industry. One might assume that the role of a project manager in a supplier firm, that develops an important number of new products every year, is quite different from that of his or her counterpart in the carmaker firm. The argument of Durand (1995) that the role of project managers is often biased due to his or her functional position will also be considered. 4.4.2 Stage Overlapping and Intensive Communication 33 PSA-Renault AQF standard (Supplier Quality Assurance). 131 As indicated above, emphasis on integration in the process of developing a new automobile means that the sequenced stages of concept generation, product planning, product engineering, and process engineering are linked through new methods and approaches to product design and process engineering including concurrent engineering, design for manufacturability, and early supplier involvement (Clark & Fujimoto, 1991). The importance of supplier involvement for product development performance has been discussed in chapter 1.4. The increasing role of the supplier as a development resource influences the product development organization, and when talking about integrating product and process engineering, this also includes integration of suppliers. The concept of integration is, however, slightly ambiguous. Some clarification of terminology will therefore be provided before proceeding with the analysis. "The need for coordination originates in the division of labour into separate tasks performed by specialized participants and interdependence of these tasks" (Karlson, 1994, p. 151). In fact, Mintzberg (1979) defines organizational structure as a division of labour into distinct tasks and coordination between these tasks. As a consequence, he sees coordination as a core activity that has to be efficiently performed for any organization to function properly. Thus, division of labour and coordination go hand in hand, but, as Karlson (1994) argues, the fundamental reason for coordination is the interdependence between separated tasks. Now, if managers wish to improve coordination and actually integrate different functions within their firm, or develop closer relations with other firms in a specific problem solving cycle (for example the process of developing a new component), the outcome will be increased cooperation between different actors in this cycle. However, Clark & Fujimoto (1991) found that increased cooperation is necessary but not enough for successfully integrating problem solving cycles. Intensive communication, which means rich, frequent and bi-directional information flows (between different steps in the product development process), is the second crucial element. Hence, they emphasize that effective integration has both a 'hard' side and a 'soft' side: organizational structure, work processes (including methods for framing and analysing problems), skill development programmes, and formal tools for analysis and communication are important, but tend to give few results if positive attitudes to change, trust between team members and between hierarchical levels, and commitment to common objectives fail to emerge. 4.4.2.1 Tacit Knowledge Karlson (1994) goes one step further, arguing that communicating intensively will be of little use if the actors do not understand each other, for example for reasons of different professional affiliations. His research shows that there is a fundamental difficulty for one particular specialist in understanding the skills possessed by other specialists. He asks a crucial 132 question: "If the skills possessed and deployed by one group of participants are not readily understood by other groups, how can we expect communication and cooperation alone to result in "true" coordination?" (p. 158). Thus, Karlson (1994) identifies differences in skills as the core problem of coordination and integration. Without explicitly addressing theories of learning, Karlson (1994) concludes that participants have to know enough about what their colleagues do and know in order to be able to ask the right questions, to anticipate problems, and provide the necessary information during the design process. Briefly, different participants must also be able to understand non-articulated or tacit knowledge possessed by their colleagues. The concept of tacit knowledge was initially formalized by Polanyi (1966). Based on his work, Jannik (1991, p. 139), quoted in Karlson (1994), defines tacit knowledge in the following way: "tacit knowledge refers to a broad spectrum of phenomena which have one thing in common; they can only be learned through action, not through theoretical studies". Thus, tacit knowledge represents the most intangible form of knowledge, and the one that is the most difficult to formalize and therefore to acquire. An interesting theoretical approach to the phenomena of knowledge can be found in Göranzon (1993). He separates three aspects of knowledge: theoretical knowledge (acquired by studies of rules and methods), practical knowledge (acquired by practising a profession), and knowledge from familiarity (acquired by reflecting on existing practice and anticipating or pushing evolution by experimentation). The last two aspects are in fact tacit, and together with the argument that the three forms of knowledge cannot be separated, this means that complete knowledge in a task, activity or profession cannot be described in a formal language - it must be acquired through practice. Also Lamming (1993) concludes that much of what is being shared in interfirm collaboration activities is intangible. He points out that the physical assets that are being shared, e.g. employees, capital, equipment, product brands, marketing skills, distribution channels, etc., are only a means to an end. He emphasizes that the implications of what he calls the deeper sharing within a collaboration (for example learning about a new product technology) are far more significant for the real outcome of a relationship, and that risks related to the fact that the partner also could be a partner of a competitor must be consciously managed. Coming back to the initial problem of how to successfully integrate product development activities that are separated into different functions and shared between different firms, the above analysis comes to the conclusion that beyond cooperation in terms of parallel operations and intensive multilateral communication, the interpretation of information -and therefore the related knowledge development- must be given particular attention. In fact, the concept of tacit knowledge indicates that the skill a person possesses mirrors the task he is performing (Karlson, 1994). This leads to the assumption that the only true way of transmitting knowledge 133 then becomes learning through the alternating of work tasks within a more or less broadly defined spectrum of activities. 4.4.2.2 The Complexity of the Design Task The complexity of design activities is something that is heavily emphasized in the product development literature (Bowen et al, 1994; Clark & Fujimoto, 1991; Karlson, 1994). It is the strong interdependency between tasks, previously discussed, that is the main reason for this complexity. It has been given a more explicit meaning by Moisdon & Weil (1992) who provide a thorough description of the activities and processes in operative design at Renault during the realization of new vehicle projects. More than a thousand people participated in a project during the four to five years that a new car model project normally lasted at Renault at the moment of their study. During the project period, more than 3,000 components were 'interfaced' in several hundred possible configurations34 . Through direct observations of the work of design engineers and technicians, Moisdon & Weil (1992) emphasize the striking variety of work tasks that these actors, organized in sub-system design teams, perform. Neighbour components must be considered (How will they be assembled with the team's 'own' component? Are the tolerances compatible with assembly requirements? etc.); styling people are contacted with regards to form and finish; process engineering is involved to discuss the way components and sub-systems will be assembled; suppliers must be contacted for design revisions, and so on in a continuous loop where one answer leads to a new question that requires another interlocutor's answer, etc. The questions that preoccupy design engineers and technicians present several characteristics (Moisdon & Weil, 1992, p. 10-11)35 : • They are very detailed, for example dealing with the clearance between a plastic box and a rubber gasket, or the access to a cable or a screw. • They are extremely numerous; a design engineer might work with a memorized list of more than a hundred questions to be dealt with. • They are very volatile; new problems that might influence already known problems arise constantly. Problems are of multiple origins for example the evolution of product specifications, quality problems identified in a more advanced project or on an earlier 34 Since that time project lead-time has diminished slightly according to interview data. 35 Moisdon & Weil (1992) identify two main categories of questions; those related to the internal function of a sub-system, and those related to the integration of sub-systems into the vehicle. 134 model, the confrontation between blueprints of neighbour components, results from prototype testing, work on mock-ups, etc., only to mention the most significant. • They are difficult to classify in a priority or hierarchy; for a technician, a problem with a door clip can be as important as bad cold starting. The importance of a question is judged essentially in relation to the influence it risks to have on its environment (design changes and modifications in a cascade). • Generally, they cannot be resolved by a technician alone; most of the time the solution will depend on a compromise between several technical demands represented by different technicians, engineers, and functional managers. • They are long to solve; questionnaires, manipulations on mock-ups, and review of blueprints are used depending on the case. Everyone concerned must meet, the problem must be explained, some will have to be persuaded of its importance, and a new solution, i.e. negotiation of a new compromise, must be found. Then, and this is often the most time consuming step, the new solution must be tested; Is it viable in itself? Will it require other compromises? This phase is characterized by tests on mock-ups and the realization of prototypes coming closer and closer to a final solution. Because of these characteristics of the design work, Moisdon & Weil (1992) conclude that no internal or external actor is able to recompose a complete design logic beyond key dates or the local planning of a sub-assembly. No one knows exactly, at a given moment, where the approximately ten design sections and process engineering sections in a new car model project are in the development process. Moreover, there are normally another three or four vehicle projects in different phases of advancement running simultaneously. The complexity of the design task implies that the increasing demands on coordination in simultaneous engineering cannot be expected to be met by employing or 'creating' generalists (Karlson, 1994). Designing a new product is usually much too complicated for a single person, no matter how competent he or she may be. Karlson (1994) concludes that as a consequence of this, project groups consisting of participants with narrow and strictly specialized skills is something that we have to live with. The question is then how to facilitate integration and organizational knowledge development between these specialists. 4.4.3 Different Theoretical Concepts for Facilitating Integration and Coordination The analysis undertaken in the previous section has emphasized the problems of cooperation and integration that persist even though structural support in terms of project organization and parallel operations (stage overlapping, i.e. simultaneous engineering in the product 135 development process) are present. The literature suggests that simply communicating more won't be enough to solve these problems. Principally based on the works of Karlson (1994), some possibilities for improving this situation will be reviewed. The following ideas and concepts all deal with integration of specialists with different skills. 4.4.3.1 Common Cognitive Ground Nonaka (1991) argues that one reason for the success of some Japanese firms is redundancy as a principle of organizational design. He defines this as "the conscious overlapping of company information, business activities and managerial responsibilities" (p. 102). Nonaka (1991) argues that even though redundancy might sound unappealing, building a redundant organization is the first step in managing the knowledge creating company36 . He further argues that the organizational logic of redundancy is to facilitate the spreading of knowledge through an organization and helps different actors to internalize new knowledge and information, also tacit. The tangible result of redundancy, which facilitates the transfer of tacit knowledge and explains the success of simultaneous engineering in Japanese companies, is the creation of a common cognitive ground. This can be interpreted as an intangible space of common knowledge and understanding. From the analysis of the works of Nonaka (1991), Karlson (1994) concludes that improved coordination and ultimately integration is possible only with redundancy of knowledge; "The point is that information transfer alone will not do the trick, since the information has to be actively interpreted by the receiver. In order to be able to interpret information, the participants need to have overlapping skills" (p. 169). Thus, only communication that fits into the common cognitive ground will be interpreted in a relevant way and be useful for its receivers. The above discussion seems to indicate that creating a broad common cognitive ground would solve many coordination problems in product design. Karlson (1994) argues that the common cognitive ground emerges quite naturally between individual participants whose activities need to be coordinated through working together, discussing, and learning from each other. Because the participants know that the problem that they are working on has to be solved in a way that is acceptable for all parties (otherwise the product is never realised), they engage in a dialogue, which hopefully results in the development of a common cognitive ground. Also Moisdon & Weil (1992) confirm the existence of a spontaneous, intensive and informal coordination between technicians; "at the basic level, everyone talks to everyone and all the time" (p. 12). The development of such interaction presupposes that technicians develop 36 The title of his paper and also of his later book; Nonaka & Takeuchi (1995). 136 networks with strong common work habits founded on social parameters such as comradeship and shared humour. The authors conclude that organizational isolation is more present in the vertical relations between design technicians and their superiors than between technicians belonging to different design subgroups or even to process engineering or supplier organizations. So, even though the first obstacle to integration -the intangible space of common knowledge and understanding, or the common cognitive ground- can be removed through overlapping skills and redundancy, there might still be problems of integrating contradictory priorities and standpoints in a product development project, above all technical and economic considerations. Some concepts dealing with these problems will be further discussed. 4.4.3.2 The Creative Negotiation and the Whetting Dialogue Midler (1993) argues that in order to succeed in inter-functional project work both rational analysis and personal commitment in taking responsibilities are necessary; "only a principal of personal responsibility can ensure convergence [in problem solving]" (p. 80). The significance of this responsibility is that everyone must identify, as early as possible, all obstacles that might hinder the successful realization of his or her commitment and inform all those concerned about them. If managed in this way, the notion of responsibility becomes a driver for anticipated problem-solving negotiations where convergence is achieved also through adjustment of initial targets. The principal advantage that Midler (1993) identifies in this line of action is that the future task executors are involved in negotiating the objectives that their downstream activities will try to achieve. Midler (1993) applies this concept of the creative negotiation both to internal inter-functional problem solving activities, and to interfirm relations between buyers and suppliers. The core argument is that the negotiation involves both operational design technicians and their most direct superiors and people responsible for cost targets, quality targets and the commercial fit of the final product. A concept related to the creative negotiation is that of the whetting dialogue (Odhnoff, 1987, 1993). This is defined as a step-by-step opening of dialogues on different issues with "a core of essentially contested concepts" (Odhnoff, 1993, p. 11). In other words, it is a process that takes place in a situation where very diverging opinions (essentially contested concepts) with regard to a problem are confronted in order to reach a common solution. Karlson (1994) applies this concept to the product development process because "In the design process the participating specialists often represent such different standpoints that it is meaningful to talk about a core of 'essentially contested concepts' at the root of the dialogue", (p.169). 137 The whetting dialogue and the creative negotiation have common objectives; sharing intellectual development, integrating the understanding of a basically common problem (developing an optimized product) and developing trust among the participants. "The participants in the dialogue all benefit from keeping it alive, not only in order to reach agreement but also to develop knowledge each one on his or her side" (Odhnoff, 1987, p. 136). Karlson's (1994) core argument, that coordination in the deeper sense must be performed by the individual specialists performing the task since they are the ones who need to create a common cognitive ground, is reflected both in the creative negotiation and the whetting dialogue. The project manager must of course be permanently present and able to understand the professional language of all the specialists (Bowen, et al, 1994) but his role should be to create a dialogue between the participants and not only to centralize and diffuse information by him- or herself. 4.4.3.3 Structural Support Following the above discussion, an important managerial challenge would be to create both formal structures and a favourable climate for the coming together of project participants. The question of trust is important in order to cope with problems such as that of "silent professions" (Midler, 1993, p. 124), which means that function representatives present in a working group do not always express their opinions but remain essentially inactive in the debate. If all possible trade-offs are not explicit, it is impossible to find an optimal compromise, and problems will occur in a down-stream crisis when an improvised solution will have to be found in an emergency. The literature does not provide any direct indications of how to manage the kind of problems that are related to trust and a strong instinct of avoiding conflicts at the operational level. Moisdon & Weil (1992) observed that design technicians had a tendency to leave such problems aside and to wait for the hierarchy to arbitrate. Concerning organizational support, the creation of project platforms by the carmakers is a recent and tangible effort for bringing together different functional specialists. A project platform is a physical space in the carmaker's design department where all concerned actors meet for working on mock-ups and blueprints. Platforms exist both at the overall vehicle level and at the sub-system level. System and expert suppliers are frequently invited to work on the platforms (Midler, 1993; Moisdon & Weil, 1992). The project platforms give a spatial identity to the project, something that traditionally was only true for functional departments (Midler, 1993). He draws up a list of several advantages with the project platforms (p. 70): elimination of administrative delays due to information circulation between different design departments and other concerned functions; direct contact between technicians instead of via notes and memos; transparency in exchanges (when working directly together it is easier for everyone to 138 envisage what can be expected of the others - in formal meetings one can only rely on appearances); concrete and integrated work on mock-ups and prototypes facilitates the reaching of an optimal compromise (discussed above). Midler (1993) concludes that coming together physically is not enough for getting to the bottom of all the design problems; the way in which people communicate must be analysed. This proves the relevance of a strong focus on the act of communication that has been argued for in the previous sections. However, the empirical evidence concerning the tangible effects (reduction of lead-time and cost, quality improvements) of the platform organization is still relatively weak, and the subject is also situated in the periphery of the present research because the platforms are an element of the carmaker's organization. Nevertheless, the perception and importance for suppliers of the platform organization will be studied in the empirical research. 4.4.4 Conclusions About the Operational Coordination Perspective A new organizational structure based on project management has emerged in the product development function both in carmaker and supplier firms as an attempt to respond to an increasing need for integrating different design activities. Different authors argue that in a development process there is often a spontaneous interaction between technicians while the difficulties of integrating different hierarchical levels as well as the marketing, purchasing and quality functions are much greater. If one can talk about a 'ground line' (Moisdon & Weil, 1992) or spontaneous organization of the development process, this is what the project organization tries to improve and complete. One of the core arguments presented in the analysed literature is that the need for coordination and integration is rooted in the division of labour -for reasons of specialization- in combination with task interdependence. It was also argued that complexity and specialization are inherent in the design process and that in order to achieve integration the participants in a design project must create a common cognitive ground in order to be able to make use of information, and communicate efficiently. The fact that only communication which fits into the common cognitive ground will be interpreted in a relevant way indicates that the development of such a ground -whether it be through an organizational design strategy based on creative negotiation, the whetting dialogue or something else- must take place before or maybe parallel to the creation of new communication channels and information transmission structures. Furthermore, because the literature indicates that important spontaneous integration takes place between operational 139 design technicians, it might be more efficient to design the product development organization around this natural work practice of organizational integration. 4.5 SUPPLYING IN A NEW CONTEXT - SUMMARY AND CONCLUSIONS The objective for this chapter was to analyse literature dealing with new buyer-supplier relationships in the automotive industry and the theoretical explanations behind the changes that these relations have gone through lately. Firstly, a review of the management literature dealing directly with automotive supply identified the main changes and trends within this area. The central importance of increased design responsibility for suppliers was confirmed, and this happens as a consequence of a strategy based on vertical disintegration together with an important reduction of the number of direct suppliers. The concept of partnership, often applied to the new buyer-supplier relations, was discussed. It was concluded that partnership is not synonymous with inter-firm relationships but one of its principal components. Moreover, to be able to talk about partnership some minimal criteria of change in the state of mind concerning the elaboration of buyer-supplier relationships must be present. Organizational restructuring of the involved firms is also a necessity for developing a partnership. The emerging industrial organization model -the tiered supply chain- was also analysed. This led to important research questions dealing with a more detailed picture of the relationships that supplier firms can be supposed to have within this model. After this overall analysis of the problem area, the literature review focused on three different theoretical approaches to buyer supplier relationships; transaction cost theory, strategic management theory and theories for organizational coordination and integration. This literature review allowed for a gradual focusing down on the core problem of the research; organizational efficiency for integrated component development. Transaction cost theory, firstly, was found to reduce many of the structural and organizational elements that are central for the present research into cost variables. Even though this approach provides a solid basis for evaluating the relevance of partnership relations for a given transaction, it cannot provide explanations for the core problem of the research. Based on an analysis of strategic management theories, it was argued that the classical strategic management concepts dealing with the link between strategy, the external environment, and the market positioning of a firm vis-à-vis its competitors are of limited use, both for explaining vertical cooperation and for analysing concrete applications of different concepts presented as strategically favourable. To approach the reorganization of the supply chain from a resource-based perspective appeared to be more useful for understanding the overall logic of outsourcing. It was argued that a strategy of core competencies does not mean 140 a driver for collaboration and partnership in itself; the transfer of activities to suppliers is likely to be productive only if it is accompanied by processes for coordinating the production chain and for sharing technological competencies. The last theoretical discipline to be reviewed was that of theories dealing with organizational coordination and integration. The analysis took the project organization as the point of departure. It was argued that the structural support in terms of parallel operations and communication channels is insufficient for ensuring efficient integration. In fact, the literature suggests that the process of achieving integration goes through three steps. Firstly, there is a need for simultaneously designing components and sub-systems and their corresponding production processes. Secondly, the diffusion of information and the design of communication channels must be taken into consideration. Rich, frequent, and bi-directional communication (between different steps and actors in the product development process) is a central element for increasing concurrency in design activities. Thirdly, the interpretation of information and the tacit dimension of knowledge must be taken into account. More communication will be of little use if the participants don't understand each other. How knowledge develops must therefore be considered and the creation of a common cognitive ground, i.e. a space of common understanding and negotiation, must be given particular attention. More than a sequenced step-by-step process, it can be assumed that the consideration of these different parameters should itself be approached in an integrated way. The literature suggests in fact that a bottom-up approach to the organizational design of the product development process -beginning with the informal communication structures that exist at the operational level- would increase the robustness of the entire structure. When bringing the resource-based perspective and that of organizational integration together, several common points can be identified. The first emphasizes that managers must consider the development of individual and collective skills and the institutionalization of such skills to a much larger extent than was traditionally the case. The second deepens and gives a more explicit meaning to the knowledge concept and emphasizes that a collective knowledge development is necessary for getting the most out of simultaneous operations and increasing information exchange. 4.6 REVISITING THE OVERALL RESEARCH QUESTIONS The literature review in chapter three and four has stimulated a large number of questions that intervene at different levels in the management and organization of an automotive supplier firm. 141 This section will revisit and further specify the research questions based on the initial ones formulated in section 1.5.2. 1. How does the emergence of new industrial principles take place in expert supplier firms? This is an overall question dealing with the establishment of the lean context in which expert suppliers are supposed to operate. In order to answer it, the evolution of lean production as perceived by the studied suppliers will be analysed in terms of different driving forces for change. The analysis of the keiretsu system opened up questions concerning what organizational structures can be developed to support more integrated work procedures. The literature review indicated that two levels of support mechanisms are likely to emerge; one more formal inherent in the lean principles that are being developed, and another more informal consisting of ad-hoc links related to the need for operational coordination at the design engineer level. The first level will be studied in the section dealing with the lean context, while the second will be analysed when focusing on operational design work. Research question number two and three will further specify the lean context 2. What is the place and the role of expert suppliers in the automotive production chain? At the level of the industrial organization and the relationships to customers and other suppliers, the empirical research will analyse customer relations of expert suppliers in terms of both physical supply and development intelligence. 3. What lean production techniques are used, to what extent, and are managers satisfied with their application? In order to answer this question the framework in table 9 will be used. In the design organization the extent to which flat organizations and project management are applied will be examined. Concerning the production arrangements, flexibility in production tool and worker skills will be studied. Furthermore, the application of JIT and quality control integrated in the manufacturing process will be examined. Finally, whether or not a continuous improvement strategy is developed, and whether or not interfunctional skill acquisition is regarded as important for career advancement will be explored. The remaining research questions deal with organization and strategies for lean component development. In the light of the literature review, research question four and five are regrouped and a new research question five is created based on the importance of knowledge management identified in the literature. 142 4. How in practice is integrated component development realized? How are work processes designed to support integration? In relation to organizational integration and coordination, the research will be focused on the structural support for integrating problem-solving. Project management and the role of project managers will be studied. Moreover, all structural support for integration, such as meetings, teams, project platforms, etc. will be studied. Special attention will be paid to the degree of participation, the satisfaction, and the transfer of experience in and with joint activities. 5. How is knowledge dissemination and creation managed in expert supplier firms? Management of the knowledge creation process was identified as a crucial element for integration in the literature review. Concerning knowledge creation, how this happens in direct design work, and what factors facilitate and block learning mechanisms will be analysed. The way in which different actors communicate will also be analysed both within the supplier firm's organization and in the customer interface. What information is exchanged and how are problems with understanding managed? 6. Are there common characteristics between companies forming a 'high performance' organizational structure for lean component development? No general answer to this very complex question is possible without combining a qualitative and a quantitative research methodology. However, a major objective identified in the literature review- for suppliers that wish to reinforce their position is to develop product engineering based core capabilities to respond to the logic of outsourcing and remain in direct contact with carmakers. A first objective in the research will therefore be to look for traits of core capabilities and analyse the way they are identified, managed and developed in the organisation. A second objective related to the issue of 'high performance' is to develop a model of a state-of-the-art example of product development strategy and organization in expert supplier firms. This model should integrate the issues identified as crucial in managing the product development process through to the qualitative research approach based on induction of data. 143 5. RESEARCH METHODOLOGY The previous sections have defined the research problem, the initial research questions, assumptions and perspectives, and reviewed related theory. The objective of this chapter is first to discuss some general aspects of methodology and the methodological choice in this research. Then the research design will be presented including a discussion of how the collected data have been analysed and some aspects of reliability and validity. 5.1. METHODOLOGICAL CONSIDERATIONS When facing a research problem, the methodological choice is as difficult as it is important. It depends on several factors, for example the context in which the research will be undertaken, the researcher's initial preference for a certain research philosophy, the research questions, the object that will be studied, the existing theory in the field, the access to data, the time available, and so on. These elements, and above all the interaction between them, have to be clearly analysed at the outset of the project. This is the objective of the present section. Generally speaking, there is a choice either to "elaborate a set of hypotheses from the study of theoretical models - developed beforehand, and test them through an important number of closed questionnaires in a way that the mass of data is adapted to a statistical analysis", or to "define an area of research and undertake an inductive line of action leading to the development of a set of propositions generated from open interviews and archival research on a limited number of cases" (Rispal 1993, p. 134). At a more detailed level, Bergadaa & Nyeck (1992) define four types of research logic: • The quantitative deductive logic, which seeks to determine if numerous objects, that are representative of the problem at hand, dispose of the properties and relationships anticipated by a theoretical model or not; • The quantitative inductive logic where the objective is to find specific relationships between a large number of objects and describe them in a model, which can be generalized in the space where the observations were made; • The qualitative deductive logic which seeks to explain the characteristics and behaviour of some existing objects following a set of predefined relationships in a model; 143 • The qualitative inductive logic which seeks to define the essential characteristics and behaviour of an object and understand to which conceptual framework it belongs. It is about developing models, concepts, or theories grounded in field observations. Hence, there is a distinction between qualitative vs. quantitative research on the one hand, and inductive vs. deductive research on the other. However, in deciding which approach to use in a specific research project, several things have to be considered, for example; the philosophical position of the researcher, the research questions, the research object, the research strategy, and the relation to existing theory (Easterby-Smith et al, 1991; Yin, 1989). These features will determine in which of the categories the research will be carried out. To illustrate this, some definitions of qualitative and quantitative research will be analysed. Then, the above features will be discussed in more detail, and a model for methodological choice will be developed. 5.1.1 Qualitative vs. Quantitative Research: An Attempt to Distinguish the Two Van Maanen (1983) defines qualitative methods as "an array of interpretive techniques which seek to describe, decode, translate and otherwise come to terms with the meaning, not the frequency, of certain more or less naturally occurring phenomena in the social world" (p. 9). This definition relates the label qualitative to the objective of the research and to the nature of the research questions - are we looking for meanings or for frequencies? Miles and Huberman (1994, p.1) discuss qualitative research in the following way: "Qualitative data, usually in the form of words rather than numbers [...] are a source of well grounded, rich descriptions and explanations of processes in identifiable local contexts. With qualitative data one can preserve chronological flow, see precisely which events led to which consequences, and derive fruitful explanations", and further, "good qualitative data help researchers to get beyond initial conceptions and to generate or revise conceptual frameworks". These propositions relate the label qualitative to the kind of data that are gathered, (which, of course, are a function of the research questions), and also to the objective of the research - to describe and/or to explain. The question of qualitative vs. quantitative research is also related to methods. Methods refer both to the techniques used to gather information - data collection methods, and to the ones used to interpret this information - data analysis methods. In the case of qualitative research, the data are words collected through observations, interviews (discourse analysis), or documentary analysis (Miles & Huberman 1994, p. 9). In quantitative research, the main data collection methods are closed questionnaires and research in data-banks. Concerning data analysis, Easterby-Smith et al (1991) argue that in qualitative research, data gathering and 144 data analysis are interactive processes taking place simultaneously. To proceed in the research and develop new questions, data must be analysed immediately. In quantitative research they argue that the process of data analysis becomes distinct from, and occurs after, the collection phase. There is no need for interaction as the questions are formulated in a way that the gathered data will either confirm or reject them. Moreover, the analysis methods in quantitative research -essentially statistics- are only relevant to use once all the data are gathered. A problem with qualitative research is that analysis methods are less explicit than in the quantitative case - the most obvious problem is that formalized statistical methods can rarely be used. This is not necessarily because of the nature of the data, but because normally there is a limited number of observations which don't allow statistical generalization. Qualitative data deal with meaning, and meaning is analysed through conceptualization. Categorizing data and making connections between categories constitutes the core of qualitative data analysis (Dey, 1993). The specific techniques for doing so will be presented in section 5.3 dealing with the research design. Thus, besides the issues of research objectives, research questions, nature of the data, and methods, the issue of quantity in terms of the number of observations for the same question must be considered. An example will illustrate how questions of a qualitative kind are coded in a quantitative form to allow for statistical analysis when they are asked to a large number of companies. The objective with a certain research project was to make three parallel surveys of small to medium-sized manufacturing companies situated in three different geographical regions (Birchall, Chanaron & Soderquist, 1996). The purpose was essentially to identify, in terms of frequency, different company profiles, different driving forces for organizational change, different ways of managing innovation, and to analyse the relationship between different management techniques and the performance of organizations. The overall research questions were: What are the main driving forces for change and how are these managed? How can different techniques be related to innovation? Are there similarities between different regions? The first question could be answered with a couple of case studies, but the rest of the questions necessarily called for quantity in terms of number of observations. Hence, a closed questionnaire was chosen as a data collection method. Now, the things that were looked for were not numbers in the first place, i.e. examples of detailed questions were: Are your products subject to rapid technological change? What is the role played by R&D in your industry? What is the capacity of managing increased global competition, shortage of key skills, new opportunities for the innovative use of technology, and so on? These questions are in fact of a qualitative nature, but asked in questionnaires distributed to hundreds of companies. In such a situation, the use of quantitative analysis methods is not only possible, but 145 necessary. The qualitative information is therefore coded in quantitative form for the purpose of data analysis. From this example one can conclude that also concerning the number of observations, the distinguishing items will be the overall research questions (if one is looking essentially for meanings or frequency/numbers), and the research objectives (describing and explaining processes vs. verifying or rejecting hypotheses). Hence, this example is one of quantitative research focused on the frequency of originally qualitative data coded into a quantitative form. At this point it is convenient to underline, as does Dey (1993), that in social science, numbers depend on meanings and that meaning is informed by numbers. In other words, qualitative and quantitative research complement each other. After this introduction to research methodology, five different phases in the process of developing a research methodology will be analysed. For each topic the position in the present research will be indicated, and a process model for methodological choice will be developed simultaneously. 5.2 FIVE STEPS IN THE ELABORATION OF A RESEARCH METHODOLOGY 5.2.1 Real World Context In the present discussion of methodology, an initial assumption is that before undertaking a research project, the researcher has some knowledge about the field - both from a practical and a theoretical perspective. Thus, the notion 'real world context' refers to (1) the industry/sector, company/organization, or managerial situation in which data will be collected, (2) the researcher's specific practical and/or orienting theoretical knowledge about current managerial tendencies and problems in an area of disciplines, and (3) the previous experience and personal interests of the researcher. That a little piece of theory is present already at the project outset does not matter for the relation to theory that will be adopted later on. Even an inductive line of action, that develops theory grounded in observations, cannot start out without some initial assumptions and knowledge about the real world context in this sense (Silverman, 1993). The real world context provides an initial and overall definition of what the research will be about. In the present project, the choice of problem area -industrial production, and industry/sector -the car industry, can be explained by the researcher's background and personal interest. Moreover, previous experience of managerial problems emphasized the difficulties in small to medium-sized companies of rendering new strategies operational and the 146 importance of designing work processes to support this operationalization in a successful way. Thus, an initial interest for organizational structure and performance was developed. Literature studies in the field of product development and supply management emphasized the important restructuring of the supply chain and the new role of suppliers in product design. However, the literature was found to deal more with the carmakers' perspective than that of suppliers', more with the development of general frameworks and strategic objectives than with their operationalization, and more with the implementation of an existing and superior industrial model than with its development through interaction with alternative existing and/or new practice. Thus, an interest in supplier problems and the dynamic evolution of industrial models emerged. Moreover, the importance of an operational perspective was confirmed. In this way, the real world context contributed to an initial formulation of the research issue as 'New practice of product development and supply management (in a production organization perspective) in the automobile industry focusing on the situation and structure of expert suppliers'. 5.2.2 The Research Problem: Philosophical Standpoint, Objective and Overall Questions Once an idea of what the research will be about has been established, some initial questions related to the identified issue, and the main objectives of the research must be reflected upon. In doing this, the way that the researcher looks at the world in terms of subjectivity/objectivity, i.e. his or her philosophical standpoint towards knowledge formation (Easterby-Smith et al, 1991), will come in as a parallel factor to the real world context and largely determine what will be the overall research questions and the objectives of a certain research project. At the basis of all academic research are two philosophical positions from which a researcher approaches a problem: the phenomenological position and the positivist position. EasterbySmith et al (1991, p. 27) characterizes these positions in the following way: • The phenomenological paradigm sees the world as socially constructed and subjective, and the observer is part of what's observed. The researcher should focus on meanings, try to understand what is happening, look at the totality of each situation, and develop ideas through induction from data; • The positivist paradigm sees the world as external and objective, the observer is independent in relation to the observed. The researcher should focus on facts, look for 147 causality and fundamental facts, reduce phenomena to their simplest elements, formulate hypotheses and then test them. When recalling the classification from Bergadaa & Nyeck (1992), the qualitative inductive logic would be the core representative for the phenomenological position and the quantitative deductive logic the one for the positivist approach. Even though the positivist/phenomenology dichotomy represents two extremes and most research 'borrows' ideas and methods from both (Easterby-Smith et al 1991; Miles & Huberman, 1994; Silverman, 1993) the basic world-view held by an individual researcher (or institute) is an important factor for the development of a research design. The philosophical position is to a large extent a personal conviction of a researcher. Different philosophical perspectives prompt a variety of methods that are a reflection of alternative combinations of political and value considerations which are part of the world in which we exist (Stiles, 1995). However, parallel to a certain determinism depending on personal values, the choice of a philosophical position is also related to the research questions, i.e. to the things that the researchers wants to find out about the chosen research issue. Regarding the issue identified above, one might ask to what extent different factors (such as inter-functional teams, competence development of personnel, development of core capabilities, etc.) are seen by general managers as important for managerial attention. Then, one might ask if there is a relation between the degree of importance accorded to the different factors and the company's performance (e.g. turnover, growth-rate, new business operations with customers). These type of questions are focused on facts and reduce phenomena to simple elements. They mirror a positivist standpoint where the researcher positions him/herself in an external 'objective' position. One might also ask questions like 'How have core capabilities emerged in the company and how are they managed?' or 'What are the difficulties in managing inter-functional teams?' or yet again 'For what reasons have customers chosen competitors to a larger extent for certain products lately?' These questions require direct experimental contact with the phenomenon under investigation; they are concerned with causal links, and the researcher must study the phenomenon from inside in order to get some understanding of ongoing processes. By asking such questions to the initially identified research issue and comparing them to the real world context, a testing of the match/mismatch between the latter and the questions will inform the objective part of the choice of a philosophical position. If the questions that are asked match the real world context, the philosophical position that they reflect will be 148 reinforced. At the same time, the matching questions will be refined and maybe further specified through readings and pilot investigations. They will be transformed into more specific overall research questions. In the present research, such a line of action argued in favour of an essentially phenomenological position. This choice was also influenced by a personal learning objective related to the doctoral research; to develop concrete insights concerning managerial work in the automobile component sector from inside organizations. One might ask what is meant by overall research questions, and why this label is used here. The reason is that it is useful to separate the general research questions, which are asked at the project outset, from the more detailed ones used in the fieldwork, simply labelled 'questions' from now on. It is also important to separate the overall research questions from the eventual hypotheses that will be determined later on in the case of a quantitative deductive logic. Overall research questions will always be present regardless of the kind of research logic that is chosen (c.f. the discussion of the real world context above). Miles and Huberman (1994, p.22) identify four characteristics of overall research questions, which also explain why it is important to define them as precisely as possible: • They make theoretical assumptions more explicit, • They focus the study on what one wants to know first, or most, in order to focus and limit the data collection, • They help in making sample decisions, one will look at some actors in some contexts dealing with some issues, • They set the rough boundaries of the research, at least provisionally. Yin (1989) distinguishes five forms of research questions: 'who', 'what', 'where', 'how', and 'why' questions (p. 17). In the light of the definitions of qualitative and quantitative studies above, the discussion of the choice between the two approaches, and the examples discussed in relation to the choice of a philosophical position, it becomes clear that the research questions strongly influence this choice: 'what' questions in the sense 'how much' or 'how many' refer to numbers, i.e. quantitative analysis, and the most appropriate research strategies are surveys or archival analysis (for example the study of financial data for a large number of companies in databases). 'How' questions, 'why' questions, and 'what' questions of an exploratory kind, are, on the other hand, concerned with coming to terms with the meaning, not the frequency, of a certain phenomenon, i.e. qualitative analysis. 149 In the present project, the literature review and exploratory interviews with managers and specialists1 identified the critical aspects of the problem, allowing for a first sorting of theoretical concepts and an identification of priorities to examine further on in the research. In other words, this phase determined the overall research questions through a process like the one described above. As can be seen in section 4.6 the overall research questions are mainly of a 'how' and exploratory 'what' kind, so a qualitative approach both in the sense of questions -dealing with meaning rather than frequency- and number of observations -a limited number of organizations to be observed in-depth- was chosen. In the process of defining a problem, the objective of the research becomes a result of the cognitive loop between real world context, philosophical position and overall research questions. One can distinguish between mainly three objectives or purposes with a research project: to explore a vague problem or a new area of research; to describe - i.e. observe and visualise the situation of certain phenomena; to explain the causality between different observations or the reasons behind a certain situation concerning the phenomenon (Evrard et al, 1993; Yin, 1989). Evrard et al (1993, p. 21) define exploratory studies in the following way: "The exploratory study is useful for exploring a vague problem in order to determine a certain number of more precise propositions or specific hypotheses, and to understand a phenomenon in depth, in all its aspects. An exploratory study can be used to obtain as complete a vision of a certain phenomenon as possible". Exploratory studies are characterized by flexibility in the use of methods in order to acquire both an in-depth understanding of a subject and a global vision. In such a case, an exploratory project is not only a preparation for other studies, but can form a complete study itself. Exploratory studies have at least three characteristics: small samples, quite heavy and costly (time consuming) data gathering processes, and interaction between the observer and the observed (Evrard et al, 1993). The objective with descriptive studies is to provide thorough descriptions and interpretations of social phenomena, including their meaning to those who experience them (Dey, 1993, quoting Tesch, 1991). A descriptive study typically tries to trace the sequence of interpersonal relationships over time, describe a subculture that had rarely been the topic of previous studies, and discover key phenomena (Yin, 1989). Explanatory studies finally, aim at establishing causal links and logical relationships between phenomena, and providing causal explanations (Yin, 1989). In an explanatory study, the researcher's objective should be to develop competing explanations for the same set of events 1 This will be described more in detail in section 5.3. 150 and to indicate how such explanations may apply to other situations - the major concern is to make causal statements (Yin, 1989,; Evrard et al, 1993). The research objective does not automatically define a qualitative or quantitative research logic. Even though an exploratory objective is almost exclusively linked to qualitative approaches with preferred methods like documentary studies, case studies and non- or semidirective interviews, Evrard, et al (1993) argue that both descriptive and explanatory objectives can comprise qualitative data as well quantitative, and small samples as well as large. Thus, without having defined the overall research questions and the philosophical position, a determination of research objective makes little sense. In the present project the objective is exploratory, and to some extent descriptive. The research questions approach the problem both in depth and breadth. The objective is to acquire as wide an understanding as possible of the lean product development in expert supplier firms. The process leading to the definition of the research problem can be illustrated as in figure 10. The cognitive loop testing the match/mismatch between real world context, philosophical position and overall research questions is made visible. After having defined the research problem more explicitly, three additional methodological choices must be reflected upon before a methodological choice can be made: the research object or unit of analysis, i.e. the unit in the real world context where data will be gathered; the research strategy, i.e. the way in which data will be gathered; and the relation to theory, i.e. a deductive or an inductive logic. 151 DEFINITION OF THE RESEARCH PROBLEM REAL WORLD CONTEXT Match / Mismatch Match / Mismatch PHILOSOPHICAL POSITION Match / Mismatch OVERALL RESEARCH QUESTIONS Match RESEARCH OBJECTIVE Figure 10. Definition of the research problem. 5.2.3 Unit of Analysis The research object - the unit of analysis - is critical for the development of a research methodology. To define the research object rigorously is still more important when the research strategy is case studies. Hence, Yin (1989) underlines the importance of defining 'what is the case?' In chapter one it has been defined (in the way described in the previous section) that the research concerns expert suppliers in the car industry and more precisely integrated product development seen in the perspective of this type of firm. A possible unit of analysis could be a component development project (a certain component for a certain model and customer). It could also be the supplier interface (all the links that exist between a supplier and its clients), or the product development process in each supplier firm. Now, possible units of analysis are compared with the definition of the research problem. If a match occurs between the overall research questions and a unit of analysis that is identified as possible to approach, the definition of the research methodology can proceed. If, conversely, the accessible unit of analysis presents a mismatch with the research questions, the research questions must be reformulated. In the present research, the most appropriate of the available units of analysis in comparison to the research questions was the product development process in the supplier firm. Focusing on such a complicated process that depends on a wide range of actors, on technology, and on administrative systems calls for a research strategy that allows for an integration of meaning and development of relevant concepts in an interactive approach: events - consequences. 5.2.4 Research Strategies 152 Yin (1989) identifies five different research strategies: experiment, survey, archival analysis, history, and case study. He argues that all can be used for the different research objectives identified above - to explore, to describe or to explain. For example, there might be exploratory case studies, descriptive case studies, or explanatory case studies. He argues that the choice of research strategy depends on three other conditions: the type of research question asked; the extent of control that an investigator has over actual behavioural events; and the degree of focus on contemporary as opposed to historical events. The first two points are of interest to examine further in the elaboration of research methodology in the present research. The overall research questions are of a 'how' and exploratory 'what' kind. In fact, the use of lean principles, the drivers for change, the structures and work processes, and the reflection upon performance are all related to organizational and managerial aspects that have an impact on product technology. A broad array of aspects related to these phenomena must be studied. Concerning the extent of control over events in the context of the research questions and the unit of analysis, this is weak in the present project for three reasons: the limited contribution from literature on the chosen problem; the exploratory and descriptive art of the research questions; and the complexity of the unit of analysis involving an understanding of a broad context. In general, case studies are the preferred strategy when questions dealing with relationship patterns and links between phenomena are being asked, when the investigator has little control over events in new and/or complex areas, when theoretical frameworks are weak, when the context has to be taken into consideration to gain understanding of the event under observation, and when the focus is on a contemporary phenomenon within some real-life context (Yin 1989, Evrard et al, 1993). Thus, an iterative line of action involving an analysis of the match/mismatch between the research problem (essentially the research questions) and the research object determined the research strategy as one of case studies of the product development process in expert supplier firms. 5.2.5 Relation to Theory Theory can be defined as a set of explanatory concepts, i.e. a set of conceptual labels placed on discrete happenings, events, and other instances of a phenomenon (Silverman, 1993; Strauss & Corbin 1990). As stated at the beginning of this chapter, two main relations to theory can exist in a research project: the inductive approach which develops new theory from data, and the deductive approach that develops hypotheses from existing theory and then test 153 them with data in order to modify or extend the existing theory. The two processes are illustrated by Pras & Tarondeau (1979), see figure 11. INDUCTION DEDUCTION Application of theory Construction of theory Theory Theoretical work Empirical work Theoretical work Empirical Generalizations Test Hypotheses Empirical work Observations INDUCTION DEDUCTION Figure 11. Deductive vs. inductive research. After Pras & Tarondeau (1979). An indication of which one of these approaches will be most appropriate is given already in the real world context. That is, the nature of existing theory in a defined research area determines the relation to theory that will be adopted, and this happens in the following way. If the existing body of literature is weak, development of new theory is essential, and the testing of existing theory might be insignificant. Conversely, if the existing body of literature is important but bias or contradictions are found in it, a testing of hypotheses generated from the existing theory in a new context, on a new population, etc., might be the most fruitful. Whether existing theory is used to generate research questions to which the answer will refer back to that same theory, or whether existing theory is used as an orientation in the problem area and the researcher then leaves it aside and develops new theory from the field observations, is very much a question of the appropriateness between what the researcher finds in existing theory and what he or she defines as a problem. However, as Silverman (1993) points out, without a theory there is nothing to research. Initially, existing theory plays an important role regardless of what relation to theory develops later during the research process. It is a common misunderstanding that in inductive research, for example when building grounded theory2 , existing theory should be left out as much as possible. What is 2 Grounded theory is defined by Strauss & Corbin (1990) in the following way: "A grounded theory is one that is inductively derived from the study of the phenomenon it represents. That is, it is discovered, developed, and provisionally verified through systematic data collection and analysis of data pertaining to that phenomenon. Therefore, data collection, analysis, and theory stand in reciprocal relationship with each other. One does not begin with a theory, then prove it (or rather prove or disprove hypotheses 154 important to distinguish is what role existing theory has in the two approaches and that these roles are different in different phases of the research process. Initially, the roles are similar (and that is where the statement of Silverman becomes relevant). Existing theories are a part of the real world context forming the research problem; they are used as a source for researchable problems, as secondary sources of data, and to stimulate questions to guide initial observations (Strauss & Corbin, 1990). The differences between induction and deduction emerge after the problem definition phase. In a deductive logic, all literature that is relevant to the phenomenon must be reviewed to discover gaps in the understanding. Then, these gaps are conceptualized through hypotheses that will be tested through the data to modify or extend the same theory from which the hypotheses originated. To be able to do so, all variables relevant for analysing the theoretical gap must be defined before data collection, from the existing theory (Strauss & Corbin, 1990, emphasis added). Conversely, in inductive research and development of grounded theory, the objective is to discover relevant variables and concepts, and the relationships between them. A list of predefined variables are likely to get in the way of this discovery. As a result of the inductive process, a new theoretical framework evolves during the research itself. If this emerging theory has some relationship to already existing theories these can be incorporated but only as they prove themselves to be relevant to the data gathered in the grounded study. The aim is to avoid all commitment vis-à-vis a certain existing theory (Van Maanen et al 1982). In the process model for methodological choice, the choice of relation to theory has been put at the end. This is because the model reflects the situation in the present project, and because the model is intended as a support for inexperienced researchers, confronted with these issues for the first time. Moreover, in management research, a real world related problem, more than a purely theoretical one, is normally the impetus for research. Management research should "lead not only to a better understanding of management, but also to a better understanding for managers about how best to go about their work" (Easterby-Smith, et al, 1991, p. 1). Hence, a conscious choice of relation to theory is made only after the process of identifying the problem and deciding on research strategy. Theoretically this is also the most 'correct' procedure. According to Strauss & Corbin (1990) the basic premise is that the research questions should dictate the method (method in the sense of building grounded theory or testing hypotheses). Many researchers have a preference for a certain relation to theory, and more experienced researchers with a thorough knowledge of the theory in a field might very well chose an inductive or deductive approach before formulating their questions. Different lines of action can easily be argued for, but as underlined by Strauss & Corbin (1990), the essential thing is that the researcher is true to the problem at hand. developed from theory, comment added). Rather, one begins with an area of study and what is relevant to that area is allowed to emerge" (p. 23). 155 Finally, the relation between research strategy and theoretical approach is flexible. The most common relation to theory in survey research, using statistical generalization for data analysis, is without doubt the deductive one, while an analysis of methodological literature (e.g. Yin, 1989 and Strauss & Corbin, 1990) show that in case studies, using analytical generalization (c.f. section 1.2), both relations are relevant. Then again, the choice refers back to the research question. Moreover, Dey (1993, p.7) points out that qualitative analysis requires a dialectic between ideas and data: we cannot analyse the data without ideas, but our ideas must be shaped and tested by the data we are analysing. In the present project an inductive relation to theory was adopted because existing literature was found to deal little with a supplier perspective on the operational aspects of integrated product development. The use of a literature review in order to delimit and guide the research and improve timeliness in the questioning means that the research design is slightly 'hybrid' concerning the relation to theory (Stiles, 1995). It comes close to the realism approach defined by May (1993)3 , as a line of action that emphasizes the need to back up and explain observations from 'inside' through the use of theoretical frameworks. 5.2.6 Methodological Choice - A Summary Only after the considerations analysed above, can a research methodology be identified as quantitative/qualitative - deductive/inductive. In this way the methodological choice becomes a combination of the research problem, the slightly more technical considerations about the research object and the research strategy, and the relation to theory. Once a consistent methodological choice has been made, corresponding data collection and analysis methods will be chosen and used. Figure 12 illustrates the completed process. The research process that takes off once the methodological choice has been made is described in the following chapter. 3 Quoted in Stiles (1995, p. 9) 156 DEFINITION OF THE RESEARCH PROBLEM REAL WORLD CONTEXT Match / Mismatch Match / Mismatch Match / Mismatch OVERALL RESEARCH QUESTIONS PHILOSOPHICAL POSITION Match RESEARCH OBJECTIVE UNIT OF ANALYSIS Match / Mismatch Mismatch Mismatch Match RESEARCH STRATEGY Match / Mismatch Mismatch Mismatch Match RELATION TO THEORY Mismatch Deduction Induction Predefinition and test of Determination of theory a theoretical model. from observations. THEORY EXTENSION THEORY DEVELOPMENT Match / Mismatch Match Methodological choice Qualitative Qualitative Deductive Logic Inductive Logic Data collection and data analysis methods, appropriate Quantitative Quantitative for the chosen methodology Deductive Logic Inductive Logic Figure 12. A process model for methodological choice. 5.3 RESEARCH DESIGN The objective of this section is to make the research process in the present project explicit by describing the research design. A research design is the overall configuration of a piece of research. Besides data collection and analysis methods, it describes what kind of evidence is gathered from where, and how such evidence is interpreted in order to provide answers to the 157 basic research question (Easterby-Smith et al, 1991). It is the logical sequence that connects the empirical data to a study's initial research questions and, ultimately, to its conclusions (Yin, 1989). The field study and analysis phases of the present research project reflect almost exactly the list of recurring features of qualitative research from Miles & Huberman (1994, p. 6-7) in combination with some terminology from Yin (1989) and Strauss & Corbin (1990). In order to answer the overall research questions: • The research was conducted through an intense contact with a 'normal' field situation reflecting the everyday life of organizations, • The objective was to gain a 'holistic' (i.e. systemic, encompassing, or integrated) overview of the context under study and to explain the ways people in particular settings manage their day-to-day situations, • The researcher tried to capture data on the perceptions of local actors 'from inside', through a process of deep attentiveness, • Field notes were analysed in order to isolate and conceptualize specific themes that were reviewed with informants and analysed and reduced step by step, • Most analysis was done with words, using specific coding procedures in order to contrast, compare, analyse, and bestow patterns upon them. 5.3.1 The Companies in which Data were Collected In chapter one it is specified that the study concerns specialized suppliers that possess specific capabilities and for that reason remain in important direct contact with the carmakers, so called expert suppliers. Interviews and case studies were conducted in automotive expert supplier firms that: • Employ between 70 and 500 persons in independently managed units, • Supply at least three manufacturers (implicitly implying supply also on an international basis4 ), • Are former suppliers on catalogue with a clear objective of being partners in integrated component development and staying aligned with the car industry, • Have an explicit strategy for R&D and organizational development, and dispose of an engineering department headed by an independent functional manager, 4 In the project, Peugeot and Citroën are considered as one company - the PSA group. Their purchasing activities are coordinated within one division, the SOGEDAC. 158 • Supply components providing a value adding function, largely depending on the suppliers' own R&D efforts, and containing externally purchased parts with low technical content. This is essentially a criterion sample strategy (Miles & Huberman, 1994) which means that all cases meet some criterion useful for ensuring the quality of the research. Multiple case sampling was chosen in order to improve the reliability of the findings; "by looking at a range of similar and contrasting cases, we can understand a single-case finding, grounding it by specifying how and where and, if possible, why it carries on as it does" (Miles & Huberman, 1994, p. 29). It is essential to underline that the issue of generalizability does not change with multiple case studies; it is still a question of analytical generalization. The choice of cases representing certain characteristics as outlined in point 1 to 5 above, reduces the problem of representativeness - "a perennial worry of case-study researchers" (Silverman, 1993, p. 160). The cases are representative of a specific group of companies in which a research interest has been identified through a review of existing theory. 5.3.2 Organization of the Empirical Research and Data Collection As specified in chapter one the research is organized around two main research topics containing seven themes for data collection, see table 11. The topic guide (appendix 2) develops questions relative to these themes as a function of the preceding literature analysis. TOPICS LEAN PRODUCTION IN THE COMPANY THE PRODUCT DEVELOPMENT PROCESS THEMES DATA COLLECTION METHODS Place and role of the firm in the supply chain. Semi-structured interviews, documents. Driving forces for change. Semi-structured interviews. Change and development tendencies in the product development process. Semi-structured interviews. Use and perception of lean techniques. Semi-structured interviews, direct observation. Evolution of product technology and product functions. Semi-structured interviews, direct observations. Organizational structure and work processes. Semi-structured interviews, documents and direct observations. Coordination activities and communication structures. Semi-structured interviews, documents and direct observations. Table 11. Outline of research topics and data collection methods. The first topic aims at setting the context for lean component development: 159 • Firstly, the place and role of the studied company in the supply chain will be identified. The objective is to map the customer relationships of expert suppliers, to analyse the situation relative to the tier model proposed in the literature, and to identify strategic and operational impacts of different customer relationships. Specific questions concern the role (including information transmission) of the supplier in relation to customers, the strategy in the tier context, and the question of increased design responsibility. • Secondly, the driving forces for change perceived during the last five to six years will be analysed. This responds to the research question concerning the emergence of lean production in the European car industry. The question concerning drivers for change was asked as an open question to General Managers and Product Development Managers. • Thirdly, current change and development tendencies in product development management and organization will be examined. The objective is to analyse the current and anticipated evolution of lean product development responding to the question of the emergence and development of lean principles and the role of expert suppliers in it. Direct questions in the topic guide concern how companies have responded to identified driving forces, the possibility of being demanding towards customers, current strategies and priorities in product development, and the ways of achieving them. • Fourthly, the use and perception of lean techniques will be studied. This theme comprises the questions covering whether lean production is seen as a global system or not, and if it is adopted to a local country and company context or not. In conjunction with the analysis undertaken in chapter three, the topic guide develops questions on time-management, flexibility, quality, organizational structure, and the relation between new concepts and industrial reality. The second topic focuses on the core problem of the research, i.e. management and organization of the product development process in expert supplier firms: • Firstly, the evolution of product technology and functions will be examined. Product technology is the ultimate outcome of the product development process. The process of developing new product technologies will be analysed in search of patterns of organizational evolution and core capabilities. Questions in the topic guide concern recent product innovations and their sources, the influence of organizational change on product technology, strategies for developing core competencies, and guidelines and performance measurements for realizing strategic goals. • Secondly, this part of the research will focus on organizational structure and work processes in product development. The objective is to answer the research question of how, in practice, integrated component development is realized, and how engineering departments, inter-functional cooperation and cooperation with customers are organized. 160 The topic guide includes questions about the typical development project in the studied companies, the structure of customer involvement in different development phases, and the description of the project organization. • Thirdly, coordination activities and communication structures will be analysed. Questions concern coordination with customers, organization and support for internal coordination, areas of improvement in integration, coordination and cooperation, and strategies for learning in product development. Based on an investigation of these research topics, where the emphasis is made on the second, the objective with the research project is to develop a model of the product development process in expert supplier firms and the principal parameters influencing it. Figure 13 provides a summary of the research activities including the dimension of time. 1993 Initial literature review, development of research proposal and topic guide 1994 Pilot interviews, testing of topic relevance Suppliers 1, 2, 3 and 4 Final version of the topic guide 1995 Formal interviews with general managers Suppliers 3 and 4 Formal interviews with product development managers Suppliers 3 and 4 Data analysis: open coding Data analysis: open coding, data reduction and development of case study topic guide First case study Supplier 3 1996 Second case study Supplier 4 Data analysis: open coding, data reduction and comparishon of concepts with case study one Second series of formal interviews (product development managers) Suppliers 5, 6, 7 and 8 Data analysis:Axial coding, comparison of findings with relevant literature Figure 13. Organization of field activities. For practical reasons, only suppliers situated in the Rhône-Alpes Region in France fulfilling the requirements in section 5.3.1 were contacted. As argued in chapter two, this region can be considered as representative for France. Moreover, the condition of supplying at least three customers eliminated suppliers that were too regionally oriented. The natural entry to a 161 company passed through the General Manager. Four General Managers in four different supplier companies, two Purchasing Managers and one Engineering Manager in three different carmaker companies, and one industry expert were informed of the research topics and asked to evaluate the project résumé (appendix 1)5 . Between one and three hours were spent in face-to-face discussions with these people in order to refine the propositions and collect valuable input concerning relevant topics. This procedure can be referred to as a relevance test of the research outline. The messages from these pilot interviews are summarized in Table 12. TOPICS LEAN PRODUCTION IN THE COMPANY THE PRODUCT DEVELOPMENT PROCESS THEMES MESSAGES Place and role of the firm in the supply chain. Supplier firms are experiencing a variety of relationships to their customers confirming a transition of their situation. A strict tier structure of the Japanese type is not very likely to emerge. Change and development tendencies in the product development process. Confirmation of the increasing design responsibilities. Drivers for change. Just-in-time delivery and delivered quality has driven change from the mid 1980s. Suppliers feel that they are performing well on these criteria. New drivers are demands for independently generating product innovations. Use and perception of lean techniques. Established use of just-in-time, project management in product development, flat organizations and interfunctional teams. Evolution of product technology and product functions Questioning of the product scope offered. Incorporating more engineering in products, development of assembly activities. Organizational structure and work processes. Important reorganization of design departments accomplished or in process. Coordination activities and communication structures. Dramatically increased information exchange with customers during the last few years. Team work is being implemented but problems in making operational people contribute actively. Flatter organizations not always followed by a shift in mentality towards greater internal openness. Table 12. Messages from the pilot studies. Two of the four General Managers in the pilot study supplier companies accepted to be interviewed in a first series of interviews following the topic guide. The topic guide (appendix 2) contained questions for General Managers and Product Development Managers, the two categories of people that were the objects for semi-structured interviews. The questions were 5 After approval of the research proposal from my research supervisors. 162 generated from the literature review in chapters three and four and completed with input from the pilot phase discussed above. In these two companies, interviews were also conducted with the Product Development Managers. These two companies also accepted a case study of their product development process to be conducted. During the case studies, the researcher interviewed, observed and discussed with the following actors: design engineers/project managers, design technicians, process engineering staff, quality managers, purchasing managers, and production managers. The case studies comprised a second series of more in-depth interviews with the General Managers and the Product Development Managers and an intensive period of observation of the product development process lasting for 40 hours in each company. After each of the case studies, a summary report was sent to the General Manager and Product Development Manager within four months after completion of the study. This allowed for verification and feed-back on major findings. Then further data analysis resulted in some additional questions which were verified by returning to the case study companies. The case studies lasted for about one year between the first interviews and the follow up presentations of findings. Data from the first case study was rapidly analysed in order to focus the topics or open up new ones. In this way the predefined set of research areas (table 11) was extended or reduced, what Miles and Huberman (1994) call "conceptually driven sequential sampling" (p. 27). This data analysis resulted in a case study topic guide (appendix 3) which was used in the second case study. The latter therefore became more focused, and questions in discussions with operating design engineers and technicians more straightforward. However, in direct observation, an openness was privileged in order to gain new valuable input to the ongoing data analysis. The case studies relied on documentary analysis (see appendix 4), semi-structured interviews, open discussions and direct observation. Observations are fundamental to understanding a specific setting in qualitative research (Silverman, 1993). The particular strength with real-time observations (during meetings, discussions, and presence in the design department6 ) is that the context in which data are collected and the process involving interlocking series of events become visible (Bryman, 1988). Concerning documents, quality manuals, customer files, performance indicator histories, job descriptions, and project status records were used both as an introduction to the context of a specific company, and, as Silverman (1993) underlines, as elements or end products revealing the practices leading to their existence and actual form. Interviews are important for receiving messages that managers want to give. Observations 6 I asked to be installed in the design office in order to always be present also during work on field notes and preparations for forthcoming interviews or discussions. 163 within the framework of a case study allow then for comparing discourse and practice; if there are differences their reasons could be sought after. During interviews a certain number of general guidelines developed in methodology literature were respected: • Firstly, semi structured interviews with open-ended questions are the most appropriate when the opinions and meanings of the interviewees are looked for; the objective being to gather an 'authentic' understanding of people's experiences (Silverman, 1993, p. 10) • Secondly, some specific techniques for animating the discussion were used. A too passive attitude from the interviewer, that might create a problem for the interviewee about what is relevant (Silverman, 1993) or force him/her to talk only to fill up a silence, was avoided. The topic guide served as a tool for conducting the interviews. Moreover, if the interviewee hesitated or gave imprecise answers questions like 'Could you tell me more about this?' 'Could you specify further'? or 'What are you thinking about in particular'? could be used (Godener, 1996). It is important in such cases not to give any suggestions to interviewees. • Thirdly, the critical incident technique was useful to start out the interviews or enter into new subjects. It is efficient when managers have difficulties in articulating answers to specific questions. It consists of asking the interviewee to describe specific events that have had a critical impact on the way a specific question or domain has been managed. Finally, as pointed out by Godener (1996), the respect of certain rules for conducting semistructured interviews calls for some experience. A reason (pertaining to the real world context) contributing to the choice of a qualitative interview/observation approach was the researcher's experience in conducting such activities in previous work. In order to further explore and specify the identified concepts a second series of interviews were conducted with Product Development Managers in four additional expert supplier companies after the completion of the case studies. This allowed both for further development and integration of counter examples to the emerging concepts. Thus, in all eight companies were studied; four in the pilot phase, two of these in the case study phase, and four additional ones in the follow-up interviews. The two case study companies that have been called COF (Company Fastening) and COR (Company Regulation) are presented in appendix 5 which also includes a list of all the eight studied companies. 5.3.3 Data Analysis 5.3.3.1 Data Opening 164 Field notes from semi-structured interviews and direct observation were initially analysed according to the open coding technique (Strauss & Corbin, 1990). This analysis process follows the following steps: Labelling Events Data are broken down and conceptualized by taking apart an observation, a sentence, a paragraph and giving each separate incident, idea or event a name that represents a phenomenon. Incident after incident in the field notes are compared in order to give similar events the same name -conceptual label- and nuance and enrich each concept. Categorising The labelling phase results in a large quantity of concepts. In this project around 80 concepts emerged from interviews, discussions and observations. Concepts that seem to pertain to the same phenomena are then regrouped -categorized- and given a conceptual name that should be more abstract than that given to the concepts grouped under it. Categories have conceptual power because they are able to pull together around them other groups of concepts or subcategories. Practically, categorising consists of going through all concepts and asking questions such as 'What is this concept about?' or 'Is this concept similar or different from the one before or after?' It consists of identifying "similar phrases, relationships between variables, patterns, themes, distinct differences between subgroups, and common sequences" (Miles & Huberman, 1994, p. 9). In this way, categories are discovered when concepts are compared against one another, and concepts become characteristic components of a category - so called subcategories. If a concept seems not to pertain to an already identified category, it is left aside and might become the entry to a new category as data analysis goes on. Categorizing is long and time-consuming. To categorize the data from the first case study took about two weeks of full-time work. This was a very important phase of the research process, because the categories generated become the basis for further data collection. They indicated what to focus on in the following research, and where to go to find complementary information. Developing Categories Concepts, categories, and subcategories have been identified above. For example, an important issue in the present research was to assess different elements of the product development structure. For example, one element that was identified as a category was group working. This category contains all information that was found in the data concerning team working, meetings, and any collective effort to realise a specific work task. Subcategories are for example managerial objectives, work group efficiency, and work group support structures. 165 Developing categories means that the properties of subcategories and their dimensions, i.e. location on a continuum are defined. For example, the sub category work group efficiency has among its properties discussion topics, and participation. Discussion topics can be dimensionalized by asking whether only problems, problems and causes, or problems, causes and possible action are discussed. Participation can be dimensionalized by asking whether discussing a point is management driven or participant driven, and whether each individual intervenes in the discussion and with what frequency. The identification of properties and their dimensionalization specifies how group working functions in a particular setting (i.e. a specific meeting or a specific team in a specific company). This makes it possible to model and conceptualize the way in which team working functions, compare the situation in different settings, and relate different sub-categories and categories to each other in order to gain understanding of the phenomenon under observation. 5.3.3.2 Data Reduction / Data Condenzation The above has been concerned with the opening of data, i.e. the techniques and procedures that are used to visualise, explain, and undertake a first conceptualization of data. The following step in Strauss & Corbin's (1990) model aims at regrouping and linking categories to each other, i.e. condensing data. Data opening leaves the researcher with a set of categories and presumed sub-categories that have to be linked together in a rational manner. The objective is to identify main categories - phenomena, and make connections between them and their sub-categories representing a precision of the phenomenon. The tool for doing so developed by Strauss & Corbin (1990) is called the paradigm model, see figure 14. The phenomenon, also called main category, is the focal point in the paradigm model. Causal conditions are the features that give rise to the phenomenon. The context is the particular set of conditions within which the phenomenon occurs and the action/interaction strategies are taken. Intervening conditions are the broader structural context pertaining to a phenomenon. Action/interaction strategies refer to the processes by which the phenomenon is handled, managed and carried out. They are procedural and purposeful. Finally, the consequences of these strategies are distinguished. (A) CAUSAL CONDITIONS => (B) PHENOMENON => 166 (C) CONTEXT => (D) INTERVENING CONDITIONS => (E) ACTION / INTERACTION STRATEGIES => (F) CONSEQUENCES Figure 14. The paradigm model, source Strauss & Corbin (1990, p. 99). The specifying features in the paradigm model give a phenomenon precision, and are, in fact, the sub-categories, discussed previously, classified under more appropriate headlines indicating their type. Strauss & Corbin (1990) label the line of action just described axial coding. It allows the researcher to build a more precise model. After open coding, there is not enough precision in each of the discovered categories. Some sub-categories might be missing, others might be put under the 'wrong' main category. Still other identified categories might be discovered as not being main categories, but sub-categories to another phenomenon. The paradigm model helps to sort this out and to ensure that the emerging model consists of well grounded and distinctive phenomena. During the process of linking sub-categories to a phenomenon according to the paradigm model, an essential activity is to ask questions in terms of the conceptual labels about how one category might be related to another. This is the moment to return to the data and to look for evidence that supports or refutes the questions. 5.3.3.3 Iterative Testing of Questions/Propositions and the Emergence of the Conceptual Framework An interaction between the activities of formulating questions, gathering and analysing data, developing concepts, reformulating questions, further focused data gathering/analysis, and modifying concepts is characteristic of qualitative inductive research (Dey, 1993; Miles & Huberman, 1994; Strauss & Corbin, 1990; Silverman, 1991), and this line of action has been referred to several times in the present chapter. This "conceptually driven sequential sampling" (Miles & Huberman, 1994), or "analytic induction" (Silverman, 1991) is a continuous process that takes place from the very beginning of data analysis and continues until the last piece of field data is collected. It takes place within the same case and between different cases. It is repeated until a universal relationship is shown. Emerging categories are then confronted with a 167 formalised body of knowledge in the form of constructs and theories that prove themselves to be relevant to the data gathered (Miles & Huberman, 1994; Strauss & Corbin, 1990). The result of this process will be presented in chapters six to eleven dealing with the results and issues raised from the field study. Before concluding the chapter on research methodology, however, some remarks on reliability and validity are necessary. 5.3.4 Reliability and Validity Reliability consists of "demonstrating that the operations of a study -such as the data collection procedures- can be repeated, with the same results" (Yin, 1989, p. 41). It is a question of documenting the research procedure (Kirk & Miller, 1986). In the research project, data from interviews, open discussions and observations exist in three forms corresponding to the outline of Spradley (1979) quoted in Silverman (1993, p. 146): • Directly taken field notes - from interviews and observations, • Expanded typed notes made as soon as possible after the field work (this includes comments on problems and ideas that arise during each stage of the fieldwork and that will guide further research), • A running record of analysis and interpretation (open coding and axial coding). Keeping a record of these different documents helps to improve the reliability of field notes. Yin (1989) distinguishes two types of validity applicable to exploratory or descriptive case studies; construct validity which means "establishing correct operational measures for the concepts being studied", and external validity which means "establishing the domain to which a study's findings can be generalized" (p. 41). The tactics proposed by Yin (1989) to reassure both construct validity -the use of multiple sources of evidence7 , the establishment of a chain of evidence, and letting key informants review draft result reports-, and external validity -the use of a replication logic- have all been respected in the present project. Firstly, each case study relies on documents (see appendix 4), semi-structured interviews and direct observation providing multiple measures of the same phenomenon. Secondly, the establishment of a chain of evidence (also relevant to reliability), which means that an external 7 Originally developed by Denzin (1970). 168 observer should be able to "follow the derivation of any evidence from initial research questions to ultimate case study conclusions" (Yin, 1989, p. 102), is ensured by the existence of the documents listed above under the discussion of reliability8 . Thirdly, the interview topic guide and the case study topic guide (appendices 2 and 3) provide input to the chain of evidence. Finally, the product development manager in the two case study companies reviewed the case study report on his/her own firm9 . External validity concerns the generalizability of a study. As mentioned in chapter one, case studies are concerned with a replication logic that Yin (1989) labels analytical generalization in opposition to statistical generalization used in surveys. Analytical generalization is a question of relating case findings to an existing or emerging theoretical body in order to extend existing theory (Yin 1989) or develop new theory (Strauss & Corbin, 1990). "If two or more cases are shown to support the same theory, replication may be claimed. The empirical results may be considered yet more potent if two or more cases support the same theory but do not support an equally plausible, rival theory" (Yin, 1989, p. 38). This is ensured through the iterative testing of questions and propositions described above, and the integration of research findings with theoretical frameworks. In the present project this process was about generalising the case findings to the theoretical discussion of emerging new production models, relationships and positions in the supplier chain, the core competence strategy, technology and operational synergy and organizational coordination discussed in chapter three and four, and to other literature directly relevant for the emerging concepts. 5.4 RESEARCH METHODOLOGY - SUMMARY In order to develop an appropriate research methodology, several steps and issues must be considered. The first concerns the definition of the research problem. It was argued that this can be seen as an iterative process between matching the real world context (i.e. the managerial situation where the research will be conducted, the researchers directing theoretical knowledge, experience, and personal interests), the overall research questions, and the 8 Available upon request from the researcher. Directly taken field notes, though, are confidential. 9 This practice can be criticised. Silverman argues that 1) respondents might have little interest in a research report and even problems in understanding it, and 2) validation is possible only if the results of the analysis are compatible with the self-image of the respondents. In the present project the first point is less relevant than the second (management research deals with timely real world managerial problems and the theorisation is limited compared to the sociological research from which this argument comes). The second is worth further consideration. Possible bias was counter-checked with the original field notes and in one doubtful case the original informant was asked to express himself on the question which permitted to reach the conclusion that the question actually was subject to ideologically contradictory standpoints by the two informants. Generally speaking, however, little divergence between original field notes and informants' reviews occurred. 169 philosophical position of the researcher. When a coherent approach has been reached between these three items, the research objective can be defined. It is to a large extent the overall research questions that determine whether an exploratory, descriptive, or explanatory study will be conducted. Once the research problem has been defined, the research object or unit of analysis must be identified. Possible units of analysis are compared to the research problem in order to find a match, if this is not possible the research problem must be redefined. Development of an appropriate research strategy is the next step in the development of a research methodology. Whether, for example, a survey or a case study strategy will be chosen depends on the match between problem definition -above all the nature of the research questions- and the unit of analysis. Case studies are a preferred strategy when questions deal with complex links where the context of investigated phenomena is important. The last step in the process of methodological choice is to determine the relation to theory that the research will adopt. Generally speaking this opposes an inductive and a deductive logic. A first indication of how this choice will be made already occurs when considering the real world context. If the existing body of literature is weak, development of new theory -an inductive approach- is essential. It is the role that existing theory plays that differs in a deductive and an inductive approach. In the former, hypotheses developed from gaps in existing theory are tested to modify or extend the same theory from which the hypotheses originated. In the latter, emerging concepts forming a new theoretical framework are compared to existing theories proving themselves to be relevant to the data gathered. Once the above process is completed a methodological choice in terms of a qualitative deductive or inductive logic, or a quantitative deductive or inductive logic is made. Then appropriate data collection and analysis methods can be employed. During the development and analysis of the above process the position in the present research was defined. Concerning the research problem the research questions emerging from previous experience, orienting literature studies, and a personal interest in examining a managerial situation from 'inside' were mainly of a 'how' and exploratory 'what' kind. The objective of the research was to acquire as wide an understanding as possible of lean product development in expert supplier firms, thus an exploratory, and to some extent, descriptive study. The unit of analysis that could respond to this objective was identified to be the product development process in each studied company. This means that the research was focused on operational development work and identified how managerial techniques and systems, internal cooperation and coordination between different concerned actors, and integration with customers intervened and influenced product development work in a systemic process perspective. 170 The research strategy corresponding best to the defined problem and the unit of analysis is that of case studies. The problem concerns complex links between phenomena, and the investigator's control over behavioural events is limited. Finally an inductive relation to theory was adopted because existing literature was found to deal little with a supplier perspective on the operational aspects of integrated component development. The second part of this chapter treated the research design which comprises a definition of data collection and analysis methods, and a description of how evidence is interpreted in order to provide answers to the overall research questions. After a presentation of the selection criteria for identifying potential case studies and interview companies the two research topics and their respectively four and three themes for data collection were presented. These constitute the underpinning logic for the data collection, and provide the link between overall research questions and field work activities. As a function of the literature analysis in chapter three and four, detailed questions were developed and presented in a topic guide under each one of the themes. Pilot interviews allowed for a refinement of the topic guide and collection of valuable input concerning relevant phenomena. Among the messages from this phase the existence of multiple customer relationships in one and the same supplier firm, the confirmation of increased design responsibilities, the application of main lean techniques, the enlargement of product technology to propose more complete functions, and an increased operational cooperation with customers can be noted. Concerning data analysis, it followed the method developed by Strauss & Corbin (1990). Field notes were firstly analysed according to the open coding technique which consists of labelling events, categorizing or grouping similar events together, and developing these categories. Then different concepts were classified in a relational structure that Strauss & Corbin (1990) name the paradigm model. Reliability and validity, finally, were managed through the application of classical means and methods developed in the research methodology literature. 6 THE LEAN PRODUCT DEVELOPMENT CONTEXT This chapter is the first of a series of six empirical chapters. It deals with the first objective in the present research project; to provide an example of the state-of-the-art in lean product development within automobile supplier firms. The presentation that will follow in this chapter is mainly based on interview data coming from the semi-structured interviews with General 171 Managers and Product Development Managers in the eight studied companies. It is also completed through interviews with Quality Managers, Purchasing Managers and Sales Managers in the two case study companies. The discussion is based on the four themes under which data were collected, i.e. the place and the role of the firm in the supply chain, change and development tendencies in the product development process, driving forces for change, and use and perception of lean techniques. In chapters 7 to 11, the empirical discussion continues with a focus on the development process. These chapters are based on the two case studies and completed with data from the last four interviews with product development managers. 6.1 THE PLACE AND ROLE OF AN EXPERT SUPPLIER IN THE SUPPLY CHAIN The central preoccupation of managers in the studied supplier firms was the customer relationship. Customer satisfaction, customer initiated change in the product development process, and customer orientation in the product development strategy were the most frequently quoted features when managers were asked to describe the difference between traditional and new 'lean' product development. Thus, customer relationships are explicitly and directly addressed as a central phenomenon in the new product development framework. Customer relations were also addressed indirectly through a number of other issues that can be related to the customer through a relationship analysis of the data according to the paradigm model. Examples of such issues were: • The need to reduce costs and lead time, and improve quality (driving forces for the development of a particular set of customer relationships), • The development of a design organization based on project management and the application of different tools and methods for improving the productivity in the design work10 (a context in which the customer relationship takes place), • The question of how to manage confidentiality between different projects running for different customers (an intervening condition shaping the design of relationships and the internal management of and organization for parallel projects), and • The question of the supplier's global product development strategy and organizational design for improving the offer to the customer and the profitability of the supplier firm (action and interaction strategies). 10 Examples of measurements of productivity in design in the studied companies were: the percentage of design studies transformed into new products, the number of prototypes realised per new product, the development lead time from first draft to first delivery of a final product, the number of new products launched per year in relation to the number of employees in the design office. 172 For the above reasons, the customer relationship emerged as the central phenomenon in the description of the lean product development context in the supplier firm. In the following paragraphs, the different relationships that were found in the studied companies will be described. All the supplier companies studied confirmed the three global tendencies previously identified in the literature (c.f. section 4.1.2.), i.e. an important reduction of the number of direct suppliers, an increased design responsibility for the remaining ones, and tighter coordination and control of the restructured supply chain - with the objective of integrated product development. These features had been playing the role and continued to play the role of driving forces leading to efforts for increased integration in the product development process. The integration comprised both an external pool -the relationships to customers- and an internal pool concerning integration between different functions and actors participating in the product development cycle within the supplier firm. The tier model presented in chapter 4 was confirmed as the reference for the current industrial organisation in the automotive production chain. However, the reality of the studied suppliers' relational structures was far more complex than this model indicates. In order to understand the place and role, and the related strategies, for a supplier in the tier context, supply must be distinguished in terms of components on the one hand and know-how or development intelligence on the other. Thus, when discussing customer relationships and the related strategies during interviews, Lamming's (1993) model of direct and/or indirect suppliers of components and/or development intelligence discussed in section 4.1.2 proved its relevance. In accordance with the definition of the companies that were studied, the only customer relationship that was sure to be found was that of direct supply of components, measured by the turnover generated by the selected suppliers in the first tier11 . At the most overall level discussing relationships only in terms of first tier, second tier, third tier and lower- the companies studied all had an explicit strategy of securing a position in the first tier. However, managers immediately related this objective to the distinction between supply of components and development intelligence. In fact, it was the latter that was central for their first tier strategy. The different relationships that were identified through the interviews are illustrated in figure 15. It summarises the most significant situations that General Managers and Product Development Managers in the studied companies described. 11 C.f. chapter two. 173 Carmaker 1 Carmaker 2 Carmaker 3 Component S2 System A System supplier A Component S1 . Component S3 Direct arm'slength supply Direct expert supply Indirect expert supply with triangulation Components Expert Supplier Development intelligence S Figure 15. Different customer relationships facing an expert supplier. As figure 15 illustrates, one and the same supplier must be prepared to manage a variety of situations in terms of customer relationships. Hence, parallel to a typology based on different kind of suppliers (e.g. Lamming, 1993), it is relevant to define a typology relative to different customer relationships facing one and the same supplier. The supplier in question, 'S' has three customers among the car manufacturers, and two among the system suppliers. The relationships can vary from permanent, close and personalised contacts between design engineers and technicians all along the development process to a 'supplier on catalogue' negotiation-situation, taking place between a sales assistant in the supplier company and a purchasing assistant in the customer company and that has evolved very little over time. In addition, the supplier typically answers to requests and/or has pilot projects going with another two or three carmakers or system suppliers. In the following the three identified relationships will be described in more detail. 6.1.1 Direct Arm's-Length Supply This situation is illustrated by the relationship to carmaker 2 in the model, see figure 16. 174 Carmaker 2 Component S2 Direct arm'slength supply Components Expert Supplier S Figure 16 . Direct arm's-length supply. Here, 'S' is a first-tier supplier but without interchange of intelligence. In this case, carmaker 2 keeps the design of system A vertically integrated, and buys a component S2 'off-the-shelf' from supplier S. At a given moment, this component has similar performance compared to the component S1 developed in collaboration with a system supplier and a carmaker (c.f. figure 15). However, it is probable that problems will occur over a longer period, if and when the integrated development effort between 'S', 'A' and carmaker 1 results in a modified component S1prim , specific for the needs of carmaker 1. In fact, supplying and buying off-the-shelf parts was characterized as risky both for suppliers and customers by the interviewed managers. On the one hand, if 'S' privileges product innovation and has a strong bargaining power, it can stop manufacturing the slightly outdated component S2 to gain capacity for new products. On the other hand, if carmaker 2 either finds that the component no longer meets his expectations (if 'S' replaces S2 with S2 prim -a new off-the-shelf component derived from S1 prim ) or that the component is outdated (if 'S' continues to offer S2, while competitors come up with more sophisticated solutions), carmaker 2 might switch supplier instantly, giving 'S' little possibility to maintain its business due to the fact that no intelligence links exist to the carmaker, something that also offers limited possibilities for suppliers to develop their product design competence. In the case where carmaker 2 continues to buy the component S2 and 'S' continues to manufacture it, carmaker 2 will suffer from competitive loss due to outdated performance in the component, compared to carmaker 1. In the relational configuration of direct arm's length supply, the evolution of component performance was described as incremental and related either to the transmission of a different detailed specification for a new component from the customer, or to the proposition of a new component on the part of the supplier. In both cases the lack of continuous 175 improvement, evolution, and adaptation of component systems might lead to a slack in competitiveness when the customer uses components that will quickly be slightly outdated in comparison to those of a competitor that works in a continuous improvement mode. It is also possible that carmaker '2' will follow carmaker '1' (a presumed competitor) and outsource the system where S2 takes part. What would happen to supplier 'S' in that case is far from clear. Whether 'S' has the opportunity to benefit from a triangulation situation depends very much on the chosen system supplier. Imagine that it is A. If the triangulation between 'A', 'S' and carmaker 1 has been successful - so that 'A' wishes to continue to work with 'S', the bargaining power of system supplier A and the evolution of the new relationship between him and the carmaker 2 will determine the future for 'S'. If carmaker 2 chooses one of A's competitors that has another supplier for the component, there is little chance that 'S' will keep the business at all, as it is only another supplier among others, without exchange of development intelligence, in carmaker 2's supply base. In spite of the above mentioned risks and problems, the majority of the interviewed managers indicated that they had long-term and profitable relations of this kind. In fact, the direct arm's length relationship seemed neither to be only a remaining part of earlier arm's-length supplier relationships, nor did such relationships seem to prevent a given supplier from pursuing a strategy of advanced product design. Instead, they were found to be a means for generating revenues from large series production at very low costs (in principle only direct production and material cost). Very probably, these relationships will remain, but the customer will to a lesser extent be carmakers and to a larger extent system suppliers. In the smallest companies that were studied, products delivered in arm's length 176 relationships had been, and continued to be, the main source of funding for investing in product development activities such as research activities in conjunction with specialized laboratories, investment in computer aided design, testing equipment and prototype manufacturing, recruitment of design engineers and technicians, employee training, etc. 6.1.2 Indirect Expert Supply What can be called indirect expert supply is illustrated by the relationship to carmaker 1, see figure 17. Carmaker 1 System A System supplier A . Component S1 Indirect expert supply with triangulation Components Expert Supplier Development intelligence S Figure 17. Indirect expert supply. For the component S1, 'S' is a second-tier supplier to carmaker 1 (who buys this component through the system A). However, in relation to this carmaker, 'S' is a direct supplier in terms of development intelligence, as the system A is designed via a close collaboration between the carmaker, the system supplier and 'S'. This practice, which was found to be very common in reality, but that is not often discussed in theory, can be called triangulation. It is mentioned by Lamming (1993) who illustrates it with a door lock supplier that designs and develops the components in close cooperation with the assembler but delivers to a first-tier door supplier. However, Lamming (1993) does not further develop the managerial aspects of such practice. In the case where the indirect supplier delivers a core technology or a specific know-how, this practice seems necessary from the carmakers' perspective in order to optimise component design in terms of functionality, manufacturability, and ease of assembly - within the framework 177 of cost targets. The present research confirms that triangulation is practised not only with suppliers of high-tech products such as advanced materials or micro-electronics, but in a much wider context involving smaller suppliers. For suppliers that have opted to develop their capabilities in engineering design, triangulation means an opportunity to keep direct contact with the carmakers through an expertise in product technology problem solving and R&D, even if the component itself will be included in a larger system. This direct contact is crucial for any supplier that wants to remain an expert. In spite of all the talk about outsourcing component design and relying on supplier development, the carmaker was confirmed as the orchestrator of the entire production process. The carmaker has still the ultimate decision power concerning different solutions and is the one that specifies functional, cost and quality targets. In the studied supplier companies, that had all gone through the revitalisation of the sector during the last decade, managers confirmed that even if they had their own ideas for improving work methods, little could have been done without the initiative coming from their final customer - the carmakers. In short, the direct contact remains indispensable for keeping up with technology, managerial competencies, and the latest development in the industry. 6.1.3 Direct Expert Supply This situation is illustrated by the relationship to carmaker 3 in the model, see figure 18. Carmaker 3 Component S3 Direct armslength supply Direct expert supply Components Expert Supplier Development intelligence S Figure 18. Direct expert supply In relation to carmaker 3, 'S' is a first-tier supplier both in terms of physical supply and development intelligence for a component S3, that is not yet integrated by system suppliers. In 178 opposition to triangulation, this situation might comprise substantial competition between complementary expert suppliers in terms of tier position. These are two or more suppliers that directly provide the carmaker with components that are technically interdependent, but where this interdependency is managed by the carmaker. The interviewed managers explained that competition between these suppliers occurs because the carmaker doesn't always want to keep two complementary experts in his direct supply base for cost or coordination reasons. There is a trade-off between fully exploiting technology synergy through an integration of the two experts in the engineering process, and the need for reducing the number of suppliers for rationalisation reasons. The 'tier dilemma', i.e. the question of which one of two complementary expert suppliers will stay in direct contact with the carmaker is therefore constantly present in this situation. In order to win this 'combat', suppliers need to make important investments in technology and know-how to develop their design capabilities, something that is difficult to finance for a medium-sized supplier company. Many suppliers also tried to enlarge their offer and supply small sub systems in which their core components were integrated. In this context, however, there might be a risk with taking on non-core assembly activities at the expense of developing engineering capabilities. Another situation occurs if the trend moves towards outsourcing of the system where component S3 takes part (i.e. carmaker 3 will turn to system supplier A, who enlarges his offer, or to one of his competitors). 'S' will then have substantial advantages over a competitor on the same product due to the fact that 'S' is very aligned with carmaker 3 from the former direct expert situation (this of course would be where the partnership between carmaker 3 and 'S' is satisfactory). Even if the carmaker chooses another system supplier than A it is likely that the carmaker will impose the choice of 'S' on the system supplier that he chooses. 'S' will then benefit from triangulation in this new structure. 6.1.4 Place and Role in the Supply Chain - Summary As has just been argued, the relational situation for an expert supplier can be quite complex. The above analysis identifies a typology of different situations that need different capabilities and competencies to be met, and that need to be differently managed. The supplier must be able to develop a structure that allows to respond fully and with flexibility to the demands in these different situations. This fact gives a new dimension to the complexity of managing a medium sized expert supplier firm which has not been explored much in the existing literature. Table 13 summarizes the characteristics of the three different relationships relative to the main points of managing buyer-supplier relationships in product design discussed in chapter four: the situation in the tier structure, the product development strategy, product technology and 179 technology synergy, operational coordination, and communication in engineering. It also summarizes the main difficulties that suppliers have to manage in these different relationships. Direct arm's length supply Situation in tier- Unstable. Little chance of structure keeping business if outsourcing Indirect expert supply with Triangulation Direct expert supply Relatively stable. Unstable - tier dilemma. Outsourcing logic applied. Great chance of keeping business if outsourcing. Product development strategy Short-term profit generation. Focus on core technology. Enlarging activities through innovative design. Technology synergy Minor High between three actors. High between two actors tier dilemma. Evolution of product technology Incremental, noncoordinated improvement of component performance. "Slack" in competitiveness. Continuous improvement of component performance, importance of cooperation with interfacing suppliers. Operational coordination Carmaker - passive, check- Carmaker - active. up on performance System supplier - active. indicators. Carmaker - active Minor Frequent and rich information exchange. Strong alignment to and anticipation of customer needs. Communication in engineering Main difficulties Limited possibilities of competence development in design. Strong risk of instant elimination from the supplier base. Frequent and rich information exchange. Strong alignment to and anticipation of customer needs. Continuous improvement of component performance, system integration. Develop cooperation with Need for investments in interfacing suppliers. technology and knowhow. Risk with taking on Even stronger cost and assembly at the expense lead-time pressure. of engineering. Table 13. Characteristics of three different customer relationships. From the analysis and summary of interview data, triangulation seems to be the best answer to the need for simultaneously balancing the following three objectives shared among carmakers (c.f. section 4.1): reducing the number of direct suppliers from which the carmaker buys components, keeping control of all the technical systems ni the vehicle, and optimising the technical, functional, and styling interfaces between all those elements and systems that go into a vehicle. However, a drawback quoted by the interviewed managers was that cost and leadtime pressure are even more severe than in relation to the carmakers. This is because the commercial context is "downgraded" one step; the system suppliers' need for margins both in revenues and time will have to be partly supported by the expert supplier. Furthermore, the analysis confirms that while triangulation is automatically associated with an important contribution of the supplier's development intelligence to the product design project, 180 first tier supply as initially defined, is not always so. In fact, two modes of first tier component supply were detected; direct supply of components and know-how (called direct expert supply), and direct supply of components with no intelligence links (called direct arm's length supply). The types of customer relationships and the nature of the products -detailed controlled parts, supplier proprietary parts and black box parts (defined in section 4.1)- are of course related. Black box parts are associated with direct expert supply and triangulation. Detailed control parts are associated with direct or indirect supply without intelligence links, so called arm'slength supply. Supplier proprietary parts can be present in all of the above customer relations. They are often present at the two end pools of a product development effort continuum; they are either commodities purchased in arm's-length relations, or new innovative products independently developed by the supplier and for which the customer has not expressed an explicit need. The investigation of the different customer relationships in the studied expert suppliers revealed two interesting results in relation to the current modelling of the automotive supply chain. The first is the confirmation of the relevance of Lamming's (1993) model of direct and indirect supply of components and/or development intelligence. In relation to this, the data indicate that the core element of a first tier strategy is direct supply of product development intelligence. The second result is that unlike what has been advanced in the literature so far, different kind of relationships can be simultaneously present within one and the same supplier firm. Moreover, it seems to be possible to manage their coexistence over time. More than a one relationship - one supplier situation, the results confirm the existence of multi-relationship suppliers. Lamming's (1993) description of supplier behaviour seems to correspond to the observed reality with the difference that suppliers do not only "manoeuvre from one relationship to another following strategies of enrichment" (p. 207), but develop and enrich several different relational modes that each have a commercial raison d'être. The study of the place and role in the supply chain also identified two particular phenomena that intervene in the product development process: the specific communication mode of very open and close to real-time (in relation to problem solving) interactions between design engineers and technicians in supplier and customer firms, and the management of the tier dilemma. These will be further analysed in chapters 7-11 dealing explicitly with product development organization and strategies. The rest of this chapter will present the research findings relative to the framework developed in section 5.3.2: 181 • The driving forces for change, where interview questions were focused both on changes perceived during the last five years and changes perceived at the moment of the research intervention; • The change and development tendencies in the product development process, focusing on the reactions to the driving forces and anticipated evolution in the industry; and • Use and perception of lean techniques where the application of and satisfaction with specific measurements such as flexibility, flat organizations, continuous improvement, and group work were examined. Finally, specific problems related to the introduction of lean practice will be discussed. 6.2 DRIVING FORCES FOR CHANGE The three classical parameters of quality, cost, and lead-time (c.f. figure 19) were quoted as the main driving forces for change by the interviewed managers. Quality Cost Lead-Time Figure 19. The 'Golden Triangle' of Quality, Cost, and Lead-Time. As discussed in section 4.1.2 a need for important delivery and development lead-time reductions, quality improvements, and cost reduction has been created by the globalization of competition and the emergence of Japanese cars with superior performance in these criteria. How these factors were perceived and managed in the studied companies will be analysed below. 6.2.1 Just-in -Time - JIT JIT was mentioned as the first intervening driving force related to lean production met by the studied supplier firms. The focus of JIT is on the flow process in operations management at the manufacturing level. JIT means manufacturing from source direct to the customer without intermediate stock, an external orientation to scheduling and, as a consequence, a different 182 adjustment of the system capacity (Wild, 1989). Early restructuring work in the companies therefore focused on the creation of flow groups, increased flexibility concerning both labour and machines, reduction of set-up time, and the creation of production teams. Following a classical scheme in lean production implementation (c.f. Krafcik, 1988), the reduction of intermediate and final stocks revealed productivity problems related to machine set-up and maintenance, and worker skills. Some companies also admitted problems with too much rectification work on manufactured products undertaken in the traditional system in order to meet defined quality requirements. Even though all of the interviewed managers expressed satisfaction with the measures undertaken, just-in-time manufacturing -balancing the just-in-time delivery currently imposed by the majority of the customers- was still not possible to achieve completely. This rejoins the findings of Sako et al (1994) discussed in section 1.4.2. The main difficulties remain in the case of the production of large series, normally concerning the simplest components manufactured by a company. In two companies this problem was studied in greater depth. If, for example, a customer orders batches of 100 components on a D minus one or two basis (i.e. one or two days before the delivery date), the suppliers' cost analysis of set-up cost, trimming cost12 , direct production cost and stocking cost determined whether a just-in-time manufacturing or a manufacturing to a final stock would be the most economic. The two General Managers with whom this question was discussed explained that such an analysis often pointed in favour of manufacturing to a final stock, mainly due to the small unit cost of the components in question and the relative inflexibility of the equipment in place which led to high set-up and trimming costs. Normally, this remained the case even after application of SMED techniques. The remaining solution would then be to invest in more flexible equipment. In that case, the cost of the investment and the related training of the workforce on the new equipment had to be offset by the reduction of the stock, set-up and trimming costs. However, data for the last two was very difficult for managers to anticipate, especially if the investment concerned tailor-made equipment. One of the companies, located in an urban and quite expensive area, identified an additional productivity gain in eliminating the stocks between production and just-in-time delivery, namely the question of how the released storage room would be used. If it could be used for value creating activities that otherwise would have needed more space, the cost estimations changed, but again such an estimation was extremely complex and therefore difficult to make with a minimum demand for accuracy. 12 Cost of time and lost products related to the trimming period before a machine produces the normal quality level after a start-up. This period can approach zero for some equipment while being considerable for other equipment; it must be analysed case by case, machine by machine. 183 It must be underlined that the problems discussed above concerned the final stock of the ready-for-delivery product. All the studied companies considered themselves successful in the elimination of intermediate stocks. Generally speaking, however, the interviewed managers in the eight companies were quite satisfied with the current application of JIT in relation to their customers. They estimated that their JIT had improved, not only thanks to an improvement in the experience curve, but thanks to an improvement in customers' scheduling. Moreover, JIT meant a complete restructuring of the production flow that was also an opportunity to renew and perfect the production tool. The main remaining problems were in the relation between just in time delivery and just in time manufacturing, as discussed above. In order to ensure the former, stocks of 2-3 weeks' production still existed for some products. In a way this can be seen as an indication of low performance, but, at the time, the concerned managers did not see any quick-fix solution to change the situation for at least two reasons. Firstly, schedules and real orders are still two different things, only under the theoretical hypothesis that customers' scheduling would be 100% realized would there be no final stocks to support for suppliers. Secondly, it is not always clear that an investment in more flexible equipment will pay off - the stocking cost is simply not always that high. 6.2.2 Quality Management Almost simultaneously to the increased pressure for JIT, quality requirements concerning the final quality of the manufactured products have become more and more severe. Quality assurance -with suppression of reception control in the customer firm- has become a rule. This focus on quality has been gradually extended to comprise the organization of the supplier company (especially the organization of the quality function and the product development activities), the traceability and analysis of quality problems, the anticipation of failures, the measurement and collection of data concerning quality indicators, the establishment of global performance guidelines for Human Resource Management (especially concerning motivation and communication), specifications of how to manage subcontractors and ensure their reliability, and productivity indicators reflecting the long-term evolution of the global performance of a supplier company13 . This has led to the establishment of formal quality standards on the part of the carmakers and some system suppliers. The objective with these quality standards is to analyse, formalize and optimize the structure of suppliers in terms of organization, work procedures, single activities and transversal processes. They oblige suppliers to explicitly specify their different activities in procedures comprising flow charts and 13 EAQ, PSA-Renault quality standard. The listed items are a selection of the areas that the quality standards treat. 184 detailed instructions of who, when, and how for each step and phase in an activity or process. As a consequence, the quality standards link individual and collective knowledge and knowhow and make a company's operating procedures explicit. After having received the certification relative to a specific quality standard, the supplier is contractually bound to respect these procedures. These kinds of contractual quality standards -elaborated by each carmaker and/or system supplier- were still the dominating references in the studied companies in spite of the emergence of the general ISO 900014 quality standards. Chronologically speaking, the studied automotive suppliers were first certificated according to a sector standard, for example Ford's Q1, then they tried to apply for an ISO 9000 label (normally 9001) mostly in order to satisfy non-automotive customers for whom auto standards would not be familiar. The importance of quality standards has increased over the years and they played a central role in the tactical and operational preoccupations in all functions in the studied supplier firms. There is a simple explanation to this; the conformity between the quality standards and the reality is regularly assessed through severe auditing procedures, and the correction of detected failures in order to remain as a supplier to the auditing customer is given very little time. As a logical consequence of this, the quality standards were quoted by the interviewed managers as the most important current driving force for change. However, significant differences in the application of the different standards were observed in the studied companies (which were all audited and accepted by different carmakers and system suppliers at the moment of the research intervention). One of the interview companies, that still has the majority of its business in direct arm's length relationships to its carmaker customers, had just started the reorganization process of the design department towards project groups. In the two case study companies, that have the majority of their business either in direct expert or indirect expert relationships, COF presented the most driven form of project organization (see appendix 5), while COR, after a successful integration of the product and process engineering functions still had to work on the relationship with running production. These differences indicate that there is a certain degree of flexibility in the auditing procedure in terms of an adaptation to the role that each supplier has for the customer. 14 International Standardization Organization. The three major standards are: ISO 9001 Quality systems - model for quality assurance in design, development, production, installation and servicing. ISO 9002 Quality systems - model for quality assurance in production, installation and servicing. ISO 9003 Quality systems - model for quality assurance in final inspection and testing. 185 The signals in terms of quality requirements coming from different customers were also found to be different. For example, in compared to PSA-Renault, Ford's Q1 standard emphasizes progress indicators to a al rger extent. As a consequence, such indicators received more managerial attention at COR (that was audited by Ford) than at COF that was audited in France only by PSA-Renault among the carmakers. As for JIT, a main priority after the application of quality standards in customers relationships was to develop similar quality management frameworks with the studied companies' own suppliers. In relation to the issues of how buyers and suppliers are linked together (c.f. section 3.3.2), and what integrated management systems can be used to build trust and goal congruence between companies (c.f. section 4.3.3), the research findings indicate that quality standards constitute the main formal link between customers and suppliers. Instead of cross equity and inter-company career paths practised in Japan, the French and European carmakers seem to rely on this contractual guarantee of a certain organization and managerial structure. The cognitive dimension of the quality standards make them particularly interesting in this context. As Duymedjian (1996) argues, the ISO 9000 standards (that are the subject of his work, but contractual automotive standards are exactly the same in this respect) invite managers to focus on problems related to the expression (as discussed in section 4.4.2 this is central for ensuring efficient communication), the memorizing, and the enrichment of collective knowledge and know-how. 6.2.3 Price Pressure An example will illustrate the daily constraints for suppliers concerning pricing of components, a problem that preoccupied all of the interviewed managers. The example concerns the relation to an arm's-length customer and comes from one of the case study companies and a meeting held by the sales engineer with the general manager, the product development manager, the project manager for the project in question, and the quality manager. For the customer, the ideal situation was to lower the price without modification of the component's specification. It was stated in the contract that prices had to decrease by 5% per year following productivity gains in the supplier's production process. However, for certain components, like the one in question, it was very difficult for the managers and technicians concerned to know whether a real productivity improvement had taken place. Within the framework of links that tied this supplier to its customers, productivity targets were fixed for certain periods of time. This particular case concerned price reductions on components that theoretically should follow upon productivity gains in the process of realising the product. The supplier in question had elaborated a proposition for modifying a component 186 to achieve a more significant price reduction. The reason for this was that the company had problems in realising a productivity gain on the component due to difficulties in measuring the cost productivity for single components. Increases in raw material costs had also altered the cost structure of the component and made a price decrease almost impossible while maintaining an identical product. For these reasons it was decided to redesign the component and change both functional parameters and material. For the supplier company, a price decrease corresponding to the productivity gains was not enough. What was important was to provide modifications that also increased the margins in order to free resources needed for process investments and new product development projects. However, the modifications in question had to be accepted by the customer. According to the rules that conditioned the relationship the supplier had the right to propose a modification in the product (and not only in the process) in order to reduce component price, but there was a strict procedure to follow. This procedure for validating such modifications was quite long and complex. It comprised the proposition of the modification, the transmission of related blueprints, and a description of the related production process to validate quality requirements and the new cost structure. This proposition then had to be validated in the customer's purchasing department with which the supplier was in contact. Then it had to be transmitted internally to the design department which then would accept or reject it. If the modification resulted in modifications in a cascade of inter-connected components the chance was very small that the modification would be accepted. The ideal situation from the carmaker's perspective was a price decrease without modification of the product, i.e. only due to process productivity gains. The sales price for the component in the example was around 30 Francs. The yearly price reduction of 5% was, thus, 1.50 Francs. In order for the customer to accept a modification, the gain had to be much more significant, otherwise the cost of the change procedure would outweigh it. In this case the supplier had to make a very careful estimation of the gain so that it left room both for a significant price decrease for the customer and an increase in the supplier's margin. The customer's purchasing office did not normally send a modification request further if the proposed price reduction was too small. The criterion used to decide upon this depended on many factors such as the quantity ordered per year, the position in the life cycle of the component, and the competing supplier offers that the customer could choose among. As a rule of thumb a ratio of 30% of the initial cost was often mentioned. The example also reveals that the more standardized a component, the greater the difficulty of increasing the margins or even keeping the same margins after price reductions. In this example one sees that the bargaining power was weak for the supplier, there was an insecurity as to how the customer would react, and dependence because the supplier would lose money if his 187 actions were not accepted. In this customer's strategy, there seems to be a trade-off between demanding price reductions and refusing modifications that sometimes might be necessary in order to achieve these price reductions. This can have a negative impact on innovativeness and on the development of suppliers' design capabilities. Finally the example shows that even if a supplier is the sole source to a carmaker for a specific component it has to stick to very formal procedures. In the case discussed above there was not a very clear interface with the customer. There was a core buyer in Germany, with whom future projects were discussed, then local purchasing agents intervened and there was no way of informally discussing technical solutions. In this case the supplier company was a first-tier supplier with very much the traditional kind of relationship. It must be asked if the problems discussed above were due to the nature of the component. Maybe the customer thought it unnecessary to develop integrated design for this particular component, or to the supplier; maybe the customer was working with someone else or planned to buy the component through a system supplier. It might also be a general practice of this carmaker not to develop too close relationships with its suppliers. The research perspective does not allow for a quick answer to these questions. It reflects the situation of a supplier who has little idea of how the customer might think and react in a particular context. Only a change towards a more open relationship could make it easier for the supplier to optimize their action strategies. 6.3 CHANGE AND DEVELOPMENT TENDENCIES IN THE PRODUCT DEVELOPMENT PROCESS In relation to the question concerning the most important changes in the customer relations, General Managers and Product Development Managers were unanimous that after JIT, quality, and cost, increased product development responsibility is the fourth heavy tendency, and that in this area, the situation is far from being stabilised. As far as general tendencies in the automotive supply industry are concerned, the continuation of the restructuring in terms of reduction of direct suppliers, transfer of design activities to suppliers and evolution of integrated structures were confirmed in interviews. Concerning the first point, the interviewed managers unanimously thought that the only direct suppliers besides system suppliers to remain would be the specialists. The rationale for this is that the design expertise for a specific component retained by the expert supplier is considered, either by the carmaker or the system supplier responsible for the system where the component will be fitted, as indispensable for an 188 optimal component solution and hence for a system's performance. Three different reasons for this were given by interviewees: • The technical functionality of the component; • The material properties of the component; • The interfacing qualities of the component. The first point concerns the product technology in terms of the function of the system. If a component represents the core technology in a system, which means that its performance is directly visible to the final user, the carmaker will explicitly maintain a direct relation. The second point deals with product technology in terms of the material. The use of new materials often leads to a complete reconfiguration both of a product's functional performance and the production process. This, in turn, affects the product and process technologies of related components . Moreover, the potential for innovation is often important when new materials are used. These factors taken together lead carmakers to preserve direct design relationships with suppliers responsible for components where material and/or process technology often changes. A typical example is technical plastic components. The third case where carmakers like to preserve a direct influence over the design is in components that are important interface elements between several different parts of a system. The main reason is that the evolution of such an interface component will have important chain effects in a functional system, effects that might be difficult to master. Without a direct contact, and an integrated development with suppliers of interface elements, there is a risk for increased coordination cost in product development. Examples of components of this kind are electric cables, fastening devices, and high precision engine components. The points discussed above indicate, thus, that there can be several logical reasons for carmakers to continue to have direct contacts not only with system suppliers, but also with specialized suppliers supplying products of the kind identified above. The interviewed managers thought that these remaining experts would take more and more responsibility for the engineering of a vehicle. The design function would begin to live a life more and more of its own - in terms of independence of company borders. More than a company identity, technical centres would have a project identity (c.f. e.g. 'on floor two, 200 engineers and technicians from 50 different companies work on the development of part X of vehicle Y'). For COF, for example, this trend was becoming visible through the fact that carmakers were beginning to entrust the development of fastening in entire parts of vehicles with the company, for example all fastening devices in the engine compartment or again in the 189 interior. This was leading COF to integrate their design work with an important number of suppliers in different tiers for the development of optimal solutions. Another identified trend was globalization. For the studied expert suppliers a global presence was an important goal that was realized in different ways, i.e. through own divisions, joint ventures, alliances or attachment to a holding company. However the global presence for these companies did not go beyond the dispatching of carmakers' engineering centres, and these tend to be more concentrated than the manufacturing facilities. For an expert supplier it will not be necessary to be installed close to each and every production site. This is only necessary for logistic reasons in the case of system suppliers supplying large and bulky components such as wheels and tyres, front-ends, dashboards, fuel tanks, exhaust systems, seats, etc. A third tendency that preoccupied management in the studied firms was the evolution related to technological progress. The couple material technology and process technology seemed to become a more and more significant area of change. Components and parts in aluminium, composites, ceramics and ultra-thin sheet metal are currently beginning to find profitable applications. One executive mentioned examples where value engineering simulations of a new component based on a new material realized in a new process would reduce direct manufacturing cost (material plus throughput time) by as much as 40%. Especially in plastics the degree of complexity that can be designed into a component is directly dependent on the manufacturing process. A frequent trade-off between the reduction of parts improving quality, facilitating assembly, and reducing coordination cost of several suppliers on the one hand and heavy investments in complex moulds for realizing a reduced number of more complete components on the other, could often be balanced through the application of new process technology. When discussing future strategic orientations for product development with General Managers and Product Development Managers as a function of the change tendencies outlined above, five action strategies were identified from the semi-structured interviews, see table 14. Action strategy Objectives Promoting learning at Building distinctive an individual and capabilities. organizational level Actions undertaken or planned On the job training for individuals and groups. Promoting collaboration and group working. 190 Difficulties/Needs Determining appropriate support structures. How to realize an organizational memory? Identification of and Reinforce the company's concentration on core expertise. Multiply the capabilities applications of core technologies and skills. Entrust more work on peripheral technologies and activities with suppliers. Assessment of transversal processes in order to redefine the global offer of the company. Vulnerability if facing rapid technological or structural shifts in the industry. Development of a Fight against waste and culture of continuous lack of rigour in the improvement organization of daily work. Formalizing best practices to avoid repetition of errors. Train operational staff and technicians in related paperwork and rigour of plan, do, check, action. Make procedures and instructions an active aid in operational work and not only a control device piled up in a closet. Increased customer orientation Align the company to Rethink and simplify the customer needs, improve customer interfaces. anticipation of customers' manoeuvres. Change of professional roles, above all those of sales agents and design engineers and technicians. Improvement of the capacity of integrating change Improve reactivity in the organization. Speed up learning and unlearning cycles. Unlearning of current customs is more difficult to realize than learning of new ones. Improvement of the general competence level in the design function. Increased customer orientation. Table 14. Five action strategies in product development management for responding to change. Behind each action strategy are specific objectives related to the product development capacity and capability that the studied suppliers were developing. The table also summarizes planned or undertaken actions and the related difficulties or needs for the different factors. When comparing the research data with the discussion of the partnership concept in chapter four, it seems, in fact, like a real shift in the nature of buyer-supplier relationships is beginning to occur. As will be further discussed later, the evolution is rapid and significant changes have occurred during the period of the research. All studied companies have developed relationships to their customers that mark a breaking point and a change in the state of mind compared to earlier practice. More of long-term and, as it seemed, permanent relations were being developed and General Managers stated that even though the carmakers have become more demanding they have also become much more reasonable. The new type of relations have also obliged organizational restructuring of the involved firms. Particularly important was the project organization in product development allowing for direct contacts between design engineers and technicians in different firms. Concerning the criterion of commonly negotiated objectives, there was still quite a way to go. The main conflicts seem to occur in the domains of pricing -as discussed above- and confidentiality in new development projects. 191 6.4 USE AND PERCEPTION OF LEAN TECHNIQUES Specific questions concerning lean techniques as identified in section 3.3, table 9 were asked during interviews with general managers and product development managers (c.f. detailed questions in appendix 2). The following will present the information gathered and relate the data to the previously analysed literature. 6.4.1 The Design Organization Short information paths and a flat organization characterised the studied companies. Generally speaking, the limited size of them reduced the need for formal group work and facilitated the development of a project organization. Concerning the project organization, the quality standards require one project manager for every design project. However, the studied companies had adapted and optimized the project team structure as a function of real customer needs. Thus, formal teams with participants from product engineering (including prototype makers), process engineering, purchasing, marketing/sales and quality management were set up only for complex design studies and new products. For more current development projects, the management of inter-functionality was entirely entrusted to the project managers. Concerning group work during execution of design work, informal joint problem solving emerging wherever the natural working practice in product development demanded it seemed more optimized and productive than more formalized meetings and team activities. For these reasons, the managerial imperative became the facilitation of such naturally emerging group activities both within the supplier company and at the interface with customers. An in-depth analysis of group work and project management in product development will be undertaken in chapter seven. Concerning the use of information technology and CAD at the interface with customers and other suppliers, there were problems with compatibility between different software used by different customers and of a lack of rapid and complete (all concerned actors) diffusion of design modifications and advancement of the design in individual participating firms. An extension of the use of information technology as a support in the communication between companies also for the exchange of more qualitative information was a need expressed by some of the interviewed Product Development Managers. All of the studied companies made specific efforts to integrate process engineering and manufacturing aspects in the design activities. As will be further analysed in chapters seven and ten, this was realized through integration between design and manufacturing in the framework 192 of a project organization and specific actions taken more punctually in order to promote learning between design and manufacturing personnel. The main advantages from 'a design for manufacturing strategy' discussed in the literature were confirmed in the study: • A reduction of the development lead-time thanks to an important reduction of design modifications discovered late in a traditional sequential process, and • An improvement of the product quality (quality of realization) thanks to simplified designs made easier to manufacture and assemble. 6.4.2 The Production Arrangement Two main areas where a strategy of flexibility was applied were identified. The first concerned flexibility in competencies. In all of the studied companies, 'on the job training' had been provided in order to train workers for a variety of tasks, and integrate quality control and maintenance with more elementary production operations. The second area concerned the production tool where all of the studied companies had applied integrated flow groups at least in their most complex assembly operations in order to balance variations in production flow and reduce lead-time and inventory. This needed polyvalent workers capable of working on several different work stations. In other words a flexible production tool needs flexible workers. According to lean principles, the development of flow cells (or flow groups) had been realized through intensive participation of the workers themselves. Once in operation, workers were responsible for continuously improving and redefining standardized operational instructions. To sum up, workshop layout and worker competencies had been adapted to the driving forces of just-in-time and quality assurance discussed previously, and task enlargement and task enrichment were applied on the shop floor. 6.4.3 Performance Measurements and Career Paths There was a great awareness concerning the importance of continuous improvement activities at the management level in the studied companies. Different formal support existed, for example weekly design meetings where the product development manager met with project managers and design technicians, and procedures for the course of events in projects. The latter corresponded to the quality standards that require detailed instructions for customer need analysis, feasibility reviews, production launch reviews, and reviews of qualification (intervening after pre-series production and complete testing and control of the final product). 193 An explicit objective with these reviews was to identify problems, propose solutions and generalize these experiences to parallel and forthcoming projects, thus realizing continuous improvement. A major difficulty in the context of continuous improvement seemed to be a lack of collective dynamism, i.e. of active participation on the part of technicians in meetings and other team activities. This was true for design technicians, prototype and process engineering technicians - it will be further discussed in chapter seven. Another difficulty was related to the mass of information that circulates in a design project. A collective memory supporting continuous improvement runs the risk of becoming either too fragmented (blueprints, patent data base, product data base, design study data base) or too loaded with information making it impossible to use rapidly. Inter-functional skill acquisition was clearly stated as an objective in all of the studied companies. Project managers were recruited based on their double experience in product and process engineering, and in the case study companies the chief engineers had all a solid production experience. For more junior staff some companies experimented with what can be called integrated product managers. These were engineers or technicians trained to be capable of assuming several different activities in the product development process. Other ways of systematically widening competencies were also found, for example through the co-location of product engineering and process engineering. As design is upstream in relation to manufacturing and as product innovation is largely dependent on the evolution of process technology, the most important learning need occurred in the direction process towards product engineering. Whitney's (1988) idea of adapting the design to an innovative production process demonstrated its relevance here. However, in relation to competence flexibility in product design, managers expressed worries concerning the trade-off between task enlargement and the need for specialization. This will be further analysed in chapter seven. An essential issue for developing continuous improvement and employee participation is the way propositions from employees are managed. The case studies made it very clear that listening to the propositions of employees is not enough, in order to progress in terms of commitment and continuous improvement employees must be informed of the managerial reactions to their propositions. Briefly, they must feel that there is organizational support accompanying efforts made (Schermerhorn et al, 1991). If propositions cannot be satisfied, or are modified at the management level, the reasons for this must be explained and alternative solutions looked for together. Otherwise employees will quickly lose interest in developing action plans and propositions. Some problems where employees lacked information about the reactions and outcomes of their propositions were identified in the studied companies. This strongly demotivated the people in question and striking contrasts were observed within the same company, i.e. 194 prototype testing staff had big difficulties in receiving relevant feed-back on their propositions while design staff were very satisfied with their managerial feed-back. An overall analysis of the problem in question revealed, however, that the issues that some employees were proposing had little relation to the current issues of customer satisfaction. If management had only explained this and oriented the preoccupations of the concerned employees differently, the demotivation problem could have been avoided. 6.4.4 Problems with Lean Techniques Concerning the question of implementing lean techniques as a best practice without considering adaptation to local and company specific context, it seems like the specific lean methods reviewed above had been quite easily integrated in the companies to the satisfaction of both managers and employees. The apprehensions expressed by certain authors (c.f. Berggren, 1994; Boyer & Freyssenet, 1995, Ellegård et al, 1994) that lean production would be implemented 'by force' in the auto industry and without taking into consideration the particularities of different companies and context seem groundless in the light of the present study. Concerning relationships between customers and suppliers, the interviewed managers whose companies all had previous experience in product design, vividly welcomed the rapprochement between design engineers and technicians in supplier and customer firms and the possibility of actively contributing to the evolution of product technology. One General Manager stated that during the last five years, the company had made enormous progress towards what he called a 'common sense way of working'. If left to purchasing agents, lacking all connection to the products, no iterative discussion of different solutions could exist in the design process. The main difficulty lay in some extreme actions taken on the part of customers to achieve price cuts, and, in this price pressure context, difficulties in paying off design studies. The first difficulty has been discussed above. The second is related to a problem with confidentiality during the design of new products and will be further developed here. The interview data indicate that specific attention should be paid to the management of confidentiality in all steps of the product development process. In the contacts with the design department of a specific carmaker, suppliers' development staff must be careful not to expose design solutions elaborated in confidentiality with another customer, or reveal specific technologies that are under development in competing customer firms. Independent suppliers, like the ones studied, work in a multi customer configuration and have to manage this internal confidentiality. It concerns mostly customer specifications, mock-ups and other customer proprietary items. Discrete product ideas, when a patent is registered by the supplier, are the 195 property of the supplier firm and customers can rarely reclaim exclusivity on a technology of this kind. It is only the integrated function consisting of the supplier technology and the customer technology in a specific and unique configuration that can be the object of a confidentiality clause. In any case, the organizational guidelines in, for example, the PSARenault or Ford quality standards oblige a supplier to organize project groups for each development project, something that separates them structurally. The project management structure adopted by the majority of the studied companies was a relatively stable one, with one project manager associated to one main carmaker customer. At COF for example, there were two project managers in the plastic department and two in the metal department, each assigned either to Renault or PSA - which were the main customers. System supplier customers were handled by the project team assigned to the final customer. Projects for other carmakers or system suppliers supplying other carmakers were normally treated separately under the project leadership of other design engineers (whose roles and functions will be treated separately in chapter eight). As the design technicians were attached to a project manager they worked essentially for one carmaker and its affiliated system supplier, but there was no detailed attachment to system suppliers. Moreover, in order to be polyvalent, design technicians regularly switched between the Renault project group to the PSA project group. Another problem, shared by all of the studied suppliers, was related to external confidentiality occurring when suppliers' ideas, product concepts, and technologies developed in the pre-ordering phase were revealed by customers to competing suppliers. These then could duplicate the design and offer very competitive prices due to the fact that they had no development costs to pay off. In the short term, suppliers protected themselves against this in two ways, either by patenting product functions or technologies at a very early stage, or by simplifying blueprints by leaving out for example detailed dimensions and material properties. In the longer term some studied suppliers tried to develop a strategy of making customers development dependent on the supplier. In some cases a severe attitude to customers practising the free competitive bidding on product solutions designed 'for free' by a supplier could pay off. One of the studied companies refused all counter-check consultations15 , i.e. refused to make its expertise available to the customer after such behaviour. Counter-check consultations must always take place, but can be applied in a more or less honest manner. If partnership rules of the game are respected they should be used only to benchmark similar 15 When a supplier is asked by a partnership customer to come up with a cost break-down for the design and/or manufacturing of a specific product. 196 efforts made by different suppliers, for example responses to a product development request, or to compare cost structures and performance criteria. Lamming (1993) proposes an ideal model for balancing and managing problems related to confidentiality. Under the condition that a specific technology is the result of the collaboration between the supplier and the assembler as equals, a supplier can work in depth with several assemblers at once and in each case the outcome will be unique. The basic technology and know-how move forward as a result of the spill-over from one project to another. A collaborative project remains secret until the car is launched, after which, either side is free to exploit the idea (notwithstanding patents, etc.) with other partners. Each assembler can expect the best from the supplier, only technology that is not yet commercialized, i.e. currently under elaboration together with other customers, will remain secret. Unfortunately, it seems clear from the data that the situation is still far away from the creative tension between cooperation and competition that has been observed in some Japanese cases. Hines (1995) describes this phenomenon in the case of Mazda's sourcing of seats. Mazda splits its purchase of seats between two suppliers on a 33-33-33 % basis that functions in the following way. Each supplier has a 33% share of the seat business and the remaining 33% share is awarded to either one of them on the basis of past performance factors, including assistance to the other competing seat supplier. The result is that both suppliers will try to improve their internal performance and to help the other as much as possible, knowing that the mutual assistance in terms of technological, organizational or managerial know-how will help the 'loosing' supplier to increase its competitiveness in relation to other seat suppliers in the supplier competition for the business of customers other than Mazda. Dyer & Ouchi (1993) describe similar procedures both with NEC, Nissan and Toyota. Two or more suppliers are continuously in competition to provide the lowest cost and the highest quality. However, a buyer does not abandon a weak supplier but helps him and obliges the stronger supplier to share information or technology in order to keep the two on a balanced level and thus enhance competition leading to continuous improvement. Of course suppliers sometimes feel it inappropriate to share their competitive advantages, but as they realise that someday they will need such assistance themselves, they view the system as an insurance. The situations observed in the present research still indicate that the European practice is quite far from the Japanese in this context. The studied suppliers could not rely on fixed shares of business or on assistance from customers, let alone competitors. Their strategic weapon was to make customers as development dependent on the suppliers as possible and actively fight against unfair copying and divulging of confidential information. 197 6.5 THE LEAN PRODUCT DEVELOPMENT CONTEXT - SUMMARY AND CONCLUSIONS To summarize, the interview data made it possible to set the current context for lean product development. Due to the limited number of interviews, no general conclusions can be drawn; however, the richness of the information makes it possible to reflect upon the different items in terms of causes and effects and outline a few ideas of how this context might evolve. The proposed model of different buyer-supplier relationships indicates that there is a continuum in the relationships between an expert supplier and its carmaker customers ranging from 'partnership sourcing', i.e. direct expert supply to 'competitive tendering', i.e. direct arm's length supply. Triangulation would constitute a third way, 'hybrid' in at least two ways. Firstly, it conserves the most important links in terms of optimising product technology, i.e. the engineering links, while the more administrative synergy links are managed by the system supplier. Secondly, it positions the supplier in a second tier relation concerning the physical supply of components, but in a first tier position in relation to engineering design. The relationship model also indicates that the situation for a supplier is dynamic, the position on the continuum is always moving. A specific supplier that at a given moment experiences a certain situation is likely to define and work towards specific strategic goals in order to optimize its role in relation to customers. In general, the results confirm the significant feature in the lean production model of increasing design responsibility for supplier firms and a tiered industrial organization. However, a better modulated picture emerges as a result of the interview and case study analysis. It was made clear that one and the same supplier has to manage a set of different relationships where each has its own particular raison d'être and needs particular managerial attention in different fields. The study allowed for the identification of a number of phenomena that are related to different relations and that influence the management of the product development process: • Learning processes related to inter- and intra-functional and company communication. The success of product development in triangulation and direct expert supply was largely dependent on frequent and rich information exchange between customers and suppliers. It is therefore of particular interest to analyse how design technicians and engineers communicate, both at an individual and collective level. Aspects of communication were also emphasized in the literature review, c.f. section 4.2.3. • Evolution of product technology and the related building of core capabilities. In direct arm's length relationships, technology evolves incrementally on the initiative of the 198 customers. This makes it difficult for suppliers to develop their expertise as there are few external incentives for doing so. In direct expert supply, tier position competition pushes suppliers to redefine the technical function of components in order to privilege system integration. In triangulation, there is a continuous improvement of component performance (as is also the case in direct expert supply) but an outsourcing logic is already applied. Here, successful cooperation between interfacing suppliers will to a larger extent than in the other cases influence technological evolution. These different phenomena largely influence the operational design work and they will be further analysed in the following chapters. The three main drivers for change in the management of the studied supplier companies were, not surprisingly, just-in-time delivery, changes in the approach to quality management, and demands for price reductions. In a time perspective, General Managers and Product Development Managers described JIT as the first important change towards lean production that their companies had been faced with and ultimately adopted. Due to important reorganization of workshops (i.e. product flows and logistics) on both sides of the supplier interface, managers expressed satisfaction with their ability to deliver just-in-time. Some problems related to internal final stocks translating an imbalance in the relation between JIT delivery and production persisted, however. The optimization of this relation in terms of cost and delivery performance is specific to each individual company. After in-depth analysis, some managers concluded that a final stock was cheaper than investing in more flexible production equipment. In the field of quality management, the main change has been the introduction and rapid application of quality standards developed by customers. This has enlarged the focus of quality management to all activities in the supplier firms, with a special emphasis on routines and procedures in R&D, purchasing, marketing and sales, human resources, etc. The contractual aspect of the quality standards and their detailed instruction concerning the formalization of all activities in supplier companies designate the former as the main formal link -somewhat comparable to the role of cross-equity in Japanese keiretsus- between European customers and suppliers. Finally, the pressure for price reductions on components was quoted as a difficult driving force to manage. An example illustrated the trade-off between the cost of redesigning components and the need for reducing their cost. When analysing change and development tendencies, the importance of technological evolution which allows for the possibility of reducing component costs (and thereby their price) was illustrated. While being difficult to manage, demands for 199 cost reductions can therefore also be seen as an important driving force behind the development of new functional solutions. Increased product development responsibility was mentioned as the most recent driving force. Here the situation is still far from being stabilized and managers anticipated even extended engineering responsibility for expert suppliers as an important development tendency. This was supposed to only consider suppliers where the technical functionality, the material properties and the interfacing qualities of the supplied components make them critical for the performance of the final product - the vehicle. Concerning change and development tendencies, five action strategies for facing the anticipated changes were identified. An analysis in terms of the underlying objectives, the planned actions, and needs and/or difficulties of realizing them showed their strong relationship to operational day-to-day activities. The strategic objectives of promoting learning, identifying and concentrating on core capabilities, developing a culture of continuous improvement, increasing customer orientation, and improving the capacity of integrating change provide a framework for the analysis of operational design work that will be undertaken in the following chapter. Important indications of how to improve management of the product development process can be supposed to emerge from this analysis. Concerning the use and perception of different lean techniques these were to a large extent perceived as something positive that has come in to support the natural way of working with component supply - above all the possibility of entering into technological negotiations with customers. In this context, no supplier expressed difficulties in having access to product platforms, customer blueprints, or mock-ups. These design artefacts and support structures underpin the common objectives between suppliers and customers such as designing an optimal function in the shortest possible time. Besides the problem of managing price reductions, a problem related to confidentiality of integrated development efforts was expressed by managers. Here reciprocal advantage from the relationship was not always evident. Suppliers had to actively manage this point in order to avoid 'stolen' designs and 'free' design consulting. A strategy of making customers development-dependent could be one possible solution. Finally, two parameters seem to be decisive for the development of closer relationships between customers and suppliers: the formal one in terms of the application of quality standards; and a more informal one around the natural ground line of developing car components. In order to better understand the latter operative design work will be the focus of the remaining empirical chapters. 200 7. MODELLING OPERATIONAL DESIGN WORK The second part of the research focused on the product development process, i.e. it focused on strategies for product development as well as the management of operational day-to-day activities in the design process. Chapters seven to eleven are based on the two case studies, and chapters eight to eleven also on a precision and refinement of the case study findings through the last series of interviews in four additional expert suppliers (c.f. figure 13). This research has tried to capture what lessons can be drawn and what strategies can be developed from the very tangible work tasks, processes and conditions that product development activities represent. The present chapter will develop a model for operative design work where four levels of work activities -individual, group, project, and systemic work- are analysed in depth. Four distinctive phenomena related to operative design work emerge through the analysis: means of guidance, design support structures, learning, and core capabilities in product development. These are then the subject of the remaining four empirical chapters. 7.1 THE CORE CATEGORY - OPERATIONAL DESIGN One of the core features that renews the product development activity in the medium sized supplier firm is that the engineering department has become more and more of an integrated part of the carmaker's development process. This means that design engineers and technicians must learn how to work in a different and much more complex context than traditionally was the case, when the supplier was a more isolated executor of finished blueprints. Through the close integration between design engineers and technicians in supplier and customer firms, the former must adapt their activities to the turbulent environment of the carmakers' product development process described in section 4.4.2. Among the different characteristics identified by Moisdon & Weil (1992) the most significant, in terms of managerial complexity compared to earlier product development activities, seems to be the volatility of the questions that design engineers and technicians are working with. Directly related to this are several difficulties: the problem of setting priorities between different questions, the need for collaboration of different actors to resolve them, and the fact that every imaginable solution will influence other design decisions in a cascade. In chapter five, section 5.2.3, the unit of analysis was defined as the product development process in the studied supplier firms. The research perspective focused on the operational aspects of product design, i.e. an analysis from inside the suppliers companies' design departments. Applying the paradigm model (Strauss & Corbin, 1990) illustrated in figure 14 to 201 the data analysis, the core category turned out to be the operational design work carried out in the design function of the studied firms. Operational design work is the central phenomenon of the product development process, i.e. of the research object. From this, different phenomena that will be developed as related categories will be identified. Their relationships to the core category integrate them into a coherent whole. Based on the case studies, it seems relevant to describe operational design following a four level model comprising individual work, group work, project work and finally systemic work. Each step mobilizes different actors and can be viewed from different theoretical perspectives, but they are all indispensable for realizing the product development process, i.e. the set of activities that transform an input (something that flows into a system - verbal and to some extent physically encoded information) into an output (something that flows out of a system physically encoded and to some extent verbal information) (Allen, 1977; Karlson, 1994). Through this definition, the product development organization is actually seen as an open system in which the four different work activities take place and are ultimately integrated. The proposed model is illustrated in figure 20. Systemic work Project work Group work Individual work in the supplier company intra-functional between design technicians and engineers in the supplier company inter-functional in the supplier company involving customers, interface suppliers and lower tier suppliers Figure 20. A proposed model for operational design work. This model corresponds roughly to models developed in the organizational learning and resource-based literature, c.f. Kogut & Zander (1992) who consider four levels of information and knowledge - the individual, the group, the organization, and the network, or Nonaka & 202 Takeushi (1995) who argue that knowledge creation takes place at three levels - the individual, the group, and the organizational levels. These models, however, have not been applied to a concrete business process such as product design. Below, the four different work activities will be described one after another. Throughout this description, different features and phenomena, i.e. related categories, specifying the core category in correspondence to the paradigm model will be identified. It is about identifying different causal conditions that lead to the occurrence or development of particular ways of organizing or executing design work, the context in which the work is done, the intervening conditions that are the broad and general conditions bearing upon actions and action strategies that are taken to develop the design work. Finally, the consequences that different ways of work might have on the strategic orientation, the management, organization and performance of integrated product development in the supplier firms will be identified. 7.2 OPERATIONAL DESIGN - INDIVIDUAL WORK The description begins at the most basic level, in the very activity of designing a component. At first glance, the design engineer or technician is working independently, at his or her work place, for example in front of the CAD, searching for information or is busy writing or drawing on a piece of paper. The point of departure is a specification, either originating from a customer or internally generated, defining a functional problem and constraints in terms of cost, lead time, functional quality, and manufacturability. The specification can be more or less complete; sometimes it only specifies a problem that needs to be solved in very general terms, sometimes it contains detailed material and dimension specifications. The description of daily design work made by design technicians in the case study companies will be taken as a point of departure for the analysis1 . Daily Individual Design Work: An Example from interviews and observations of design technicians in the two case study companies Generally speaking, the guidelines2 deployed by Product Development Management is that a functional solution should be as simple as possible. It is not 1 The example integrates data from interviews with five design technicians, two in SAR and three in SAF, and from observation of their operational work. 203 easy to describe the process of realizing a component. The design technician will begin by looking for information and ideas using several sources; the product catalogue, existing blueprints of previously realized components, the advice and knowledge of colleagues - i.e. design technicians working in the same area, other ongoing design studies, etc. Already at the specification stage, the technician normally has quite a good idea of the product family in which a new design will fall, so the catalogue or CAD research can be limited to a certain product family. In front of the CAD there are finally, only two things that can help: one's own experience or that of a colleague's. When, from his or her own experience, the individual has the feeling that something similar to a new specification has already been done, he or she checks in the blueprint library or in the product catalogue. There was, however, no storage device containing all design studies. What the design technicians dispose of are blueprints or CAD drawings of products that have been put into serial production. Based on an existing blueprint, he or she then begins to change a dimension, a flap, a thickness, etc. Finally it becomes an entirely new product. Sometimes, however, it is better to begin with a blank screen or a white paper for a first draft. The first method is good if one has difficulties in getting started - in finding an initial idea. The line of action will depend on the type of product that is about to be designed and previous personal experience. One does not search merely for the sake of searching; for some demands there is a very small chance that anything exists, thus one starts the design study immediately. Once the design study has been initiated, it is very much of a trial-and-error process: will a change of the degree of an angle promoting an easy assembly or the flow of a liquid influence the strength and the life of the component? Will the use of a different material allowing for a more complex form make it possible to eliminate one or several components in a system and thus simplify the product? There is a permanent application, testing, and validation or rejection of ideas. It is for example about (1) realising a blueprint that theoretically is estimated as defining a product answering the specification, (2) realising a prototype that allows for the testing of the design, and (3) modifying or abandoning the first concept and developing a second blueprint, etc. Collaboration with colleagues is very important. If a colleague is able to transmit an extract of an existing blueprint, or of one under realisation, this can lead to a very large gain in time because the colleague will be able to explain the pitfalls 2 The French word used was lignes directrices. 204 and point out problems already encountered in this type of design. Unfortunately, if a product is not yet in the catalogue, it sometimes happens that a design technician proposes a solution very similar to something that a colleague is working on simultaneously. There is not always the time or the occasion to talk to everyone and to take into account every project that different colleagues are working on. The highest risk for doubling a design effort is during the period before a product becomes known in the system. A real-time update would in other words reduce this kind of waste. If there is nothing particular about a project that makes it generally known through meetings, corridor talk, or conspicuous mock-ups, it is chance that makes an individual design technician discover a particular ongoing design study of a colleague. There is indeed the person responsible for patents that sees all the blueprints and he will discover redundancies, but by this stage the double work is already done. The engineering manager also signs all formal customer offers, but if he discovers redundancy at that stage it will again be too late. The sales people also may discover redundancies, but they do not always have a transversal vision of different components. When a first draft to a solution is ready, the design technician often asks, informally, either a more experienced collaborator or the design manager to covalidate the solution, propose modifications or another solution. In this way knowledge in design accumulates through time and there is a transfer of experience to the individual design technician from senior colleagues and of new ideas from him or her to senior colleagues. The example above highlights several concepts that condition, intervene or emerge as a result of the product development work: • First, the importance of guidelines was underlined. In fact, different means of guidance strongly influenced how operational design was executed; • Support structures in the form of specifications, mock-ups, idea banks, etc. were also an important part of the product development context; • Building of knowledge and transfer of experience were consequences of different ways of organizing and executing design work; they depended on different action strategies that might be undertaken in order to actively promote learning or unblock natural learning mechanisms. 205 • The guidelines translated specific strategies for product development that, through the dynamic process of learning, could evolve into specific capabilities constituting a competitive advantage for each different firm. In chapter 6, sections 6.2 and 6.3 the origins -in terms of driving forces- of different strategic orientations were analysed. The identified strategies were specified and analysed in table 13. These different features emerge immediately as categories related to operational design work. They constitute a first indication of how a model of product development might be developed based on the paradigm model, and provide a framework of elements that the continued analysis must look out for and specify further. The example also provides some elements of description of individual design work that are interesting to relate to different models of product development work proposed in the literature. A scientific view of design activities emphasizes the analytical side of the work (c.f. Shigley & Mishke, 1989; Dertouzos et al, 1989). In this school of thought, Pahl & Beitz (1988) have developed a six phase sequential process model of problem solving in design that focuses on the doing in design work. According to them, problem solving consists of confrontation, information, definition, creation, evaluation/check, and decision (p. 38). Karlson (1994) points at the rationality and linearity in these kinds of models. The design work is viewed as a process that moves from A to Z, and where a satisfactory result will be achieved only if certain explicit rules and procedures are respected. In a different perspective, Bucciarelli (1988) describe the solitary part of the job of design engineers e.g. where they explore variations in a concept, run through analysis, catalogue parts options, debug a piece of software, construct a test plan, or draft a report or documentation, as a 'navigation' within object worlds. The latter are defined as patterns and beliefs grounded in the object being designed and the technical expertise that this design requires. Bucarelli emphasizes the conceptual or reflecting side of individual design work: "participants sometimes work alone within 'object worlds' - conceptualizing, analysing, sketching, etc." (p. 167). Both the doing and the reflection perspective are illustrative of operational design work and rather than being contradictory they complement each other. Kim (1993) integrate the two in a model of individual learning that comprises two types of learning: operational (corresponding to the doing side of product design) and conceptual (corresponding to the reflection side of product design), see figure 21. 206 Individual Learning Conceptual Elaborate/Plan Implement Assess Observe Operational Individual Mental Frameworks Routines Models Figure 21. Individual learning, after Kim (1993, p. 40). Even if this model is intended to explain individual learning it is applicable to the individual design work discussed here. The design engineer or technician implements an idea for solving the problem he or she has at hand. Then he or she will observe the result -in the form of a blueprint or a prototype- before moving into the assessment of the result. As a function of the assessment, he or she elaborates a modified solution that will be implemented, and so on. The advantage with this model is that the design work is seen as an iterative circular process and not, as in the scientific view, as a linear one. This corresponds much better to the observed reality in the case studies. This model has one important shortcoming, however. It makes an artificial distinction between thinking and doing. This might indeed be very useful, even indispensable, for analyzing what is happening in order, for example, to improve the organization of product design, but it is not very true to reality. Through the concept of reflecting-in-practice, Schön (1983) integrates doing and thinking: operational activities are permanently supported and guided by an inner picture of what the result will be. Therefore, operational implementations and observations will also be affected by conceptual reflections. This consequence is that a practitioner, for example a design technician, will be unable to give a complete and accurate description of or justification for what he or she is doing. The design work is in fact a practice consisting of a large part of tacit knowledge and insights. Karlson (1994) defines this view of design work as the 'art of the practitioner' perspective. Schön's (1983) inner pictures seem to correspond to Kim's (1993) mental models that are developed through accumulation of experiences from conceptual and operational activities as illustrated in figure 21. In a similar way that the inner 207 pictures support and guide operational activities, mental models influence both them and the conceptual ones by more or less explicit routines and essentially tacit frameworks respectively. In the example above, the tacit component of the design work is translated by the fact that there was no formal and generally applicable procedure to follow, for example, when starting out on a new design project. Sometimes the design technician did research among existing blueprints, sometimes he or she took an existing design delivered by a colleague as a point of departure, sometimes he or she began with a blank screen or a white sheet of paper. It was the technician's inner pictures and mental models formed by previous experience that determined what line of action was applied. The fact that design technicians could give few rational explanations of certain choices or undertaken actions was often perceived as a problem at the management level - a lack of rigour that could lead to a problem of traceability. Even though much can be done in order to formalize best practice, the 'Art of the Practitioner' perspective suggests that a part of the product development work will always remain a tacit practice. This tacit part will always be present as it precedes standardized procedures; the art of a profession evolves in the front-line of new learning. Even if innovation always starts with the individual (Nonaka, 1991), product design is not only an individual affair. Design engineers and technicians frequently engage in formal or informal groups, product design is in other words a social process (Bucciarelli, 1988). If, as the analysis suggests, operational design work is best viewed as an art of the practitioner comprising an important dimension of tacit knowledge and if this art begins at the individual level but at the same time is a social process, the problematic of knowledge building and knowledge transfer becomes crucial for improving product development performance. In the literature review, tacit knowledge was related to the building of a common cognitive ground and the altering of work tasks. The data analysis reactualizes the question of how this can be done and adds emphasis to the link between individual and collective learning. Also the example of individual design work referred several times to interaction with colleagues. The next level in the proposed model therefore considers group work. 7.3 OPERATIONAL DESIGN - GROUP WORK This first widening of the perspective on operational design concerns problem solving that is still mainly focused on the aspects of product technology because it takes place within the 208 design function in a company. Operational design at the group level is here defined as actions taking place between design engineers and technicians, i.e. in a group with shared values and similar professional languages executing the same basic profession. If design work is illustrated by Kim's model analysed above, the group perspective means that each of the single activities -implement, observe, assess, and elaborate/plan- can be performed collectively in collaboration between several design engineers or technicians. 7.3.1 Informal Group Work It seems useful from the case observations to make a distinction between informal and formal group work in the design function. To begin with informal group work, it was found to occur whenever it was natural according to the work situation, but it also depended on each individual's professional behaviour and on structural supports for collaboration. As the example explained above, it occurred for example: • During idea generation in the elaboration/planning phase, when an individual turns to colleagues for advice on how to solve a specific problem; • In the implementation phase whenever there is a need for real-time advice during application; • In the observation phase when outcomes are presented to colleagues for their comments and reactions; and finally • In the assessment phase where outcomes are analysed on a more abstract level focusing on causal links between the outcome and the way of getting there. Support structures for informal group work were found to have an important impact on its frequency and intensity. At COF, for example, all the design technicians had been regrouped in one common office allowing for spontaneous common problem solving, idea exchange, and asking for and giving of advice. When asked the very straightforward question whether integration and cooperation between design technicians was beneficial for design productivity, the unanimous answer was yes. Both project managers and design technicians expressed great satisfaction with the reorganization. It had much facilitated problem solving in collaboration with colleagues and thus individual and collective learning. This organization also seemed to reduce product development lead time -at least this was the impression of the interviewed design technicians, but no formal measurements had yet been made-, and to reduce redundant work in terms of repetition of previously committed errors. This last effect was purely implicit and the outcome of a learning process that this organization had started off. This was a particular form of learning, where operational design technicians worked in small groups on a problem that occurred during a design work session and where 209 the joint problem solving most often occurred in real time, i.e. it followed directly upon the discovery of the problem. The possibility of reducing costs related to the reinvention of solutions that has also been discussed theoretically as one of the main driving forces behind a focus on organizational learning, c.f. e.g. Goodman & Darr (1996). When relating these observations to Kim's (1993) model of individual work and Schön's (1983) ideas of interwoven conceptual and operational activities, there is here an additional collective reflection-in-practice integrating the entire design cycle. This multiplies the chances of finding relevant solutions. When colleagues are informed about each other's problems, they will unconsciously treat them mentally while working on their own respective design problems. Suddenly an idea for a solution comes about while solving another design problem. However, in the absence of some type of information dissemination and organizational memory device (meeting, meeting protocol, mock-up, blueprint, checklist, data base, interactive software) that could capture new learning and make it available to the entire design organization, binary or small group learning remains with the involved actors at the individual level rather than at the organizational one (Goodman & Darr, 1996). Project managers, other design technicians (for example at COF there were two shifts of design technicians, one in the morning and one in the afternoon), design engineers allocated to specific projects, and sales engineers will not be part of this learning process. A further specification of how such devices might function will be presented under the discussion of project and systemic work. 7.3.2 Formal Group Work A formal work group is "created by the formal authority of an organization to transform resource input (such as ideas, materials, and objects) into product outputs (such as a report, decision, service or commodity)" (Schermerhorn et al, 1991, p. 221, quoting Herold, 1979). Formal work groups within the design function were discovered during the case studies, especially in the form of weekly design reviews where the following objectives were more or less explicitly stated: • Identify the state of advancement in different ongoing projects; • Inform about and discuss new development projects; • Determine priorities between different projects managed by the same project manager; • Inform about and discuss specific problems such as customer complaints. These meetings, that in both case study companies were led by the product development manager and where project managers participated systematically and design technicians 210 occasionally, did not seem to represent an optimal use of time. One reason might be that they were too oriented towards issuing directives by the design manager. Operational people principally seemed to wait for new instructions concerning project priorities and did not see opportunities to draw much useful information concerning each other's problems. Another problem was that these meetings seemed to lack what can be called 'broad-band feed-back structures', i.e. a registration of all that is discussed during a group session and where participants afterwards have the possibility to make comments and observations concerning the evolution of different points during the time between two different meetings. In the absence of such a support device, participants did not pay much attention to what was said by another participant, they did not express themselves very vividly, and seemed to be mostly preoccupied with their own problems. Thus, while design technicians fully recognized the usefulness of collaborating, project managers seemed to share this opinion to a lesser extent. When observing the work of project managers and talking to them during execution, three reasons for this emerged. Firstly, their work tasks were much more heterogeneous than those of design technicians, this was of course due to a much wider spectrum of activities. Therefore, their preoccupations could be quite different at a given moment, for example at the time of a weekly meeting, leading to a lack of interest in each other's discourses. Secondly, their personal work methods were also more heterogeneous than those of design technicians. Some project managers were oriented towards a rigorous planning of every single project activity and preferred to manage their work to a large extent from their desk in an analytical, almost scientific way. Others were more relational-oriented and practised a management-by-walking-around approach3 . In other words there were some problems of compatibility between the management methods of project managers, making their incentives to learn from each other weaker than in the case of design technicians. Thirdly, a certain career competition between project managers could be registered. This was not a very apparent or significant problem, but for this reason, project managers seemed less open to each other than design technicians. In order to improve the productivity of this kind of formal group work, and make it more interactive and a real support for learning and improvement in the development process, it might be useful to analyse the effectiveness of these meetings and other group work activities. Schermerhorn et al (1991) quote Likert's (1961) eight point list of characteristics of an effective group as classic: 1. The members are loyal to one another and the leader. 3 Both approaches have their pros and cons The best way is probably a combination of the two. This is also what is proposed by the project management literature (e.g. Clark & Fujimoto, 1991, Bowen et al, 1994). No judgement in favour of either of the two is made here. The research did not include measurements making it possible to recommend one method rather than the other. 211 2. The members and leaders have a high degree of confidence and trust in each other. 3. Group values and goals express relevant values and needs for members. 4. All activities of the group occurs in a supportive atmosphere. 5. The group is eager to help members develop their full potential. 6. The groups knows the value of 'constructive' conformity, and knows when to use it and for what purposes. 7. The members communicate fully and frankly all information relevant to the activity. 8. group's The members feel secure in making decisions that seem appropriate to them. Following the field observations, group work effectiveness focused on three items taking up or covering some of Likert's (1961) characteristics: (1) the central theme or subject treated, (2) the way participants work and discuss (implication, atmosphere), and (3) formal structural supports for the group work. In order to obtain a more specific and analytical picture of group work effectiveness, these items could be developed as follows: 1. What is central for work and discussion in the group? • Problems, • Causes, • Solutions - individually or in group, • Verification of actions taken since the last working session, feed-back concerning issues and questions raised during the last working session, 2. In what way do the participants work and discuss? • Is the participation strong or weak ? • Is the discussion management-driven or participant-driven ? • What is the mode of the discussion? positive (seeking for solutions) negative (complaints and piling of problems), 3. What are the supports? Meeting review. Example: weekly re-actualised project listing where a comparison of the series of lists makes it possible to establish a historical context for each project in order to pinpoint different problems, and look for their causes. • 212 Interactive meeting review where participants communicate their opinions on each discussed point on a circulating document or on one put in an information system network, • • Diagnosis and performance measurements of the discussed problems including risk indicators to classify the questions in a hierarchical order in order to choose the most important ones, • Verification of undertaken actions in order to assess the relevance and efficiency of action plans and reduce gaps between planned and realized ones, • Formal procedures integrating found and confirmed solutions with the objective of enriching the shared knowledge-base in the company. When discussing such formal procedures with Product Development Managers and design staff, their propositions ranged from simple information notes circulating within the function, via a continuous evolution of written procedures for design work, to an integration in design data banks and different computer-aided systems for information transfer. The project managers at both COR and COF had improvement objectives for the weekly meetings. They wanted to see them as a vehicle for (1) materializing a kaizen (continuous improvement) effort by searching for the causes behind all kinds of problems, discuss them and propose possible actions, and (2) promoting learning by integrating experience of problem solving, design solutions, and customer behaviour from each individual's work into the common knowledge base. Creating more of collective dynamism in formal group work requires a long-term cultural change that needs structural support. A first step could be to analyse meeting and group work efficiency by an assessment of three levels according to the framework developed above. 7.4 OPERATIONAL DESIGN WORK - PROJECT WORK The two case studies allowed for a close observation of how a project-based organization in product development works. For COR and COF the project organization was principally seen as a way to balance the trade-off between the need for specialist skills and functional integration. Focusing on project work means an introduction of inter-functional work and interaction between different professional groups in the design work. In section 4.4.1 a project organization in product development was defined as an organization where all the necessary resources for developing a new product come under the direct control of a project manager for the project life time, i.e. from idea generation to product launch. Furthermore, the project group's character must be broad, i.e. beyond product engineering it should comprise functions such as marketing and sales, purchasing, quality management, process engineering, and manufacturing. Project work thus means a transfer of the focus from task to process. 213 COR had gone the furthest in their approach to integration within project groups by experimenting with knowledge transfer through the alternating of work tasks within a defined spectrum of activities in the sense discussed in section 4.4.2. In COR this had been a part of the training for project managers. Project managers were normally chosen from the design department. In order to prepare them for their project management task, they were first trained and assigned as process engineering technicians in order to achieve a complete knowledge of this function. The result was that in projects needing less specific competencies in different fields, technicians with a common basic platform of knowledge both in design and industrialisation could alternate their work tasks. Project participants with this profile can be called integrated product managers. This example illustrates how the problem of tacit knowledge is managed through the complete integration of different professional skills from the practice of different professions. It shows that the theoretical conclusion reached in section 4.4.2, that the only true way of transmitting knowledge is learning through alternating work tasks within a defined spectrum of activities, is relevant. However, the observation concerning the complexity of the design task -hindering the use of generalists- also proves its relevance through the case studies. At COF, the project manager considered the specialised skills needed in design and process engineering too different to make an exchange productive. A direct comparison between the two case study companies provides some explanations for this. COF products were generally produced in much larger quantities and the production tooling was more specialized compared to that of COR. Thus, more specific expertise and a longer experience was needed for process engineers at COF. A design technician during apprenticeship would simply not be productive in any process engineering task. It must also be underlined that at COR the job rotation was applied only to less complex projects, e.g. modifications of existing designs and incremental process improvements. The study indicates that a structure with few project managers and design technicians drives the introduction of integrated product managers, and, thus, the development of tacit knowledge leading to improved coordination. In the COR structure there were only two design technicians, one design engineer, one commercial engineer (whose role was very important in the design interface with customers), and three process engineering technicians (the project management role was alternated between these actors). This small structure meant an important need for competence flexibility in order to balance the work load in product and process development. However, in spite of the objections against this kind of task rotation at a technician or engineering level, its interest remains for at least three reasons, identified through the case studies, that could be beneficial also in larger organizations: 214 • Alternating work tasks is the only manner to acquire complete knowledge in a field. Direct operational experience allows for the development of tacit knowledge; • Based on the acquisition of knowledge concerning the conditions of interfacing work tasks, commitment to inter-functional considerations in individual tasks develops more naturally thanks to a better understanding of different priorities and constraints, and the perception of how guiding visions influence the opposite side of an interface; • Through task enlargement, the status between different professions can be better balanced. Bowen et al (1994, p. 46) underlines the problems related to viewing different employees as more or less important. This implicit classification of professions in dominant and non-dominant disciplines is generally most visible between design, process engineering, and manufacturing staff. Design staff tend to have the highest status in engineering companies (Bowen et al, 1994; Karlson, 1994). The increasing importance that the design function has gained in the automotive supplier sector in recent years tends to reinforce these tendencies. Task enlargement can then be a means for coping with integration problems (above all design for manufacturing) that follow an unbalanced status of different professions. The issue of dominant and non-dominant disciplines also highlights the question of who has the right to formulate problems, i.e. from what perspective and in which language problems will be presented. Karlson (1994) argues that true coordination between specialists assumes that this right is shared, at least in part. To sum up, task enlargement and job rotation in the product development process can eliminate or reduce several problems related to different understandings, false priorities, and lack of knowledge and inter-functional understanding. The analysis indicates that the lean practice of internal and external career management by rotation also has a practical relevance besides its theoretical founding. Through this line of action, different people, occupying for instance product and process engineering positions in a given moment, share the same experiences and a common language. However, the study did not provide any general indications that a true shift between professions, as proposed for example in the context of the lean enterprise (Womack & Jones 1994), is practised in a formalized way, neither intra between different functions in the same company, nor inter between different firms (suppliers and buyers). Single examples of internal systematic rotation were found, but no data indicate tendencies towards inter-company rotation. If people are moving between different levels in the tier structure interview data indicate that this was seen more as a 'hostile' career rotation: a shift to a lower tier supplier could easily be regarded as a down-grading; the departure to a higher tier supplier or a carmaker as an escape to a 'better' situation (a little less pressure and stress at work). 215 Project managers have an important role to play in facilitating and nourishing inter-functional cooperation. Comparing case study observations with the role and qualities of a project manager discussed in section 4.4.1 two issues emerge as important: • The role of project leaders is, apparently, not always clearly defined. In the case study companies, their image was to a large extent that of a design engineer with supplementary administrative tasks, something that gave less credibility and possibility to intervene in commercial, process, and production issues; • The fact that project managers were operating design engineers meant that they could not coordinate projects full time. This meant difficulties in following a project through all the way to running production. Often, project managers expressed problems in managing their time. They perceived a lack of time in both operational and managerial tasks. However, there is an important advantage in this situation; coordination is ensured by operational people and not only by liaison people. Clark & Fujimoto (1991, p 104) argue that project managers that influence engineers and technicians only through liaison meetings and not through an operational participation in the development process (from concept generation to intermediate or end use by customers) have little chance of ensuring a relevant quality in project functionality, coherency and fit. This double nature of the role of the project manager -more or less imposed by the modest size of the kind of companies that were studied- might then, in fact, bee seen as an advantage. General management and engineering management must take this into consideration when developing means of guidance for product development and project management. Project managers must be convinced that the conflict and stress that are caused by the double role of being a design engineer and a project coordinator can be positive and not only negative. The following two positive outcomes were particularly visible in the case study companies: • Integrated problem solving frequently occurred between a project manager and one or several of his or her attached design technicians. In this way, technical parameters were balanced with more general ones in terms of customer satisfaction such as component cost and development lead time. This way of work could minimize design changes and conflicts between project managers and technical staff; • The fact that project managers shared operational work increased their credibility; firstly their instructions would be more realistic, secondly they themselves had to work under their own conditions, something that encourages other project participants to do the same. The classical conflict for project members between belonging to a functional department and participating in a cross functional project was not present in the case study companies. In fact, the project organization was seen as a natural way of working with product development based on the naturally occurring spontaneous interaction discussed in section 4.4.4. 216 The existence of this spontaneous interaction was confirmed in the case studies, especially in COF. However, an example illustrates that this matter could be differently understood in different hierarchical levels. At COF, one of the two project managers for metal components was located, together with one of his attached design technicians, in the tool-making office right next to the workshop. The project manager expressed some worries related to this organization in terms of a limitation in design creativity due to a too important impact of manufacturing considerations in the design process. This opinion was not shared by the design technician. On the contrary, the closeness to toolmaking allowed him to become aware of the toolmakers' opinions concerning the manufacturability of different designs in real-time. Being located separately, as was normally the case, and needing to call or walk over to the tool making office, never allowed for the same frequency and continuity of information exchange. The organizational arrangement of co-location also provided an excellent opportunity for interfunctional learning. The natural and spontaneous organization of product development work called for joint problem solving between product and process engineering technicians. In relation to the inter-functional integration problems, this example also indicates that smaller companies have an advantage in inter-functional working due to their more organic organizations. In these companies, asking how people manage to work together is asking the wrong question. Instead, the relevant question is how the formal organization can be adapted to better support integrated work and make it more explicit. Karlsson & Åhlström (1996) argue that in order to make teams function effectively, basic changes in systems, practices, attitudes, and behaviour must occur in the entire organization. Data from the present research confirm their opinion that a change towards a cross-functional focus in the higher level management perspective must take place in order to develop project structures and give them the status in the organization that their efficient functioning requires. The means of guidance -from the guiding visions to the daily performance measurements- must be coherent with cross functionality to make it possible for efficient project structures to emerge. As a direct consequence of the limited size of the studied companies a cross functional focus, at least on the technical side, was facilitated through the existence of an inter-functional directorship for product and process engineering. This was the case at both COR and COF. 7.5 OPERATIONAL DESIGN WORK - SYSTEMIC WORK 217 Looking at design work in from systems perspective means taking on a wide perspective including all actors and phenomena that participate, intervene, or are involved in the design process, and focus on relationships and dynamics of interaction between different actors and phenomena (Hanna, 1988). At the beginning of the present chapter, the product development process was referred to as an open system. Morgan (1986) lists the following key issues that an open systems approach focuses on: • The environment in which an organization exists. The perspective in this study will focus on the 'task environment' defined by the design organization's direct interactions with customers and interfacing suppliers; • Interrelated subsystems in an organization. In the present study this means an analysis of the relationships between individuals, functional groups, and project groups, i.e. the entities examined above; • Congruency and elimination of dysfunction between different subsystems. This is also an important part of the present research; being a management research project an important objective is to identify problems and propose ideas concerning their improvement or elimination. In the example presented in section 7.2 the design technician had difficulties in explaining of what exactly his work consisted. This is a common observation in studies of design work (Bucciarelli, 1988; Karlson, 1994; Moisdon & Weil, 1992). The reasons are numerous; work tasks are interdependent of other work tasks and other actors, they take place in a chain, and one task leads to another - something that often breaks a planned logic. Systemic work, the last level in the proposed model (c.f. figure 20), attempts to clarify this complexity. It covers the three steps developed so far and adds the very important parameter of interaction with customers and interfacing suppliers. However, it begins with a look inside the design function, at the systemic nature of internal development work in order to point out the importance of integration between different teams and project groups. In COF, three specialized teams dedicated to specific development projects coexisted with the project teams responsible for the current developments. This organization seemed to be particularly productive for promoting the evolution of product technology and product innovation4 . The competencies and roles of the heads of these three specialized teams closely resembled those of so called consulting engineers (Bowen at al, 1994). These are defined 4 An innovation is defined as "the elaboration and the diffusion in the economy of new and improved products or processes" (OECD, 1970 quoted in Chanaron, 1991, p. 17.). In this research the distinction is made between evolution of product technology -which comprises improved materials and/or functionality based on an existing product, and innovation which means the diffusion of a new product, new both to the comp any and to the market. 218 as permanent employees of a company having the following characteristics (Bowen et al, 1994, p. 131-132): • Development projects of products with a high degree of novelty are concentrated on them; • The technological evolution of the company largely depends on them. They become the internal drivers for innovation because they take on extensive responsibilities to choose and implement new technologies in new generations of products; • They materialize or articulate the core competencies of the company due to the fact that their projects integrate technologies that are critical for several existing product families previously separated; • They are expected to become 'champions' in their speciality and they are given the resources needed to achieve this. Like the consulting engineers identified by Bowen et al, (1994) the heads of the dedicated teams and their associated staff at COF ensured a technological leadership within the product engineering function. Besides their strong focus on innovative technologies and product systems, they played an important role as integrators and forecasters. For these reasons it was very important that they were able to draw on existing design know-how on the one hand, and on the other transmit their new achievements in product technology to the entire design staff for integration in the more basic design activities. A first structural condition for this two-way interaction (existing design knowledge - product innovations - integration in the basic design studies) was an organizational cohabitation between the consulting engineers and the project managers and design technicians responsible for the more current engineering projects. The result of this was an interlocking cycle between individual implement-observe-assess-elaborate/plan processes of design technicians and consulting engineers respectively. In spite of important differences in work context and strategic visions guiding their work (rapidly developing a large number of product solutions proposing incremental improvements in functionality and cost on the one hand; privileging innovativeness and quantum leaps in product functionality, material and process technology in one specific functional area over a long project lead time on the other), neither could succeed in his tasks without the other. This emphasizes systemic interdependence between these actors and their activities. Besides their important role for product innovation, i.e. development of new functional solutions and use of new materials, the consulting engineers provide an interesting example of how to deal with the trade-off between integration and the need for specialized skills. The majority of the interviewed product development managers expressed worries concerning this problem. As mentioned earlier in this chapter, COR experimented with task enlargement 219 between product and process engineering, but many design managers in other companies were worried that such a dispersion of professional know-how and knowledge would have negative effects on both product and process innovation capacity. The consulting engineers at COF were, if not senior, at least very experienced design engineers. Bowen et al (1994) compare the ideal competence profile of these people to a 'T'. The T-shaped engineer is a person with deep expertise in one area and broad experience in several areas. This specific competence profile and the functional inclusion of consulting engineers in the design department -they were not isolated in a separate research function- prevented the isolation of experts pulled away from the central missions of development projects (that are value added for the customer in terms of product quality, product performance, manufacturability, ease of assembly). Moreover, these facts ensured both a possibility of working in-depth on specific technologies and the dissemination and integration of these specialized skills in more current design projects. The identification of people susceptible to develop such a competence profile and take on the role of 'consulting engineers' is therefore a critical issue both for engineering and top management. Turning to the external side of systemic work, customer-supplier integration is central for the new product development map (c.f. section 4.1). Besides formal intervention, i.e. through the transmission of specifications (that could concern both detail-controlled parts and black box parts), the case studies confirmed a permanent information exchange between the design departments in supplier and customer firms. This information exchange took place essentially at two levels; the project one and the individual one. The interlocutors in customer firms were design technicians or engineers responsible for the detailed design of the system or interfacing components where the components of the studied companies would fit in. They were also project managers responsible for coordinating the development of vehicle parts, systems or technologies where the studied suppliers had a role to play. Information exchange at the project level took place between one or several of these interlocutors and the project managers in the studied companies. In this case, he or she would be the only interlocutor between the studied companies' design functions and those of customer firms. The internal diffusion of the information then became a question of the internal project management structure in place. The case studies revealed some problems related to this way of functioning. If the interfaces that a piece of information had to pass were too numerous, there was a risk for distortion of the initial message coming from the customer. In the case study companies it was difficult to define what information different actors, i.e. sales engineers, project managers, and design technicians, needed to efficiently execute their tasks. It was also difficult to define 220 the different domains of intervention of these different actors. It is important to underline that the problem of information transmission wasn't related to deliberate withholding of information, but to a distortion due to the fact that different actors did not perceive a question in the same way. If an initial message, from for example a design technician in a customer company, passed via a sales engineer, then via the project manager, before reaching the design technician concerned, the final message would inevitably be different. This caused several problems: • The design technician in the supplier company would not return information and questions that were 100% relevant in relation to the original request; • Implicit elements in the initial message and in the answer might disappear; • When project managers visited a customer's design department they would not perceive a problem that a design technician would explicitly look for if he or she went there him or herself. Two ways of reducing these problems seemed possible in the studied companies. Firstly, the tasks of each actor could be better defined, the project managers should not be too involved in detailed technical design questions outside their own particular design activities. Secondly, it seemed possible that the design technicians could be allowed to have more direct customer contacts in order to exchange detailed technical information more rapidly and with improved precision and relevance with their counterparts in customer firms. The interface problem and the related recommendations are intimately related to the question of how information is transmitted between customers and suppliers during product development - there is a question of how different types of communication take place in the customer-supplier interface. When the supplier is working on black box or own proprietary development projects (in a triangulation or direct expert situation) there is a need for very frequent information exchange for mutual adaptation and evolution of interfacing components. Field observations and the problems analysed above indicated that information exchange at the project level was not enough for ensuring optimal integrated development. Through the case studies, a specific way of communicating at the individual operational design engineer or technician level could be observed. This exchange of information with customers was very frequent and broad banded, i.e. concerned issues that at a first glance sometimes looked quite distant from the actual activity performed by the engineer or technician. It was particularly visible in indirect and direct expert supply, c.f. table 13. Through this information exchange, participants on both sides of the interface learned about new practice that could be transferred to other parallel projects. Other important actors also come into the picture of designing black box parts, namely suppliers responsible for interfacing components evolving simultaneously. For the studied 221 companies, their integration in the development process caused more problems than that of integrating with customers. It seemed that the system coordinators did not fully master this complex integration process -that only recently had emerged- but that tended to become more and more important. In addition, the individual participants, like the studied companies, were normally less focused on their horizontal colleagues than on their vertical customers. The research findings indicate that the systemic nature of design work draws managerial attention to integration between participants in the design process. With an increasing number of interfacing actors, both external and internal, the demands for efficient coordination also increases. Through the change in the industrial organization towards more inter-company collaboration and concurrency in the development process, suppliers have been taught the need of integrating both internal operations and external inputs the 'hard way'. Integration has become the inevitable means to an end of remaining an expert supplier to the carmakers. However, at the same time that systemic work is a driver for integration it is also a facilitator: peoples' minds will turn towards more of an outward looking perspective because they will understand the sense of communicating and informing through their own need of information coming from others. The above discussion emphasizes that managerial focus should be on supporting the systemic nature of design work and simultaneously avoiding a too scientific way of managing product development in favour of a mode that sets creativity free and supports the natural working practices in product development of coordination, integration and learning between different participants and functions. It is through systemic work that core capabilities take form, product technology evolves, and inter-firm and intra-firm learning are connected. At the same time, the negative side of systemic work, namely that everyone tends to be preoccupied with anything at any time, must be recognized and mastered. Design technicians often expressed a problem with too dispersed work tasks that hindered them from working with concentration during longer periods on new design solutions. In fact, it might be fatal to lose the thread of a concept idea, above all if the design study is in its starting phase. Their difficulty can be interpreted as if breaking up the problem solving cycle of implement-observe-assess-design/plan is not an optimal way to work. Another example illustrating the problem of fragmented work tasks shows up from one of the case study companies where problems of feeding back relevant information to individual design work occurred due to a fragmentation of the product development process. This concerned modifications of existing component designs made by design engineers or technicians other than those responsible for the original design. If relevant feed-back structures or informal communication links were lacking between these actors, the design technician responsible for the original design would neither be aware of the modification, nor of the 222 reason behind it. The result was that he or she continued to use the same solution for other similar functional problems without knowing that it was not optimal. To avoid this, project managers proposed that the same design technician should follow through each design study from A to Z during its whole life cycle. 7.6 THE PRODUCT DEVELOPMENT PROCESS IDENTIFICATION OF RELATED CATEGORIES - SUMMARY AND 7.6.1 Summarizing the Four Step Analysis of the Product Development Process The analysis of individual design work introduced several important basic assumptions about how the product development process functions. Design work is characterized by interlocking cycles of doing (implementing and observing), and reflection (assess and plan/design) - in other words it is a 'reflecting in practice' to a large extent guided by tacit knowledge and insights. Through its cyclical nature, individual design work can also be seen as an individual learning process where incremental improvements of functional solutions and new ways of working are developed. The individual design project participant is also strongly influenced by instructions and performance measurements established by the hierarchy. The nature, content, and inherent coherency of such means of guidance will determine how individuals will perform their work tasks. The analysis of group work emphasized the importance of informal collective work, that seemed to be more productive than formal group work because it emerged naturally if and when there was a need for it. In fact, it was found that design work called for the creation of temporary micro-organizations that involved all actors concerned by a specific problem. Such organizations would constantly be created, support the realization of specific activities, be dissolved, and then recreated with different participants for a different end. The analysis also identified a specific learning situation between small groups of people having the same professional background. However, the absence of transfer mechanisms of such acquired learning seemed to be blocking its more general diffusion even within the design community. This calls for a focus on the role and importance of support structures -playing the role of a collective memory- as a context to the operational design work. Such items were also identified as important idea sources in individual work. 223 The fact that the most interesting cooperation seems to be the informal one does of course not mean that management of the design work should be left unattended. On the contrary, the evidence suggests that it should focus on how to support the natural working practices in product design and how to unblock spontaneous mechanisms of learning. The research findings confirm the theoretical conclusion from chapter four that a bottom-up perspective should complete the top-down one in organizational design of the product development function. The analysis of project work focused the attention on the problem of inter-functional integration and the transfer of tacit knowledge between different professional groups. Again, learning was identified as important; in this case the central point was how to develop intracompany, inter-functional learning. This problematic emphasized the importance of adopting the formal organization to a natural working practice in product development work. Management must develop a cross-functional focus and translate this through adequate strategic visions and performance measurements. Finally, it is through a consideration of systemic work with crosswise communication patterns that the notions of collective learning and evolution of product technology takes shape. Even though innovation begins at the individual level, the findings suggest that evolution of product technology takes place through a systemic process integrating different actors within the company as well as customers and interfacing suppliers. Systemic work also focuses managerial attention on the identification and development of core competencies. In section 4.3.1 core capabilities were defined as collective knowledge and aptitudes embedded in the structure and specific to each enterprise. They were identified as built up of combined basic resources, skills, management systems, and physical systems, and formed by technological development, collective learning and organizational processes. Through this definition it becomes clear that core capabilities relevant for product development depend on a systemic nature of design work. For example, the capability of COF to patent a large number of product technologies was due to a capability of integrated innovation based on interaction between existing technology and design knowledge in the company, and specific dedicated research activities. As will be analysed further in the following chapters, this capability could be developed by integrating the patenting and patent scanning activity in the design function, and by reinforcing process scanning activities. These findings indicate that managers should support the systemic nature of design work systemic work was found to be both a driver and a facilitator for integration between an increasing number of internal and external participants in the product development process. A main problem in systemic work was the fragmentation of work tasks. In order to overcome 224 these problems, alternative ways of structuring and organizing product development work must be examined. This will be further discussed in chapter nine dealing with design support structures. 7.6.2 Categories Four distinctive phenomena related to the core category -operational design analysed abovehave been clearly identified through this analysis. These categories are: • Means of guidance, • Design support structures, • Learning in product development, and • Core capabilities. These categories are all possible to relate to the core category following the paradigm model (c.f. figure 14). Below, the characteristics of these categories will be recapitulated briefly and their position in the paradigm model will be defined. The following empirical chapters will then discuss these categories in greater detail. 7.6.2.1 Means of Guidance Strategic visions were frequently quoted as important for guiding operational work. Performance measurements also influenced the work carried out by different participants in the product development process. Such different means of guidance need particular attention because even though strategic visions are general and transversal in relation to functional limits, performance measurements are specific depending on different functions. In order to avoid discrepancies between strategic intentions and operational outcomes, particular attention must be paid to the coherent management of these means of guidance. Using the paradigm model, means of guidance are defined as causal conditions to the core category. They guide and influence how operational design is executed. 7.6.2.2 Design Support Structures 225 Another concept emerging from the study was design support structures. Different types were quoted in different situations occurring in operative design. For the individual work, product specifications are the first tangible artefact from which a design study takes off. Besides this, verbal support, prototypes, mock-ups, and already existing components were quoted as important information carriers. Concerning support structures, the analysis of group work emphasized the importance of spatial organization through the positive result of the co-location of design technicians in a common design office. In project work similar actions of co-location of product and process engineering technicians was a causal condition for inter-functional learning. Additional support structures were computer-aided information systems for information transmission and storage. Seen in the paradigm model, design support structures are defined as pertaining to the context in which operational design takes place. 7.6.2.3 Learning When analysing individual, group, and project work, different learning needs and situations emerged in each of these types of work. The individual problem solving cycle of implement observe - assess - elaborate/plan can be extended to take place collectively in a group of colleagues within the same profession. The learning is then focused on a collective development and sharing of best practice in design problem solving. At the project level, learning is seen more as a means for bridging gaps in knowledge and understanding between different professional groups. The main objective there is to improve coordination in order to respond to driving forces such as the need for reducing product cost and design lead time. At the systemic level, finally, learning was identified as a vehicle for integrating customer needs into the component design activities and for connecting operational day-to-day design and more specific research activities. The more and more interactive nature of design work also increases the opportunities for learning. This learning must be collective so that incremental improvements, best practice, and new ways of work are quickly disseminated in the entire development system, i.e. to all its different individuals dispersed in different functions and companies. 226 Seen in the perspective of the paradigm model, the learning concept is defined as an intervening condition to the core category. This is because the outcome of design work will be influenced -incrementally improved or more radically changed- as a function of what learning takes place. 7.6.2.4 Core Capabilities Learning, emphasized heavily in all types of design work, is not an end in itself. Core capabilities were found to take shape from an interaction between what happens at the most basic level -in individual design work-, the guiding visions deployed by management, the activities of individuals or teams working on long and medium term research projects, and last but not least the influence of customers in terms of direct product development specifications and more indirect interpersonal interaction during the development process. When analysing and comparing the notions of core capabilities and learning in the light of the paradigm model, learning emerges as an action strategy for building core capabilities. Thus, core capabilities can be seen as one of the objectives of learning. In relation to operative design work the emergence of core capabilities is a consequence of a well managed product development process. 227 8 MEANS OF GUIDANCE Through the interviews with General Managers, it was made clear that product development was seen as a strategic problem in all of the studied firms. A more strategic approach to product development -via visions and objectives complementing detailed specifications- has also been much argued for in the related literature (c.f. Takeuchi & Nonaka, 1986; Bowen et al, 1994; Karlsson & Åhlström, 1996). Takeuchi & Nonaka (1986) point out that new product development projects can act as catalysts for bringing about change in the entire organization. Bowen et al (1994) argue that in spite of product specifications and strategic plans, critical uncertainties, where development team members look in vain for guidance, emerge throughout the product development process - a focus on strategic guidelines for product development are therefore argued for. Karlsson & Åhlström (1996) identified a changed management focus as an important priority for improving coordination in product development. A strategic approach to product development means that it becomes a top management issue and that the role that a supplier will play in product development in a medium and long-term perspective is subject to the greatest attention. In the studied companies, this new and additional focus for top management was due, on the one hand, to the historical role of a development intensive supplier that the companies have played (but development had traditionally been undertaken more independently from customers - the studied companies were traditionally focused on supplier proprietary parts), and to the outsourcing tendencies emphasizing development capabilities that have been acting as driving forces for change during the last five to six years on the other. Parallels can be drawn with the change in perspective concerning the management of quality. Juran (1981), for example, argues for making quality a top management issue and installing a top-down approach to quality management. This has today not only become general practice, but an obligation for companies that want to obtain an ISO 9000 certification; the first chapter in these quality standards is entitled 'Management Responsibility'. Much points at the same evolution for product development. Through the case studies, two different types of means of guidance playing the role of mediating factors between strategic objectives and operational reality emerged. The first concerned general guidelines providing a framework for how to think about product development. Using the learning vocabulary (c.f. figure 21 and the discussion in section 7.2), these frameworks seemed to play an important role in managing the creation, evolution, and 228 modification of individual and collective mental models of product development work. This kind of means of guidance greatly resembled a concept developed by Bowen et al (1994) that they name guiding visions. This term will be employed in the following analysis. The importance of such frameworks, constituting the most general form of means of guidance for the development work, was much emphasized both at the management and operational levels in the studied companies. The second type of means of guidance that was identified concerned the more detailed operational level, and more precisely the different performance indicators developed to measure the efficiency and productivity of product development work. That performance indicators are decisive for the outcome of operational work, and that greatest attention therefor must be paid to their evolution when lean production techniques are implemented has been strongly argued for in the relevant literature (c.f. Clark & Fujimoto, 1991 and Womack et al, 1990). The two identified means of guidance will first be analysed separately, beginning with the guiding visions. Then the interconnections between guiding visions and performance measurements will be discussed leading to the conclusion that they must be managed in an integrated way. 8.1 GUIDING VISIONS FOR PRODUCT DEVELOPMENT - ORIGINS AND CONSTRUCTION The search for guidance in operational work discussed by Bowen et al (1994) was reflected in the example of the work of design technicians introducing the analysis of operational development work in chapter seven. Prioritizing, choosing between a solution privileging ease of assembly or optimal functionality, and choosing between using or not using a new material allowing for a more complex form, knowing that this will increase production costs but reduce the number of components needed in a system, were all examples of situations where project managers, design engineers and technicians had to make rapid decisions based upon the product development strategy of the company. Bowen et al (1994) define guiding visions in the following way: "In the context of new product and process development, a guiding vision is a clear picture of an operational future, an organisational or project destination that serves as a referent and focal point for current decision making" (p. 59). They continue: "An appropriate end point [for a guiding vision] could either describe a future organic system or an external standard toward which the company can 229 strive." (p. 62). Bowen et al (1994) qualify the use of guiding visions as conceptual leadership. These clear pictures must be anchored in the global strategy of the supplier company. Examples of such strategic objectives in the studied companies were discussed in chapter six, c.f. table 13. The strategic issues of increased customer orientation, promoting learning at an individual and organizational level, and identification of and concentration on core capabilities all call for the development of clear guidelines in order to be operationalized. When discussing with project managers, design engineers and technicians, such statements as 'play the game of partnership' and 'be a force in product design' translated a consciousness of business guiding visions of customer orientation and product development expertise guiding operational day-today decisions. How, then, were guiding visions formulated and how were they elaborated? Two clearly expressed guiding visions for product development were identified in each of the case study companies. They were formulated by General Managers and Product Development Managers as an answer to the questions concerning basic ideas and priorities that guide the way the company has to work to achieve management's goals in terms of product development capabilities (c.f. appendix 2, questions 5.5 - 5.7). During discussions with operational staff, it was confirmed that these guiding visions were well disseminated in the organization, even though they were not always properly understood. In COF they were: A component design should be as simple as possible, and A component design should go beyond the proper function, quality, and cost of a component and reduce the global cost of a component solution through innovative functional solutions. In COR they were: Extension of product functions, i.e. development of mini systems should be a result of innovative design and not only a result of integrating assembly, and Ultimate prioritizing in operational tasks should be made according to customer satisfaction. 230 Through interviews and discussions with general managers, product development managers, and other functional managers in the two case study companies it became clear that guiding visions were generated largely as a function of external driving forces. In the language of the paradigm model they were in a way an action strategy for responding to driving forces, something that made them causal conditions in relation to the core category operational design. In the case study companies guiding visions were permanently elaborated by functional managers and general management in common using, as Bowen et al (1994) also proposes, all kinds of market and technical information (e.g. direct customer contacts, patent and technology scanning, and benchmarking with competitors' products). Both those situations where guiding visions clearly contributed to orient development work in a desirable direction, and those where operational people might feel confused about how to act were encountered in the case studies. An example of each will be discussed here. The first more successful one considers the guiding vision for developing components that provide more to the customer than its proper function, quality and cost. The second, where some problems were identified, considers the prioritizing between different parameters of cost and lead time in order to optimize customer satisfaction. 8.1.1 Guiding Vision: Reduction of Global Component Costs To be an expert in a specific field of technology and a partner in product innovation and development was a product development strategy shared by all of the studied companies. In COF one guiding vision corresponding to this strategy was to 'reduce the global cost of a component solution through innovative functional solutions'. In operational development work this meant that design engineers and technicians were invited to focus on a broad array of characteristics that the developed component ought to satisfy. This included functional performance1 , internal manufacturability, system fit2 , and potential for simplifying a system. If the first item was related mostly to product performance (both of the single component and of the system in which it was contained), the last three translated a cost concern in terms of reducing internal production costs, improving the ease of assembly of the component system in the customer firm, and reducing the number of components in a system. A successful example of this expert strategy was observed in the development of a new product in one of the dedicated teams at COF. Here, the consulting engineer responsible for the project was involved as an expert and development consultant in the customer firm (which was in this case a system supplier). His role was not limited to work on the functional 1 Such as capacity, strength, resistance to temperature and vibrations, etc. 2 The ease of assembly of the component into the system where it will function in the vehicle. 231 performance of the component family in question, but it also comprised active analysis of the production and assembly process in the customer firm in order to gain process productivity and quality improvements in the customer's processes by means of the characteristics of the new component. This could only be achieved through an analysis of existing solutions in the application environment and a mutual adaptation of the latter and the new replacing component. An objective in COF was to generalize this expert approach to more current development projects. However, in the dedicated team it was conducted in an environment where one person -the consulting engineer- had total control over all of the activities around the project, including product technology, process technology and commercial actions. When applying the guiding vision of reducing global costs through innovative solutions to all development efforts it would become dependent on the support of other concerned actors within the company. For example, a total cost reduction approach to product development would have to be supported by appropriate communication on the part of the sales force in their interactions with customers' purchasing agents. As Bowen et al (1994) underline, guiding visions are truly useful only if they are shared by all those working in a team and in support of a team. To achieve this objective seemed to be a very lengthy process, however. The problem was also raised by Product Development Managers during the last series of interviews, something that confirmed that the lack of rapid and relevant methods for analysing total component cost (including elimination of components, leading possibly to elimination of suppliers, and increased ease of assembly) was a shared problem in expert supplier firms at the time of the study. The development of an expertise based on total component cost should therefore be seen as a long-term project. Consulting engineers played an important role as catalysts of this guiding vision. Through their long-term projects, conducted in close collaboration with customers' development and research departments, they could slowly but surely develop this image in customer firms. If this guiding vision could not be realized in directly ordered components, an alternative line of action observed in the company was to try to take the opportunity of developing an innovative solution independently of customer demands and propose it 'off the shelf '. This strategy was practised also in COR and confirmed in the last series of interviews as a way of making customers aware of a broader spectrum of design capabilities that the suppliers could offer. 8.1.2 Guiding Vision: Prioritizing as a Function of Customer Satisfaction 232 The second example concerns the guiding vision of prioritizing operational tasks as a function of customer satisfaction. The difficulty of following this guiding vision can be illustrated by an example from COR where it was made clear that different customers accorded different importance to the balancing of a complete testing of prototypes and the reduction of development lead-time. One customer, A, preferred to test prototypes himself if he only had the component a few days earlier than planned. With another customer, B, the opposite situation dominated. On more than one occasion this customer accepted a longer lead-time in order to have the prototypes completely tested by the supplier. This different customer behaviour needed to be well known and understood by all the participants in the product development process so that the right decision leading to optimized customer satisfaction could be taken in every different situation. In this example, at least three different possibilities existed: • to always respect the lead-time, • to always complete prototype testing before shipping, • to let the customer decide in each individual situation. If a global strategic goal is customer satisfaction - as the guiding vision indicated, the latter option should be chosen. However, it was very likely that this choice, normally taken by a project manager, would be contested by other actors. Sales representatives would insist on the strict respecting of the contractualized lead-time - this was customer satisfaction in their view. Quality personnel would argue for the strict respecting of the complete testing of prototypes - they had difficulties in understanding that a customer could be satisfied with the company's offer without fulfilling this requirement. The lesson that COR needed to learn from this experience was that what satisfies the customer can only be determined by the customers themselves. It is very easy for employees to confuse their own professional priorities with real customer needs. Therefore, the guiding vision of customer satisfaction needed to be supported by means of bringing together different actors inside the company and customers. For example, management considered letting quality control personnel occasionally attend customer meetings as part of a training programme. Performance measurements also needed to be revised, as will be analysed below. 8.2 THE DEPLOYMENT DEVELOPMENT OF GUIDING 233 VISIONS FOR PRODUCT As it has been made clear earlier, there were different customer relationships and different parts developed in the studied companies. Customer relationships ranged from direct arm's length to direct expert supply, and concerning product development all companies worked simultaneously with detail-controlled parts, black box parts and supplier proprietary parts. Therefore, the guiding visions for product design and development had to integrate all these different situations. In both COR and COF product development work was divided by product family and customer so that a natural mix of different kinds of products, black box and detail-controlled parts, were designed by all design technicians. At COR this organization had emerged naturally. At COF an earlier organization that separated the design blueprints (that were used in the design interface both with customers and internally in the project groups) from the manufacturing blueprints (that were used in the internal production interface) had been unsatisfactory. In this organization, design technicians felt that they were divided into an A (product) and a B (production) team. To avoid a similar situation related to the design of different kinds of products demanding different levels of creative design skills, management opted for an integration of the design of different groups of products. This leads to an increasing task diversity that guiding visions must take into consideration. Detail-controlled parts or very simple design demands with a small financial margin cannot be developed with the objective of proposing innovative designs. Such projects need to be launched as quickly as possible with as low an indirect cost as possible in order to provide the cash flow needed for developing more complex products. The vision guiding such design tasks, clearly expressed by the Product Development Manager, was one of using standardized techniques and devoting as little time as possible by choosing the simple solution. The analysis above indicates that both the messages of individual guiding visions and their differentiation must be well understood at the operational level. This leads the analysis towards the question of how to deploy guiding visions in the product development organization. The role of guiding visions is to provide the link between business strategy and operations; the critical parameters of guiding visions are their visibility, quality and acceptability (Bowen et al, 1994). Moreover, guiding visions should be defined in operational terms rather than as static goals. "Profit-based numerical goals do not translate into innovation in the laboratory or on the factory floor, goals based on competitive issues or internal capabilities do" (Bowen et al, 1994, p. 62). New guiding visions constantly need to be developed in an organization, for example, guiding visions for developing a perspective of seeing product design constraints and communication 234 links as learning opportunities were at the embryonic stage in both of the case study companies at the moment of the study. Besides the question of how to articulate a vision corresponding to these objectives, the question of how to deploy it and make it accepted therefore preoccupied managers. An understanding of the process of deploying guiding visions could be developed by looking at the example of the well accepted vision of total cost reduction as a consequence of innovative design. Four points were identified as decisive in the successful deployment of this vision: •• Firstly, strategic goals and their raison d'être in the company were explicitly communicated to employees. This was done through memoranda, notes, information meetings on the part of general management, training, etc. •• Secondly, the top-down information diffusion approach was complemented by a bottomup return of information where management ensured a comprehensive analysis of operational situations based on feed-back from employees. •• Thirdly, a key to successful deployment of guiding visions was for management to prove the seriousness of communicated guiding visions through concrete interventions in the organization. For example, the creation of a specific dedicated team with a product champion in the form of the consulting engineer materialized the realization of the guiding vision of promoting innovative design. •• Fourthly, employees saw a concrete example of success for the guiding vision through the fact that the project advanced and the customer seemed to adopt the new product solution. There was a need for successful examples in order for the mental models of how to look at product development to change. By examining, through the case studies, how an organization really operates, an assessment of the match or mismatch between the official discourse in a company and the operational reality could be made. If a mismatch occurred, as in the example of the customer satisfaction guiding vision discussed above, a first approach could be the verification of the application of the points identified above. Finally, if, as argued in the previous section, guiding visions are developed largely based on customers' expectations, it is important to ask how driving forces from customers are received and who receives them ? The case studies indicate that they are often received through those that are in direct design contact with customers, i.e. product development managers, project managers, design engineers, and design technicians. In order to develop guiding visions that are coherent with the customers' expectations, relevant feed-back of the information collected at all levels must reach the designers of the guiding visions. This information can consist of direct customer input, product technology information contained in specifications, mock-ups and internal design data bases, and material and process technology scanning. Thus, different 235 support structures also play an important role in deploying guiding visions. As Bowen et al (1994) underlines, such visions do not need to originate at the top of the corporate hierarchy, but can spring from creative minds anywhere in the organisation. If this is true these signals must be captured, analysed, and developed which includes spreading them to all concerned actors and integrating them in a company's organizational routines. 8.3 GUIDING VISIONS - CONCEPTUAL LEADERSHIP Bowen et al (1994) refer several times to guiding visions as a way of exerting conceptual leadership: "Guiding visions are the tangible output of conceptual leadership" (p. 86). However, they do not further specify what conceptual leadership means. When relating the concept of guiding visions to the analysis of product development work undertaken in the previous chapter, it seems clear that they have strong links with the cognitive part of development work. The example of how COF succeeded in becoming a development consultant, analysing also the customer's internal manufacturing process, prove that guiding visions actively contribute to the altering of individual and collective frameworks for how to look at development work. The guiding visions introduce a reversed and complementary approach to the reflection-inpractice discussed in section 7.2 namely that of an influence on and a modification of mental models that precede practice - a cognitive assessment and elaboration of design activities through leadership. Danielsson (1986) develops an interesting perspective on leadership that allows for a clarification of what conceptual leadership is. His model can be developed in three steps: •• Individuals that work in a specific profession normally have quite a clear picture of what the dominating problems are and what the right solutions to these problems could be. This was also true for General Managers, Product Development Managers, and design staff in the case study companies. •• These pictures can be called basic supporting ideas and they guide the actions that are undertaken both to manage and execute operational work. An example of such a picture that was undergoing change both in COF and COR was that customer contacts in design should be managed only by project managers. •• Basic supporting ideas are important for designing and elaborating action plans, but in order for actions to be undertaken there is a need for releasing factors. As discussed earlier, guiding visions were found to be driven by external forces such as customer pressure and specific competition emerging as a result of the reconfiguration of the entire 236 automotive industry (tier dilemma). These external forces correspond to the releasing factors in Danielsson's (1986) model. Based on Danielsson (1986) conceptual leadership can be summarized as interpreting, influencing and ultimately integrating these three steps. It is firstly about detecting, analysing, explaining, and communicating releasing factors in an organization. In COF, for example, explaining and communicating was done through different dissemination mechanisms as described in section 8.2. It is secondly about developing and deploying new basic supporting ideas and managing the 'unlearning' of old ones as a function of releasing factors. Both Danielsson (1986) and Kim (1993) argue that basic supporting ideas that have had an important impact in an organization persist over a long time period - often several years or even decades. Competing new ideas also need a long time to be developed and achieve cogency. In the example of the guiding vision of customer satisfaction, the problems encountered were related to standardized behaviours formed by existing performance measurements that partly were becoming outdated. Thus, Danielsson's (1986) conclusion that conceptual leadership over basic supporting ideas must be exercised continuously proves its relevance. As guiding visions are about preceding actions through influencing mental models they can correctly be seen as the tangible output of conceptual leadership. 8.4 PERFORMANCE INDICATORS Performance indicators were identified as influencing operational decision making in several situations and in different ways in the previously undertaken analysis of product development work, and guiding visions. Ideally, performance indicators should be the explicit translation, in measurable terms, of guiding visions (Bowen et al, 1994). Sometimes, however, they are not well aligned to the general strategic discourse in a company, something that causes incoherence between management's objectives and the outcome of operational work. Hanna (1988) presents an example of a company where in spite of formal rhetoric and much top management attention concerning profitability, the dominating operational priority and the domain where much attention was paid to performance measurement was market share. The underlying logic was that profitability would follow sales volume. But the company lacked an analysis of what segments were really profitable which meant that growth was encouraged also in non profitable markets which, in turn, was disastrous for profitability. The less the company was profitable, the more attention was paid to volume in a vicious circle. This example shows how performance measurements can dominate over non-anchored guiding visions. To change 237 the situation a new performance monitoring system would have to be developed in line with the profitability guiding vision. As Hanna's (1988) example indicates, too little attention is often paid to the importance of performance indicators. Defining performance indicators in different functional areas is therefore an extremely important issue as they must reflect both the global strategies of the company and more detailed functional guiding visions, for example those influencing product development work. Performance indicators can be oriented both towards evaluating individual performance and more global performance at a functional level. In both of the case study companies several performance indicators were related to the development and launch of new products, for example the number of new products released, and the turnover generated by new products compared to total turnover - measurements that were dispatched per customer, per product family and per type of output, i.e. only prototypes or mass production launch. There were also measurements concerning the turnover generated by new customers compared to total customer turnover. These performance indicators were in line with a strategy of maintaining and reinforcing an expertise in product development. Then there were different efficiency measurements. Product Development Managers often had quantitative objectives, for example to increase the new product turnover or the number of launches with 10% over a one year period. Another objective could be to reduce the number of design studies and prototypes that did not result in launched products. At a first glance this might seem to be a reasonable criterion for improving productivity in product development. However, it might lead to the contrary if, as a consequence, design technicians become reluctant to test out new concepts and to innovate in functional solutions. A guiding vision common to the two case study companies was that product development managers and project managers ought to become more selective in the choice of development projects. Generally this was translated at the performance measurement level as prioritizing the most important projects in terms of potential turnover. Such an objective has both a financial and a technology/organizational side and it was present in both of the case study companies. If a company follows a strategy of building core capabilities it should select projects in line with, for example, the guiding vision of developing small systems as a function of innovative design. In this context there was a problem of determining the 'right' product or process diversification that was in line with, or could be useful for, developing existing or emerging capabilities. Otherwise there was a risk that diversification could split the design organization and distort more than stimulate learning. The main risk with giving priority to large volume affairs was that they still often concerned detail-controlled parts and direct arm's length relationships that could undermine an expert strategy. The main risk with prioritizing black box components and 238 products with an important design content was the important lapse of time between the cost of the development and the return on this investment through a mass produced component. It is interesting to come back to the example of prototype testing and delivery discussed above and look at the role of performance indicators in it. In the concerned supplier company, a global strategy and a strategic discourse was to listen to the customer and try to optimize customer satisfaction. At the same time, however, two performance indicators in product design were the respect of development lead-time and the percentage of completely tested prototypes before delivery. As these criteria were the closest to the operational staff, the latter would give greater importance to these criteria than to considering what real customer satisfaction could mean. Satisfying customer A and delivering untested prototypes earlier than planned would penalise the testing criterion. Satisfying customer B and overriding the lead time while competing testing would penalise the lead-time criterion. In this example, satisfying customer A and B would give negative scores on the performance criteria. Two reasons behind the discrepancy between strategies, guiding visions and performance indicators were found. Firstly, if the latter was imposed by one specific customer (Ford's Q1 quality standard for example proposes a set of obligatory indicators), there was a strong chance that they were not relevant for all customers. For reasons such as those discussed above this could result in deteriorated customer satisfaction even though performance criteria were fulfilled. Secondly, if performance indicators were developed within each function without a transversal reflection on what they could mean for other interrelated personnel, there was also a risk of contradictory behaviour between different project participants jeopardizing product development performance in the supplier firm. 8.5 MEANS OF GUIDANCE - SUMMARY AND CONCLUSIONS The analysis in this chapter has allowed for a clarification of the leadership dimension in product development management. Guiding visions, representing a conceptual leadership, were found to be very important for deploying, sharing and developing a long-term adhesion to a company's strategic goals. The other part of the leadership dimension, performance indicators, were found to influence the operational side of development work and sometimes enter into conflict with the more cognitive side influenced by guiding visions. Through the analysis of a successful example of the development and deployment of a guiding vision, four factors of success were identified; explicit communication of strategic goals, 239 analysis of feed-back from the operational level, concrete intervention in the organization related to resource allocation, and 'return on investment' in the form of examples where pilot actions proved to be successful. Even if these recommendations are respected, difficulties in managing guiding visions can occur if these conceptual messages are not backed up by coherent performance measurements. Guiding visions are only credible if they are reflected in systems and performance measurement. An important role played by the analysed guiding visions was to encourage inter-functional activities in the product development process. In the light of the analysis of integration and coordination undertaken in section 4.2. it therefore becomes clear that attention must be paid to how guiding visions are expressed. This is also emphasized by Bowen et al (1994). If guiding visions should inspire ownership across disciplines and functions, the language in which they are expressed must be consciously selected. If a guiding vision aims at promoting interfunctional collaboration for example between product engineering, process engineering, and marketing sales, it must be articulated within the common cognitive ground that exists between these actors. Concerning performance measurements, the analysis leads to the conclusion that too much focus on performance criteria runs the risk of leading to a distortion of the deployment of strategic goals and even a blindness for real problems. Individual performance criteria will often have a stronger impact on employees' actions than guiding visions even if they are in opposition to a company's global strategy. It seems that this happens most often with external performance indicators which are not always grounded in a company's culture, profession, or in management's internal guiding visions or strategies. When reflecting on means of guidance and when elaborating strategic plans and related guiding visions managers must ask themselves what role existing performance criteria play, what they represent, and what the reactions to and operational results of changing performance measurements will be. The analysis emphasized the chain of consequences that performance criteria create that must be analysed in order to optimize the means of guidance for product development work. From the problems discussed above it becomes clear that guiding visions and performance criteria must be aligned. If strategic goals are to be reached, performance criteria need to be consequently developed and adapted once guiding visions have been determined. 240 9 DESIGN SUPPORT STRUCTURES During the analysis of operational design work in chapter seven, and that of means of guidance in chapter eight, the issue of different design support structures -specifications, data banks, group work, spatial organization of work areas, and information systems- were addressed. In the data analysis of different concepts, design support structures correspond to the definition of a context to a core category given by Strauss & Corbin (1990, p. 101): "The context is the particular set of conditions within which the action/interaction strategies are taken to manage, handle, carry out, and respond to a specific phenomenon [in this case operational design]". Some of these factors have already been analysed in the previous chapters: • In section 7.3, group work was analysed and it was shown that an organizational action of co-locating design technicians, i.e. project participants belonging to the same function, was beneficial for innovative thinking and for reducing problem-solving lead-time. Furthermore, it was argued that in order to improve efficiency during formal group work, support in the form of meeting reviews (ideally interactive), performance indicators, and formal procedures for integrating new solutions to problems in the shared knowledge base of the company had to be further developed. Managers anticipated an important role for computer-aided information transmission systems in this context. • In section 7.4, the importance of physical co-location of project participants belonging to different functions for collective learning and evolution of the supplier company's product offer was emphasized. In section 7.5 the same conclusion was made concerning the relationship between consulting engineers and the rest of the design staff. The case studies identified this kind of formal organizational action as important for the creation of a common cognitive ground, identified in chapter four as crucial for improving coordination in product development. Group work structures and aspects of physical reorganization will therefore not be further addressed in this chapter. Attention will be focused on the other identified structural factors, i.e. specifications, and support structures for information transfer. In addition, the issue of technology scanning related to the increasing importance of process technology in the development of new product technology (c.f. section 6.3) will be addressed. 9.1 SPECIFICATIONS 235 A central element supporting and guiding the development process is the product specification that will be defined here -in accordance with the perspective of the interviewed design engineers and technicians in the case study companies- as consisting of blueprints and different text documents. One specific case was the subject of particular managerial interest, namely the role of specifications in black box engineering. The analysis here will therefore be focused on this situation. In black box engineering, the text documents mentioned above contain a variety of information. In the case study companies, more precisely, they contained: • The scheduled order quantity for the finished product, • The design study lead time, • Quality demands (for example ppm level), • The unit price for the ordered quantity, • A written functional description, • A description of the liaison between the part that will be designed and the interfacing parts, • Technical data such as capacity, strength, temperature, vibration and chemical resistance, • Internally generated budget and cost price, • Commercial and technological interest -such as potentials for entering a new market, or developing new product technology possibly leading to patents-, and finally • Competitive situation. As the above description illustrates, the specification as defined by product development staff is already the result of a very large number of interactions and adjustments. Even though the specification was often seen by design technicians as one relatively fixed document, the observation in the case study companies of the activities taking place between the initial receipt of a specification and the moment when it was considered as completed, indicated that specifying is a process taking place at the systemic level of operational design work. It always needed further interaction with the customers, cross-functional consultations inside the supplier company, and often also the advice of experts such as sales engineers or consulting engineers. These observed tendencies have strong parallels to recent research on the use and management of specifications. Kaulio (1996) argues that in opposition to the traditional view of specifications as a single document that someone else has written, that is hard to take in and that limits creativity, the specification is a series of generations of documents that change both over time and in relation to the users. Moreover, Kaulio (1996) found that its primary role is that of being an arena for cooperation among the actors in the design process. Kaulio's (1996) research focuses on the specification of complex final assembled products (e.g. a new vehicle for street-care, a new ambulance for paramedical treatment), service centred products (e.g. administrative computer systems), and simple single piece components (e.g. a new coffee and 236 teapot for the cabin crew at SAS). The present research confirms these results also in the case of designing car components -especially the idea that the specification is an arena for cooperation. The specification was in fact found to be not only a support structure for the technical problem solving at an individual level, but a vehicle for inter-functional and intercompany collaboration improving coordination in different interfaces and creativity in product technology through the exploration of several different ways of resolving problems. These tendencies were confirmed in the last series of interviews where one Product Development Manager described the specification process of black box parts as "an open field for technical negotiation". He emphasized the impact on process technology that a less rigid product specification leads to; technology shifts in process technology are becoming more and more frequent. Black box engineering necessarily redefines the role of the specification - as the data indicated, the specification process must become more interactive than traditionally was the case. Moreover, design technicians had to take an active part in this process. Some difficulties were related to this changing role. In the case study companies, design technicians sometimes complained that specifications were incomplete, something that considerably delayed the operational design work in their perspective. However, behind the expression 'the specification is not complete' there was the question of whether the specification ought to be more complete, i.e. if information has been let out by error at some stage in the transmission of the specifications, or if the specification was intended only as a guideline that the customer didn't wish to specify further. This problem was clearly related to the transformation more towards general guidelines and guiding visions than detailed instructions in the customer-supplier interface in product development - identified as black box engineering. Karlsson & Åhlström (1996) make similar observations in their case study of an international manufacturing firm producing mechanical and electronic office equipment: "Requests for detailed specifications were regularly observed in the different projects. [...] The specifications that were requested should contain fairly detailed descriptions of the product. Specifying the functions the product should perform was not enough" (p. 290). Karlsson & Åhlström (1996) further observed that design staff had a tendency to project the responsibility of generating specifications onto other functional staff, e.g. marketing. In other words, the transformation from this traditional view of specifications that they should contain concrete descriptions of the desired design principles of a producttowards a limitation to visions and guidelines as specifications was, as in the present research, perceived as difficult to handle by project members. Without further specifying what this means or how it can be done, Karlsson & Åhlström (1996) suggest giving continuous concrete expression to visions in order to move the product development process ahead. In the present research, two factors facilitating this were identified (c.f. chapter eight, section 8.2): 237 • Management must prove the seriousness of communicated guiding visions through concrete interventions in the organization, • Successful examples related to the application of a guiding vision must appear - for example in terms of improved possibilities of satisfying performance indicators. The present research provides a more balanced picture of the process of specifying in black box engineering. Indeed the study reveal a general problem for design technicians to determine in which category different specifications belonged, but project members at COR and COF did not have any particular problem with inventing new product solutions out of only partially framed specifications. What was perceived as a problem was knowing when a design demand concerned a black box type of product or a more detail-controlled one. Several examples were quoted when a design problem was seen as pertaining to the first category, but where the reaction to proposed solutions would be negative on the part of customers with the pretext that the specifications had not been respected. In order to reduce these problems, clearer information on the part of the customer could be asked for. Faced with an ambiguous situation, a general tendency observed in all of the studied companies was that suppliers had to become more proactive in their customer relationships in order to obtain the information they needed to do their work correctly. Moreover, the experience of Product Development Managers at COR and COF was that the more active the supplier was in searching for information on his own initiative, the greater the chance of profiling the company as an expert supplier. The internal performance of the supplier company in terms of information transmission and a relevant understanding of transmitted information was also informally assessed in black box engineering. A false interpretation of customers' demands could sometimes be a result of too many interfaces in the information transmission process (c.f. section 7.5). Once these obstacles had been eliminated, however, the creativity and capacity for new suggestions was not lacking in the case study companies. In COF, for example, around 20% of the design studies undertaken yearly in the company resulted in new launched products. What actually happened in black box engineering was that a set of different specifications was developed based on the initial framing guidelines transmitted by the customer. This contained nothing revolutionary in itself; design technicians generally developed several product solutions and prototypes for every new development project, so the 'raw material' for such a line of action already existed. What was changing in both COR and COF at the moment of the study was that these parallel solutions (in the form of blueprints, prototypes, and as complete a set of text documents -described above- as possible), were successively presented to the customers. Traditionally, the different developed solutions were evaluated inside the supplier company and only the final chosen solution was presented to the customer. In the new approach, slowly 238 emerging in some black box projects, customers intervened much earlier in the assessment of component solution propositions. Some solutions were abandoned depending on customer feed-back, and the remaining ones were refined until the process converged into a final and single component solution. The process described above is only an attempt to conceptualise observations of the way design engineers in a very limited number of cases seemed to work. What is interesting, however, is that a recent case study (Ward et al, 1995) of the product development process at Toyota points in the same direction. At Toyota, capable suppliers are provided only with approximate specifications or targets. Then, they work with sets of solutions which are gradually narrowed, converging to a single solution. The application of this process at Toyota has resulted in less time spent in formal coordination meetings, and less need for very frequent communication also at the engineering level. The important thing is that broad band and direct communication channels exist between supplier and customer design technicians so that consultation can take place whenever needed. Ward et al (1995) found that this practice was refined over time: the more the supplier became integrated with customers, the more he would be able to anticipate customer needs. Thus, the set-based approach represents a contradictory practice compared to that of minimizing the number of design studies and prototypes that do not result in launched products, a performance criteria generally dominant in the studied firms (c.f. section 8.4). If a transition towards a set-based practice takes place, the related performance indicators must therefore also be adapted. Besides the possibility of freezing hard specifications later, the observed use of the set-based approach confirmed its presumed possibility of promoting innovation. When working in this way, customers became informed of a broader spectrum of the supplier's design capabilities. This made it possible for the supplier to further interest them in the possibility of proposing solutions that went beyond the perfection of single component performance - for example mini systems or different technology for resolving a specific functional problem. These observations confirmed a somewhat contradictory practice compared to the general perception of 'lean' product development. Suppliers were indeed involved early in the development process, but hard specifications tend to be fixed as late as possible to promote innovative solutions. Thus, instead of talking about early design involvement, the case studies suggest a permanent design involvement and a late component procurement. The main difficulty in black box engineering was the trade-off between the perpetual change of specifications during the development process -leading to the development of a set-based approach, and the demand for short development lead-time that weighed heavily upon 239 suppliers. In order to facilitate product development work under these conditions, the field observations indicated that it might be useful to divide the design process into two parts: • The functional concept. This is the product technology that will solve the functional problem that the customer presents; • The dimensional definition. This consists of the dimensions and the form of the component. If suppliers learn to work with 'sets' of functional concepts according to the principles outlined above in a first phase and only subsequently devote time to the dimensional definition (after intensive exchange of information with customers and interfacing suppliers concerning the functional concepts), suppliers would be able to reduce development lead-time and redundant work, and at the same time respond to the evolution of specifications in a flexible manner. This vision of design work in two phases might also contribute to a concentration of the design efforts on the functional solution, something that was asked for by several design technicians in the case study companies. As discussed in section 7.5, they considered that they often had too many design studies going on simultaneously and work tasks that were too fragmented - for example interrupted by changing priorities - to fully concentrate on more complex design studies. In order to operationalize these ideas, there is a need for flexible procedures allowing for the rapid realization of dimensional definitions once a functional solution has been defined as sufficiently interesting for pursuing the development in a specific direction. Design technicians at COF mentioned the possibility of working with standardized modules or pre-defined standard geometries that it would be possible to combine in a flexible way in order to speed up development time. However, increasing the rapidity at this level is useful only after having fully exploited coordination between different actors working on a component system. This will be further analysed in the discussions of computer-aided systems for information transfer below. 9.2 COMPUTER-AIDED SYSTEMS FOR INFORMATION TRANSMISSION Project managers and design technicians in the two case study companies expressed problems related to the time that it could take before a return on requests or information was received from customers. The demands on system coordinators (system suppliers and/or carmakers) concerning the coordination and optimization of the individual product development processes in different supplier firms that contribute to a system have been increasing steadily over the last years (c.f. e.g. Automobile Management, 1996). In order to facilitate the optimization of the internal development processes in different suppliers engaged in simultaneous engineering, it is first of all very important that each actor is informed about who the other participants in the 240 system are and what their respective roles are. It was clear from the case studies that this was not always the case. Secondly, all system participants ought to be informed about the modifications undertaken by everyone involved, and, ideally, as close to in real time as possible. The research findings indicate that it could be useful for suppliers participating in the design of a system to put more pressure on the system coordinator concerning the rapid diffusion of all information treated in the individual binary relations between the coordinator and individual suppliers. It was the conviction of the design engineers and technicians in the case study companies as well as the Product Development Managers in the last series of interviews that this would reduce the number of late modifications (when a dimensional definition has been initially accepted), shorten the periods of uncertainty during which the supplier doesn't know if his solution will be accepted, and prevent suppliers for pursuing work on a concept that would become obsolete following simultaneous evolution and increased precision of the system specifications. Through the research perspective focusing on specialised suppliers, the problem with simultaneous engineering was not viewed from the perspective of the system coordinator, but from that of singular system contributors. A central finding based on data from all of the studied companies was that the problem was not to simultaneously engineer products, but to simultaneously manage the related information flow. The system coordinator does not have the sole responsibility in this process. Suppliers ought to specify their needs more clearly and take their own initiatives for improving the situation. One way of coping with these problems, suggested by the interviewed Product Development Managers was enlarge the demands on the information transmission that CAD systems handle. This vision could be one of a computer-aided system for learning (Goodman & Darr, 1996). They identify three benefits with such a system that come very close to the needs expressed by the design technicians in the studied companies: • Fast and efficient communication, • A memory shared by all organization members, • A mechanism whereby multiple members can dynamically exchange solutions and update solutions to problems1 . The central idea related to the problems in simultaneous engineering and addressed through the case studies and interviews was that the CAD systems could be supplemented with more qualitative information that goes beyond blueprints and material specifications, and, moreover, 1 The software discussed by Goodman & Darr is of Lotus Notes type, i.e. groupware which enables users of network hardware to share databases. 241 allow for informal dialogue between design engineers and technicians working on a system. An information system that simultaneously treats qualitative and quantitative information would also be one way of facilitating the development of a set-based approach to product engineering and the realization of design studies in two phases -functional solutions and dimensional definitionsdiscussed in the previous section. As argued there, these approaches will be profitable only after having achieved integration between different actors working on a component system something that this kind of information system might facilitate. Moreover, this kind of system might facilitate the constitution of inter-functional and inter-company micro organizations for example for the life time of a specific project. This redefinition of the computer-aided systems that support the design process will not take place without difficulty. However, due to the lack of standards in the European auto industry concerning the existing CAD systems -for example, Renault, PSA, Opel and Volkswagen all use different systems- CAD experts that were interviewed anticipated a future redefinition and harmonization of the systems, something that might give an opportunity to consider expanding to more qualitative information exchange and instant updating of data. The implementation of such a redefined system should only be undertaken after the fulfilment of three conditions (Goodman & Darr, 1996; Duimering et al 1993): • The system must be legitimized in the organization and the latter must be able to make use of the information; • The users themselves, i.e., in this case, design technicians, engineers and project managers, must specify the demands on the system; • The system must allow for an alignment of goals of different participants. The first point means that if the need for a new or modified information system is expressed by the future operational users of such a system themselves, the probability of success increases. If it is imposed by management, or, for example, customers, it might easily hit a wall of resistance to change. In both of the case study companies, operational people showed a certain resistance to change towards everything that did not directly stem from their own initiative2 . Moreover, before undertaking modifications in the information system, an assessment and, if necessary, a modification of the internal organizational structure can improve the effective use of the possibilities that the new system might offer. This rejoins the discussion in section 7.5.1 where the central importance of the internal project structure and the configuration of the customer-supplier interface for efficient information transfer were emphasized. For example, a powerful combined qualitative and quantitative 2 The resistance to change was of a 'normal' intensity compared to the researcher's previous and current experience from consultancy and research. 242 information transmission system will be of use only if customer-supplier information exchange at the individual operational design technician level is allowed for. Consequently, this statement implies that a partnership relationship will be a requirement for developing more sophisticated information systems in the sense discussed here. The importance of assessing organizational and relational structures before implementing new information technology and computer-aided manufacturing systems has also been much argued for in the literature (c.f. Duimering et al, 1993; Klein, 1991). Karlson's (1994) argument that simply communicating more will be of little interest if the interlocutors understand each other poorly is also relevant in relation to information system development. Hence, one of the central arguments of Duimering et al (1993) and Karlson (1994) is that information transmission does not equal integration and that, therefore, organizations must be redesigned and a common cognitive ground between specialists with different professions must be developed before new information systems are implemented. The second point rejoins the first, but goes one step further. In fact, legitimizing a new information system is not enough for making it function properly in operational use. In order to make sure that it functions at least in the majority of day-to-day situations that the future users encounter, Goodman & Darr (1996) argue that the latter should actively participate in the specification process and on an equal basis. An example illustrates this issue. In both of the case study companies a design database existed, but it only contained design studies that had been launched for mass production. Product development managers and project managers, who had participated more in the design of these data banks than design technicians, focused their interest principally on design studies that were launched in mass production. Design studies that remained at the prototype stage were of secondary interest. For these managers, the central importance of all design efforts, including sketches and preliminary outlines, for developing optimized functional solutions by design technicians, was not obvious. From the perspective of the design technicians, the definition point for entering a study in the database i.e. when a study and prototype was transformed into a mass produced component- was simply not the best. Their argument was not to enter every idea and sketch into the data base, but to allow the possibility of entering projects that they identified as important to share with their colleagues on a formal basis. Finally, the third point specifically emphasizes the role that an information system might have as a vehicle for integrating dominant and non-dominant disciplines. In section 7.4 it was argued that this was an important objective in order to reduce problems and false priorities due to a lack of inter-functional understanding. As will be further developed in the following chapters, physical systems such as a design database therefore play a central role for deciding how 243 learning will take place in an organization and what learning will occur. In this way, support systems also ultimately influence what core capabilities a company will develop. Potential difficulties in the expansion and redefinition of the information exchange according to the case study based ideas outlined above were anticipated by the interviewed Product Development Managers in the last series of interviews. The main concern was the reliability of the information transmitted by different participants that would be a function of the trust that the participating firms have in each other. This trust will, in turn, be a function of how well different individuals (design technicians and engineers, project managers) know each other and of previous behaviour of different actors towards each other (c.f. section 6.4.4). To cope with this problem, it seems necessary to refer to the first point discussed above and use the information system in accordance with the customer relationship and the related organizational situation that is at hand at a given moment. The information system should be applied only between firms where the design people already have personal contacts. As these relationships evolve and are spread to a larger number of participants (according to the generalization of the partnership approach) the development of information systems will probably follow. Personal ties will motivate the information exchange. Moreover, the problems of delays and lack of interface information is common to all participants in a multi-firm development project, something that probably will increase the probability of finding a common interest in developing this way of working. Another anticipated problem was related to the fact that it can be both strategically risky and politically embarrassing to acknowledge problems. In fact, the kind of information system discussed here will very quickly unmask problems in the participating firms if, for example, the lead time for a design specification expected by the other partners is exceeded. This problem is confirmed in the study of Goodman & Darr (1996). Relating this problem to the theoretical analysis of buyer-supplier relationships (c.f. chapter four, section 4.1) it seems that only a true partnership approach can reduce this negative effect. In such an ideal context, a company having trouble ought to inform the other development partners in order to make it possible for them to assist in finding a solution. The competition between complementary suppliers identified as the tier dilemma (c.f. sections 6.1.3 and 6.1.4) still makes this problem difficult to manage, however. Comparing with the Japanese experience (c.f. the example of Mazda's split of supplier business in section 6.4.4) it seems that it is up to the carmaker, i.e. the global orchestrator of the development process to define rules making progress possible in this area. 9.3 TECHNOLOGY SCANNING 244 A constant flow of information concerning technological evolution and development was quoted as an important input in operational design by the interviewed engineers and technicians in the case study companies. Both information on product and process technology was important and there was a need for specific scanning activities beyond the core technologies used and analysed in operational development work. Therefore, a specific organization of scanning activities seemed to be a necessary support structure for developing innovative product solutions in the functional areas covered by the suppliers. Technology scanning can be defined as the systematic exploitation of technological information (Jakobiak, 1991). Jakobiak (1991) emphasizes that technology scanning implies three different actors: the observers, the analysers, and the decision-makers. He further specifies that technology scanning comprises scientific information (laboratory research, theoretical data), technical information (patents), technological information (manufacturing processes, product technology), technicaleconomical information (capacities, market opportunities), and economic information (sector statistics, macro economic data). That technology scanning is important for development of new products was made clear through the case study of COF and its technical plastic component business3 . Design engineers and technicians in COF identified new process technologies and new materials allowing for change and evolution of product functions as one of the principal sources of product innovation in plastics. Examples of production technologies that had allowed for the design of much more complicated products over the last years were gas injection, extrusion, and biinjection. However, a problem with technology scanning was expressed in this context. In fact, technology scanning was sometimes perceived as insufficient by design engineers and technicians. The plastic design staff at COF wished to know more, and as rapidly as possible, about moulding processes, newly patented process technology, and new material combinations. They also wanted to be better informed about quality problems that had appeared with pioneer users of specific process technologies. Thus, beyond the traditional activities of technological information scanning in data banks, specialized and summary journals, specific publications (including dissertations) from laboratories, universities, and other research institutes, organisms specialised in information searches, patent journals and archives, etc., the scanning activities that operational design staff were looking for comprised benchmarking with users of new production technologies both inside and outside the auto industry. 3 Generally speaking, technical plastic parts are an important domain for expert suppliers (c.f. De Banville et al, 1997). 245 Besides this need for widening scanning activities, a more profound analysis based on focused questions revealed that the problem was more one of exhaustively and rapidly disseminating the results of technology scanning activities than the efficiency of the scanning itself. The present analysis will therefore focus on this part, while considering that the choice of information to be collected, the organization of the very scanning activities, and the tools used for scanning were satisfactory. The productivity of scanning activities was found to depend on efficient dissemination of the obtained information in order to facilitate analysis by operational design staff and decision-making at the level of Product Development Managers and even General Managers. The organization of this dissemination and feed-back remained for the most part to be developed, however. Design engineers and technicians were thinking in terms of a computer-based information system for answering this need. They would prefer an interactive system in order not only to facilitate access to information but also to make discussions, comments, and questions concerning it possible and available to the entire product and process engineering organizations, as well as to manufacturing. In a true cross functional perspective this information should also be made accessible to operative staff in marketing and purchasing. It is important to note that such a system, in contrast to a support system for design, ought to be isolated from customers and competitors. Information related to scanning of new technology and patents typically pertains to the long-term strategic orientations of a supplier company. It might seem exaggerated to develop a computer-based information system also for technology scanning. However, the field interviews and observations in the case studies, especially in COF, indicated that an informal oral transmission as well as a more formalized transmission through reports and meetings was insufficient for keeping different actors rapidly and systematically informed of new product and process technologies and their different aspects. Oral transmission ran the risk of distorting original messages and not reaching all those that might have a potential interest in the information. Reports had a tendency of ending up in piles or closets instead of circulating among the concerned people. Moreover, the report format did not seem to invite a joint discussion between design staff, let alone between people from different functions. Meetings, finally, were not always attended by all participants. As discussed in chapter seven, section 7.3.2, formal meetings did not always function in a very efficient way; participants did not express themselves very vividly and seemed to have difficulties in laying aside their momentary preoccupations in order to participate actively in discussions of common interest. A computer-based system could reduce or limit these negative effects through its possibilities of real time reaction to information, a 'coffee-room' style discussion of ideas and opinions that would also be stored and be available to consult in the future. In fact, this was still an example where there was a need for creating a flexible and temporary micro organization for resolving a specific problem or realizing a specific activity. 246 In the case study companies, technology scanning activities were mainly conducted by product development managers, consulting engineers and sales engineers. Besides the demands for a better diffusion of technical information, the productivity of scanning activities increased if the people responsible for the scanning also made an effort to bring operational R&D staff such as design engineers, project managers and design technicians closer to the source of the technical information - for example the designers and developers of new process technologies and materials. In cases where design staff had been able to integrate directly with the developers of a new moulding technique, and discuss the new technology related to concrete design problems, the assessment of the technology in question, and, ultimately, the decision whether to further investigate it, had been much facilitated. A better knowledge and a more systematic exploitation of patents were also asked for by design technicians. Patent information (i.e. concerning patents obtained both by the employees of the company in question and by its competitors) needed, like technology scanning information and operational design information, to be transmitted between design technicians as close to in real time as possible. In the case study companies, it happened quite frequently that a design solution was dropped because it infringed patents held or applied for by competitors. Patent scanning was found to be an important and rather particular activity that needed an extensive knowledge of different functional solutions that had been patented over the years -by the company in question and its competitors- in a more or less wide area of functional applications. In COF this area was so wide that the patent scanning activities needed to employ a specialist. The case studies indicated that the more patents applied for within a company and by its competitors in the concerned business, the more relevant knowledge and rapid updating of patent information was important for design engineers and technicians working in operational design. At COF, a project of creating a check list, or dictionary, of the basic patents that the company possessed was ongoing. The objective was to facilitate the comparison of new ideas with existing patents, centralise new product solutions around these patents, and avoid redundant design of solutions that either already existed or could be susceptible to infringing competitors' patents. The patent dictionary was also aimed at facilitating integration of new staff as quickly as possible into the design process. The case studies indicated that in order to capitalize on the information collected through patent scanning activities and contained in patent data banks, patent management activities ought to be better integrated in the design function. This could begin with as simple a measure as co-locating patent scanning staff and design technicians. 247 Another factor, related to support for the technical/scientific side of product development was that of technical calculations. In the case study companies, a theoretical anticipation of product performance was related mainly to the use of formalized calculations and calculation software in the dimensionalization of components. One of the interviewed Product Development Managers described the problems of defending an expert position in projects where technological complexity increased or quality requirements (for example tolerances) were made more stringent, due to a lack of rigorous calculations. If the dimensionalization was based only on experience and extrapolation of already functioning solutions, problems might occur if one day it were necessary to go back into the design process and justify a particular dimension or strength parameter. There would be traceability problems and problems of reproducing a development process or adopting it to minor modifications of an existing design. Thus, a lack of formalized calculation procedures might lead to a deterioration of the technical mastery of designed products. 9.4 SUPPORT STRUCTURES IN OPERATIONAL DESIGN - IN SEARCH OF AN ORGANIZATIONAL MEMORY General Managers and Product Development Mangers in the case study companies liked to see more capitalization of earlier design efforts and less waste related to work on design solutions that either had been rejected earlier, but where this information was not available to the design technicians, or where competitors' patents hindered an exploitation of the design. The notion of wasted design effort also emerged at the operational level. Much of the discussion in this chapter has been centred around the simultaneous storage and availability of information concerning previous experiences, accumulated knowledge and news in terms of patents, and material, process, and product technology. However, from a different perspective, incomplete design efforts seen as a problem of waste could be transformed into useful experience if only there were some mechanisms or structures able to integrate these experiences. If the notion of memory is defined as "the mechanisms through which a certain acquisition is held available and can be recalled and reutilized" (Reuchlin, 1990, p. 173), it seems obvious that what both managers and operational people were looking for in the case study companies was an organizational memory making it possible to store, search for, and apply different kinds of information at any moment when this was suitable for the activity on hand. Different devices for capturing individual and collective experiences in product development existed in the studied companies. These presented two main problems, however: their lack of dissemination functions, and their fragmentation, i.e. blueprints, patent data base, launched 248 product data base, design study data base, etc. did not exist in an integrated form. Briefly, these different support structures acted merely as horizontally separated storage devices. This problem was shared among the Product Development Managers in the four supplier companies where follow-up interviews were conducted. Kim (1993) argues that support devices allowing for a learning process to take place should be transversal and active structures should be able to affect thinking processes and undertaken actions. As argued in chapter seven, an important part of the knowledge employed in product development work is tacit. Non formalized knowledge retained in the behaviour, minds, and memories of individuals and groups was often quoted as a central capability in the race for improvements and innovations in product technology. In order to share tacit knowledge with colleagues, both intra- and inter-functional, and facilitate its development with younger staff the necessity of introducing different means of formalization acting as memory was identified through the case studies. Two examples will be discussed here. The first concerns product data bases that were present in both of the case study companies. However, these contained only blueprints and specifications for design studies that have resulted in commercialized products. It would certainly be useful to expand this data base to all design studies that have resulted in a blueprint, i.e. also those that had not been prototyped, produced and commercialized. This would reduce the redundant work on design propositions that have been rejected earlier and considerably expand (at least double according to estimations of one design technician) the bank of data in which to search for ideas when faced with a new design demand. The second example concerns the visualisation of prototypes, mock-ups and defaulting products. According to the interviewed design technicians, the tangible object remains the best means for anticipating a function and a technical solution. The tangible object is a projection of ideas and in this role it contributed largely to the advancement of the ideas concerning a specific functional problem. In this way, different tangible design artefacts were identified as a means for developing product technology - they become a part of the organizational memory. Something that was quoted as very favourable for inter-personal and inter-project learning was to have as many mock-ups and prototypes as possible present in the design department. A design technician in COF explained that when he had a prototype or a mock-up exposed on his desk, everyone that walked by commented on these objects; the objects drew the attention to a specific problem, invited discussion -also inter-functional ones- that could be very useful for solving different kind of design problems related to the component in question. Moreover, the presence of such objects helped to make the related design project known within the design community and within other functions in the company. 249 The notion of the object is interesting to develop a little further. In Karlson's (1994) conceptualization of the product development process, the object plays a central role; "[The object] is the outcome and purpose of the process, it is the end. But it has a peculiar duality; it is also the means by which this end is reached. The object, or at least the image of it in the minds of the participants, is the medium by which the design process is carried through. It is the entity on which the work is performed, the reference and focal point, and the centre of all the design discourse. Individual activities become meaningful only with reference to the object. It becomes both end and means for the process" (Karlson, 1994, p. 133). Bucciarelli (1988) put an equal emphasis on the object-product qualifying it as an "icon within the culture of the firm" (p. 160). It is true that the object played a central role in the product development process in the case study companies, but the dominating discourse both at the management level and at the operational one was that of the function that the supplier company's products (objects) would satisfy. This was especially true at COF where top management's persistence on the role of the company as a provider of functional solutions to fastening problems, and not specific fastening components, was an example of a successfully deployed guiding vision -it was shared by all the interviewed actors who had a relation to product development. When relating Karlson's (1994) and Bucciarelli's (1988) perceptions of the object, the more central role played by the function in the case study companies, and the discussion of the resource-based theory and the development of core capabilities in chapter four, section 4.3.1, the following conclusion emerges. As the changes in the auto component sector are very turbulent at the moment, too much focus on objects could imply a risk of marketing myopia. The capability based competition, that showed clear tendencies of emerging in the studied sector, emphasizes the function that the expert supplier companies' products are supposed to satisfy more than the products themselves. The present research confirmed the object as a carrier of information, but Karlson's (1994) interpretation did not seem completely appropriate in the context of expert supply. The function replaced the object as the centre of design discourse. 9.5 DESIGN SUPPORT STRUCTURES - SUMMARY AND CONCLUSIONS Three different kinds of design support structures playing an important role as a context in which operational design was carried out were identified through the case studies. These were specifications, computer-aided systems for information transmission, and technology scanning. 250 The role of specifications in black box engineering was analysed in particular. Black box engineering necessarily redefines the role of the specification compared to the case of detailcontrolled parts. In black box engineering, specifying was seen as a process taking place at the systemic level of operational development work - it needed input from customers, from all the project participants inside the supplier firm, and from experts such as consulting engineers. Due to its process character, the specification was found to be a vehicle for design creativity through the exploration of several different ways of resolving a problem. This also meant that design technicians must take an active role in the process of specifying - something that could be difficult to assume rapidly. The present study revealed a general problem for design technicians to determine when a specific design project pertained to the category of black box parts. This indicated that suppliers should become more proactive in their relationships to customers and actively search for the information that they needed. Once this had been determined, design technicians did not have any particular problems with inventing new product solutions out of only partially framed specifications. The research data also showed that the more active a supplier was in looking for customer information, the greater the chance of profiling the company as an expert supplier. A result of specific interest emerged through the analysis of specifications and the process of specifying. It was found that several different parallel solutions were developed in each black box project and that these were assessed in common with the customer -final carmaker or system supplier- instead of, as traditionally, only inside the supplier firm. This line of action has been conceptualized in the literature as the set-based approach. In order to function properly, this approach requires a close alignment to customer needs and a capacity of anticipating the latter. In the case study companies this approach appeared only in situations were design technicians had direct communication links with their counterparts in customer firms. The set-based approach was also found to promote product innovation. Through the close interaction during the early stages of product development, customers became aware of a broader spectrum of the supplier's design capabilities, something that could result in the development of solutions beyond the performance of isolated components. The set-based approach also modifies the generally accepted idea that design changes must be minimized at any price (c.f. Bowen, et al, 1994; Clark & Fujimoto, 1991; Karlson, 1994). The case studies provided examples both of development projects where coordination and integration functioned very well -for example in triangulation where customers, system suppliers, and interfacing expert suppliers worked together on integrated problem solving- and in situations where coordination was poor. Late design changes intervened in both settings, however. In the case of triangulation this was due to the nature of the development work undertaken -black box engineering- that, as argued in this chapter, by definition implies a 251 series of modifications until different solutions are narrowed down to a final definition of different parameters. The set-based approach proposes extensive prototyping in order to make the necessary and normally numerous design changes as early as possible in the development process. However, once a parameter has been tested and confirmed as correct no later modifications are accepted (Ward et al, 1995). Working with several sets of possible solutions makes it possible to stick to this decision. If the value of a parameter is very uncertain, several sets representing different values will be kept a bit longer in the development process. Finally, it was argued that in order to balance the trade-off between continuous change of specifications during black box engineering and the demands for shorter development lead time, it could be useful to divide the design process into two parts; the functional concept and the dimensional definition. If suppliers apply the set-based principles and develop several parallel functional concepts and only in a second phase -after intensive joint problem solving with customers and interfacing suppliers resulting in a final solution- devote time to the dimensional definition, lead time and redundant work could be reduced. The ideas outlined in relation to specifications implied a high level of integration in the interfaces between an expert supplier, its customers, and its interfacing suppliers. This oriented the analysis towards information handling, and more precisely computer-aided systems for information transmission. In fact, the central problem in simultaneous engineering was to handle the related information flow. The case studies indicated that the existing inter-company CAD systems ought to be completed with more qualitative information and allow for interactive informal dialoguing between design engineers and technicians working on a system. This would typically facilitate the development of a set-based approach discussed above. However, before starting out on projects for developing these kinds of systems, the internal project structure and the configuration of the customer-supplier interfaces of a company need to be assessed and maybe modified. Moreover, an information system needs not only to be legitimized in an organization, but the latter should also be able to make use of the system. One of the central arguments in the organizational integration theories, namely that communication is relevant only if the interlocutors share a common cognitive ground, must also be taken into consideration when designing information systems. On the other hand, information systems could, for example, be explicitly designed for integrating dominant and non-dominant disciplines or different functions such as design and marketing in order to support the development of common cognitive grounds. In this case, however, the information system would start out more 252 as a vehicle for organizational evolution than one for improving efficiency in operational problem solving. A central advantage of these kinds of information systems is that they allow for the instantaneous creation of micro organizations that the natural working practices in product development was very often found to require. A problem was related to the fact that the kind of information systems discussed here will unmask problems inside different participating firms, something that can be both strategically risky and politically embarrassing especially in a situation where competition between interfacing suppliers in terms of tier position is present. The third identified design support structure, technology scanning, was particularly important in sectors where the product performance was directly dependent on advances in material and/or process technology, such as in technical plastic parts. The results indicated that technology scanning must be extended to benchmarking with users of new process technology both within and outside the car industry, that particular attention must be paid to the dissemination of obtained information, and that the productivity of scanning activities increased if those responsible for scanning also made an effort to bring operational R&D staff closer to the source of the technical information. The latter could facilitate analysis of scanned information, and speed up the following decision-making process of how to approach the technology in question. Concerning the dissemination of obtained information, this could be facilitated and made more reliable through the development of an internal interactive information system allowing for a real-time reaction to information and making it possible to store information, related analyses, and decisions in a user-friendly format. Patent scanning was a particular part of technology scanning. Due to the fact that existing competitor patents could be easily infringed by new product solutions developed by design engineers and technicians, it was argued that patent scanning and management activities should be seen as an integral part of operational design. Moreover, a supplier's patent assets were found to play an important role in centralizing past and present product development capabilities. It was finally concluded that a common characteristic of the identified design support structures was that of an organizational memory. Holding acquisitions in terms of design capabilities available to all potential users, and making it possible to recall and reutilize them were central objectives in specifications, computer aided information transmission systems, and technology scanning. The main problem with existing support structures was that they lacked appropriate dissemination functions thus hindering the active participation of all the concerned actors, for 253 example relative to a specific product design problem in a particular customer relationship, and that they were too fragmented. True support structures for learning should be transversal and active. 10 LEARNING IN PRODUCT DEVELOPMENT 254 Through the development of the core category operational design in chapter 7, learning emerged as a transversal issue in the entire development process, i.e. in individual, group, project, and systemic work. Learning as a transversal phenomenon has also been recognized by recent scholars, e.g. Hughes & Chafin (1996), who argue that "continuous interaction around discovery learning, certainty in knowledge, consensus, and shared responsibilities improves the new product development process as we move into an era of continuous innovation" (p. 103). Also Hawkins (1994) stresses the integrative nature of learning in everyday work. According to him, learning is not knowledge banking of factual information and something that is done separate from, and normally prior to, doing. It is sharing of skills and insights integrated in operative work. In individual design work, individual learning through conceptualization of an operative trialand-error line of action was identified. In group work, an intra-company intra-functional learning situation was identified. This learning occurred between different design technicians, between senior and junior design technicians, and between project managers and design technicians. The objective and outcome of this kind of learning was the dissemination of best practice in operative design. In project work, the situation changes as the professional languages of different participants are no longer the same. Here, the identified learning situation was one of intra-company interfunctional learning between staff from product engineering, process engineering, manufacturing, marketing and sales (including sales representatives), purchasing, and quality management. The literature review on communication and coordination undertaken in chapter four concluded that only communication that fits into a common cognitive ground will be interpreted in a relevant way. Moenaert & Caeldries (1996) illustrate this by arguing that communication should be aimed at developing convergence between the information source and the information receiver. In chapter four, the common cognitive ground was identified as an intangible space of shared knowledge and understanding. In order to enter into this space there is a need for inter-functional learning. The particular learning need that occurs for anyone wanting to enter into a common cognitive ground was exemplified in the case studies by the integration problems in project work. Finally, in systemic work there was a problem of integrating both other internal functions in the supplier company and product development personnel in other supplier firms and in customer firms, i.e. design personnel responsible for neighbour components designed elsewhere. However, the distinctive learning situation in systemic work emerged as the one taking place between design staff in the studied supplier firm and their counterparts in interfacing supplier and customer firms. In other words an inter-company intra-functional learning situation was identified. 255 At this point it must be underlined that the ambition here is not an exhaustive analysis of the learning problem. The analysis will build on the learning situations identified in the case studies. The inter-company inter-functional mode is deliberately left aside as it was not a significant input in the product development process in the studied firms. Figure 22 illustrates the different functional actors that intervene in the product development process and the identified learning situations. Intra-company, inter-functional learning Process Engineering Purchasing Sales Intra-company, intra-functional learning Design function in customer firms Design function in interfacing supplier firms Inter-company, intra-functional learning The design function in a supplier firm Figure 22. Actors and learning situations in the product development process. This model of different learning situations, developed directly from the case studies show similarities to the levels of knowledge creating entities -individual, group, organizational, and inter-organizational- in Nonaka & Takeuchi's (1995) theory of knowledge creation. In their model, knowledge created by individuals is organizationally amplified -through an organizational support context- within an expanding community of interaction which crosses intra- and inter-organizational levels. Throughout this process, knowledge is created through a continuous interaction of tacit and explicit knowledge. In the present research, this interaction was established as a characteristic feature of problem solving in product development (c.f. section 7.2). The parallels to learning models that were drawn in the analysis of operational design work means that the perspective in this study moves away from a static input-process256 output one and approaches the paradigm of dynamic organizational knowledge creation (Nonaka, 1994; Nonaka & Takeuchi, 1995). Through its positioning in a wider business process context -that of product development- the analysis of learning situations will focus on the development related mechanisms that underpin this organizational knowledge creation or organizational learning process and on managerial action that can facilitate the process. The approach to analyzing learning in the product development process will here be the following. First a discussion of existing approaches and definitions of learning will be undertaken. The focus will be on identifying the differences and relationships between individual and organizational learning. From the case studies leading to the four-step model of design work presented in chapter seven, it became clear that these two processes are intimately interrelated and that collective or organizational learning is a natural outcome of the group, project, and systemic work that product development represents. In the literature review undertaken in section 4.4 on operational coordination in the product development process, learning emerged as a possible solution or means for resolving the core problem identified in product development - that of integrated problem solving. However, the learning process was never explicitly addressed. The central role of learning in the product development process emerged through the case studies and not through the analysis of literature dealing with design collaboration between suppliers and customers. A debriefing of the field of learning will therefore be necessary in order to provide a framework of analysis and reflection to apply to the case observations. Generally speaking, the literature dealing with learning is mostly conceptual and only rarely is the learning phenomenon related to a realworld managerial problem4 . Second, the observations of product development work will be analysed from the angle of individual and organizational learning within the framework. It will be applied to a running description of three different learning situations identified in the case studies. The objective of this analysis is that of a gradual integration of different parameters in order to build a description that corresponds as closely as possible to the transversal and continuous realworld learning process. 4 One exception is Charue & Midler (1994) who analyse organizational learning in automobile stamping plants. Conversely, the work of Wheelwright & Clark (1992) on learning in product development is only briefly anchored in learning theories. 257 10.1 INDIVIDUAL AND ORGANIZATIONAL LEARNING: TWO CONCEPTS As the research deals with the tangible issue of developing new products, learning is defined as taking place only when new knowledge (know-how and/or know-why) is translated into different behaviour that is, at least theoretically, reproducible (Argyris & Schön, 1978). It is a question both of operational learning -what people learn- and conceptual learning -how they understand and apply this learning (Kim, 1993). The knowledge that people acquire includes facts, principles, experience-based insights, working procedures, research findings, and ideas (Glaser et al, 1983, quoted in Moenaert & Caeldries, 1996). It seems clear both from the case studies and from the product development literature (c.f. Karlson, 1994; Moisdon & Weil, 1992; Wheelwright & Clark, 1992) that an enlargement of individual competencies is important but not enough for continuously improving product development performance. Wheelwright & Clark (1992) argue that the learning that is of specific interest in product development is a learning that integrates tasks and capabilities that cut across functional boundaries, and critical linkages at the working level between different disciplines, functions and departments. Organizational learning (OL) was also an explicit goal set up besides individual learning by several General Mangers and Product Development Managers that were interviewed in the present study. OL is also a subject that has been frequently treated in recent managerial literature (e.g. Argyris, 1994; Garvin, 1993; Hawkins, 1994; Hughes & Chafin, 1996; Kim, 1993; Nevis et al, 1995; Nonaka & Takeuchi, 1995; Schein, 1993; Senge, 1990) but where the origins can be found much earlier (e.g. Argyris & Schön, 1978; Cyert & March, 1963; Simon, 1969). Important efforts in clarifying these origins and setting new research agendas in the field have been conducted by, for example, Fiol & Lyles (1985), Hedberg (1981), and Shirvastava (1983). A general problem in defining organizational learning is to distinguish it from individual learning (Hedberg, 1981; Kim, 1993). Hedberg argues (p. 6): "Although organizational learning occurs through individuals, it would be a mistake to conclude that organizational learning is nothing but the cumulative result of their members' learning. Organizations do not have brains, but they have cognitive systems and memories. As individuals develop their personalities, personal habits and beliefs over time, organizations develop world views and ideologies. Members come and go, and leadership changes, but organizations' memories preserve certain types of behaviour, mental maps, norms, and values over time." Through the research, the importance of such types of behaviour, norms and values for an organization's daily life were clearly confirmed. Several features were identified: 258 • Means of guidance, for example strategic visions such as being a partner in product development, and operational performance measurements such as the lead-time between design start-up and prototype delivery; • Instructions and procedures, for example how to interact with the customer during the development of a new product; • Professional habits and concerns, for example a passion for technically optimized products on the part of design technicians, or a strong focus on customer satisfaction on the part of project managers, consulting engineers and sales engineers; • Support structures, for example design data bases and physical cohabitation of different functional staff. The fact that all these different features are connected -through an organization that defines specific links between its individual participants5 - creates reciprocal influence between them. Thus, when an individual that is a part of an organization learns a new or different type of behaviour, this will be registered and spread in the organization through existing procedures and support structures. Two outcomes are then possible (Argyris & Schön, 1978; Fiol & Lyles, 1985): • The learning restricts itself to detecting and correcting errors within a given system of rules, this 'single loop' learning occurs through application and repetition, and maintains the central features of an organization's 'theories-in-use'; • The learning aims at adjusting overall rules and norms; this 'double loop learning' occurs through the use of heuristics, skill development and insights, and develops differentiated 'theories-in use'. Both processes can be qualified as organizational learning; the first adjusts in part what the organization does, the second influences globally how the organization thinks. What qualifies learning as organizational is then if the outcome of individual learning provokes reaction and action in the entire organization through a modification of collective behaviour and a modification of norms on the one hand, and mental maps and values on the other. Often, some type of crisis is necessary for provoking double-loop learning: a new strategy, a new leader, or altered markets (Fiol & Lyles, 1985). However, there is not only the individual learning that influences organizational parameters in a one-way direction. The original learning at the individual level is influenced both by current episodes and retained sediments of past learning in the organization (Danielsson, 1983; Hedberg, 1981). Thus, individual and organizational learning are interconnected. 5 Schermerhorn et al (1991) define an organization as "a collection of people working together, in a division of labour, to achieve a common purpose", p. 593. 259 Based on Senge (1990), Kim (1993) emphasizes the notion of mental models for understanding individual and organizational learning and their relationships. These are defined as "deeply held internal images of how the world works, which have a powerful influence on what we do because they also effect what we see" (p. 39 quoting Senge, 1990 and Forrester, 1971). Kim (1993) pursues: "Mental models represent a person's view of the world, including explicit and implicit understandings. Mental models provide the context in which to view and interpret new material, and they determine how stored information is relevant to a given situation. They represent more than a collection of ideas, memories, and experiences...[they are]...the manager and arbiter of acquiring, retaining, using and deleting new information [...] Mental models not only help us make sense of the world we see, they can also restrict our understanding to that which makes sense within the mental model" (Kim, 1993, p. 39). In his model, individual mental models consist of two parts; frameworks related to conceptual learning, and routines related to operational learning, c.f. figure 21. Individual double loop learning occurs when an individual modifies his or her mental models as a function of learning and when this, in turn, affects future learning. Thus, the mental models seem to be much the same as the theories-in-use mentioned before. It is through the concept of mental models that Kim (1993) relates individual and organizational learning: "the cycles of individual learning affect learning at the organizational level through their influence on the organization's shared mental models" (p. 43) and further "Organizational learning is dependent on individuals improving their mental models; making those mental models explicit is crucial to developing new shared mental models" (p. 44). As stated above, individual mental models consist of frameworks and routines. In Kim's (1993) model their equivalent at the organizational level are world views and standard operating procedures. The relations between the two levels are the following: the organization's world view slowly evolves to encompass the current thinking -i.e. the current frameworks- of the individuals within the organization. Individual routines that are proved to be sound over time become standard operating procedures. 260 10.1.1 Individual and Organizational Learning - Conclusions to the Literature Analysis Individuals learn facts, principles, and working procedures through operational implementation and observation of different ways of work. These factors become routines that represent learning at the procedural level. Parallel to this, individuals develop experience-based insights and ideas through a conceptual assessment of different ways of working and an elaboration of alternative ways. This results in the differentiation or development of frameworks that represent learning at the conceptual level. Together, these routines and frameworks constitute individual mental models or theories-in-use. Following the new insight, either purely operational, or also conceptual, the individual will act and the environment respond. This environmental reaction will influence future learning. Now, if the individual is a member of an organization, where his or her activities both depend on those of others and influence those of others through the links between participants that the organization defines in order to fulfil the purpose for which it exists, individual learning will both be a function of the organization's behaviour, mental maps, norms, and values, and at the same time influence these parameters. As the central question here is how organizations learn, the focus will be on the last process, but without forgetting that existing organizational routines and values affect what each individual is learning. The transfer from the individual to the organizational level takes place through existing procedures -physical support structures- on the one hand, and interpersonal communication -intangible information exchange- on the other. Thus, the impact of individual learning on organizational learning is of two kinds: new formal ways of working where individual routines influence organizational ones; and new cognitive insights where individual frameworks influence organizations' world views. Together, organizational routines and world views constitute shared mental models or organizational theories-in-use. As in the individual case, the organization will act differently following its modified world views and routines. For someone external to the organization, altered organizational action will be perceived as a systemic change in the global relationship to the organization. Following the environmental reaction to this change, individuals within the organization will alter their operational routines and/or conceptual frameworks which will ultimately influence the organization's shared theories-in-use and organizational learning will proceed. Single-loop learning occurs if actions and their consequences do not influence the existing theories-in-use. Here, operational execution will change but frameworks and world views remain. Double-loop learning occurs when actions and their consequences are accompanied by a conceptual assessment and where the latter redefines existing theories-in-use. 261 10.1.2 Materializing Learning: Towards a Framework for Analysis How, then, is organizational learning materialized in companies? Based on Huber (1991), Nevis et al (1995) identify three main stages in an organizational learning system; (1) the acquisition of knowledge, (2) the sharing of knowledge, and (3) the utilization of knowledge. Acquisition is about the means for appropriation of knowledge. This can happen through the translation of physically encoded information, for example, in the case of design engineers, through the analysis of a competitor's product, or through interpersonal communications (Allen, 1977). Sharing takes place through communication, either verbal or through the dissemination of written text, images or other artefacts (including mock-ups). The aspect of sharing indicates that organizational learning comprises an important dimension of social interaction, and that this, in turn, comprises interaction between tacit and explicit knowledge (Nonaka & Takeuchi, 1995). Utilization concerns the objective of the learning process. Hawkins (1994) emphasizes that learning is a means to an end "its value is dependent on where your learning is taking you" (p. 72). Steensma (1996) argues in the same sense; evidence of learning would be present if an organization can alter its system in response to changing environmental stimuli and needs. The utilization of learning is often summarized in outcomes that get encoded in routines leading to transformational change (Hawkins, 1994; Kim, 1993). The three components are interlocked; organizational learning is a cyclical process based on experimental trial-and-error, development of insights, and structural and other outcomes. Communication is central both for acquisition and sharing. If learning is utilized for deploying change, the means for acquisition and sharing will be permanently altered, and so will the central element of learning; the factual knowledge that is originally acquired. As stated before, knowledge concerns both know-how and know-why and includes facts, principles, experience-based insights (essentially tacit), working procedures, research findings, and ideas. Concerning the utilization of learning, some authors focus on OL as a process of selecting and retaining knowledge based on environmental demands (Anderson & Tushman, 1990, quoted in Goodman & Darr, 1996). This perspective is based on the idea of the alignment of an organization to its environment to remain competitive and innovative (Fiol & Lyles, 1985). Fiol & Lyles (1985) argue that alignment implies that firms must have the potential not only to learn, but also to unlearn or relearn based on their past behaviour. It is important to emphasize that the idea of alignment is more than passively following customers' needs and competitors' manoeuvres. Learning is both adaptation -defensive adjustment to reality-, and differentiation offensive use of knowledge to improve the appropriacy to the environment- (Hedberg, 1981). In the case studies, the last point dominated the learning problem. Generally speaking, the 262 companies had gone through an adaptation phase, and learning was beginning to be seen strategically as a way of creating competitive advantage. Nevis et al (1995) propose a model of organizational learning composed of seven learning orientations and ten facilitating factors. Learning orientations are defined as the values and practices that reflect where learning takes place and the nature of what is learned. These are dimensionalized on a binary scale where a company will be positioned in one of two opposed cases. The facilitating factors are the structures and processes that affect how easy or hard it is for learning to occur. The learning orientations and facilitating factors are presented in table 15. The model developed by Nevis et al (1995) is more concrete than most of the previously analysed models. It contains tangible check-points and is developed in a real-world managerial language. It therefore seems to be an appropriate tool for benchmarking the research findings after the analysis that will follow below. Learning Orientations Facilitating Factors 1. Knowledge source: Internal - External 1. Scanning Imperative 2. Product-Process Focus: What? - How? 2. Performance Gap 3. Documentation mode: Personal - Public 3. Concern for Measurement 4. Dissemination Mode: Formal - Informal 4. Experimental Mind-set 5. Learning Focus: Incremental-Transformative 5. Climate of Openness 6. Value chain Focus: Design - Deliver 6. Continuous Education 7. Skill Development Focus: Individual - Group 7. Operating Variety 8. Multiple Advocates 9. Involved Leadership 10 Systems Perspective Table 15. Seven learning orientations and ten facilitating factors (After Nevis et al, 1995, p. 77). Another interesting contribution which directly concerns learning in development projects is that of Wheelwright & Clark (1992). They emphasize that in order to learn from development projects one needs to understand the sequence of activities and critical events that the product development process comprises. Thus, they argue for a research perspective close to that of the present research project - an understanding of the behaviour of the development system. From their analysis, Wheelwright & Clark (1992) identify five critical events for potential learning from development projects. What makes an event critical is, they argue, its connection to an aspect of the development process that drives performance. Their five critical events are (p. 288): • Recurring problems linked to critical performance dimensions - does the organization retain solutions and make them permanent? 263 • Crucial individual activities/tasks and associated capabilities - does one measure the right information about tasks? Does one have the skills needed? • Working level linkages (e.g. engineering - manufacturing) - are there processes, skills, attitudes and values that drive integration? • Design-build-test cycles - are the right people, tools, and supporting resources available? • Process for making decisions and allocating resources - are the right people involved at the right time with the right information; are projects correctly planned? These items and the associated questions provide an additional focus for the analysis of learning situations and individual and organizational learning in the studied expert supplier companies. The findings from the analysis undertaken above can now be summarized in the form of a framework of questions that will be useful to examine in the light of the research data in order to achieve a better understanding of learning processes: • What are the transfer mechanisms between individual and organizational learning? • What in practice makes the difference between whether single or double loop learning occurs? • What blocks learning in different phases and stages in the learning process? 10.2 THREE DIFFERENT LEARNING SITUATIONS - AN ANALYSIS The three different learning situations identified in the beginning of the chapter provide the basis for this analysis. Intra-company intra-functional learning inherent in individual and group work, intra-company inter-functional learning inherent in project work, and inter-company intrafunctional learning inherent in systemic work will be examined in detail. As a point of departure for this analysis, however, a general example of the learning cycle in design work will be presented. It represents the typical process of developing and transmitting product design solutions observed in the case studies. The learning cycle in design work - a general example based on interviews and observations of metal project managers and their attached design technicians in COF The learning process starts at the individual level with a design technician developing a new solution, or resolving a particular type of functional problem, 264 through a trial-and-error learning process that typically mixes individual activities and direct inputs from colleagues involved in the design process. As long as the work concerns a functional problem these inputs will come from other design technicians and design engineers - for example consulting engineers previously identified in the present research. When the design technician has reached a satisfactory solution according to his or her own judgement, the proposed solution will be presented to the project manager who then will submit it to the customer in accordance with the existing organizational routines or standard operating procedures (Kim, 1993). However, before presenting the solution to the customer, the project manager will evaluate it in his or her perspective, a perspective that normally is more focused on customer satisfaction and manufacturability concerns than that of the design technician. While the first learning situation, at the individual and group level, concerned both the how, i.e. the process of arriving at a solution, and the what, i.e. the solution itself, the evaluation of the project manager will be concentrated on the latter - the what. If he or she finds problems of manufacturability or high material or manufacturing costs and communicates this to the design technician, the latter will look for ways of improving his or her understanding of the problems for example in purchasing, process engineering, and manufacturing. He or she will enter into inter-functional learning processes with staff from these areas. The reason for which inter-functional learning was often a strategic objective was that a direct integration of cost and manufacturability aspects with the functional ones had been identified as a way of reducing design rework and modifications and thus of speeding up development lead-time. If a solution passes the examination of the project manager, it will be presented to a customer -always in accordance with the standard operating proceduresand this will add an environmental reaction to the proposed solution, and, hence, influence the original learning.. When the customer receives the blueprint and related documentation specifying the component he will, as did the project manager, focus the analysis on the what i.e. function, fit into a larger system, and price of the component. After analysis, customer feed-back will concern (1) the functional performance where the design technician must intervene, (2) the system fit, i.e. the ease of assembly into the system where the component will function, where the design technician, process engineering and manufacturing must intervene, and (3) the price - where all concerned actors in the supplier's product development process must intervene under the coordination of the project manager. This emphasizes that the how perspective inside the supplier company, i.e. how the entire company went about elaborating the complete offer 265 around the new solution, is very important for learning as a function of customer feed-back in order to constantly improve the offer of the company. The reaction of the customer will not only concern the design technician, but also the project manager, the process engineers, the design manager, and the colleagues of the design technician - all those who were formally or informally aware of the new solution. In this example it is the organization's routines or standard operating procedures that transfer the individual insights to an organizational level and back again. 10.2.1 Intra-company Intra-functional Learning The analysis begins with the initial event, when the design technician experiments with different solutions in order to find a specific technical solution to a functional problem in a component design, and analyse it in the light of the framework developed by Kim, (1993) (c.f. figure 21). At the most basic level, the development of a new solution is a process of individual operational learning where the technician implements an idea and observes its outcome. The art of design work is, however, not only operational. In order to progress, the observation must be reflected upon in terms of an assessment of what has been implemented. In this step, the process moves into the conceptual side characterized by why questions and an 'if-what' reflection. First, the technician will question the outcome of the implementation and try to find causes and sub-causes that lead to specific problems and advantages in the obtained results. Here a first remark can be addressed based on observations in the two case study companies of how design technicians worked in operative problem solving: if the outcome of a specific problem solving is an immediate success, at least at the individual level, the learning cycle will stop at the assessment level -which in fact can be very short and almost automatic. This means that it can hardly be characterized as a conceptualization of the operational phase. In such a case there is an important risk that the process leading to the new solution will be forgotten. The solution will be transmitted to the project manager and the customer, but there is a strong risk that the traceability of how it was developed will be difficult to restore. If the first solution is not satisfactory, i.e. the result of the assessment is a rethinking or a design of a new solution, the conceptual phase of the learning cycle will be completed. When the elaboration of a new solution has been completed according to the technician's own appreciation, a new implementation will be tried and the process moves back to the operational side. In this case the problem will occupy a larger place in the mind of the design technician - he or she has reflected upon the first effort and developed a second solution. The chances that the way of reaching a solution will be remembered increases compared to the first 266 case. Also Wheelwright & Clark (1992) argue that episodes where things go wrong are the main raw material for learning. This example shows that the first problem blocking individual learning is that the quicker a solution is found, the greater the risk of bypassing learning. This risk increased with the heterogeneity of work tasks; the risk of forgetting seemed to increase if the subsequent tasks were very different from the first. A design technician who has not reflected upon a newly found solution will not draw parallels to other work. It is here important to notice that even a quickly found solution often was observed to be the result of a complex web of experiencebased insights and randomly intervening ideas or inputs from colleagues or customers. This problem was also reinforced by the fact that, as discussed in the example, feed-back from project managers and customers was found to be focused on the solution itself and not on the process of how the solution was obtained. However, quick satisfactory solutions are one of the central objectives in product design. One way of managing this trade-off, proposed by the interviewed design technicians themselves, was to keep a trace of all design activities, through recording sketches and notes developed during design work. It could also suffice to explain the traceability problem formally, for example in a weekly meeting, and instruct design technicians that every time a solution works according to individual judgement, special attention should be paid to a brief documentation of the line of action employed. The impact on individual learning from the environment is also an important parameter to consider. As mentioned in the example, this impact reaches design technicians in day-to-day design activities through colleagues, project managers, and customers. If the feed-back from the environment is positive one will try to know how the solution was elaborated in order to repeat it and make it a new best practice. If the answer is negative one will also try to know how the solution was elaborated in order to avoid repeating the same mistake and ultimately unlearn an undesired line of action. As customer satisfaction was a dominating guiding vision in all of the studied companies, a positive reaction from customers to a new product, for example, would immediately modify mental models in favour of the process leading to the satisfying result. Besides the problem of neglecting the assessment side of problem solving, two additional blockages to learning were identified in the case study companies. The first concerned the fact that once a solution had been presented to the customer there was often a long lead-time before the individual received feedback that could confirm or lead to a modification of the related proposition. The second concerned the risk for distortion of messages sent out by design technicians when several interfaces had to be passed between the one in the supplier company proposing the new solution and the one in the customer company whose work would be affected by it. Both these points re-emphasize the importance of relevant support 267 structures. Design notes ought to be stored at least until the environmental reaction is fed back. Moreover, the use of such notes would make it possible to correct distorted information. When going back to the very problem solving in the example above, one can ask the question of how the design technician went about finding the product solution, i.e. how did he or she work in implementation, observation, assessment, and planning? What is interesting is to see whether he or she applied recognized and well known routines for his or her job, or if he or she applied alternative methods outside the institutionalized routines of the organization. If the latter was true and on condition that the internal and environmental feed-back was positive, this implied that a change in the individual mental models (frameworks and routines) would take place. If the design technician then continued to work in this different way and the outcome continued to be positive, it would be spread within the design community and ultimately to the entire organization whose routines would also change. A double-loop learning would occur. If the design technician applied existing routines for resolving the problem a single-loop learning would occur as there would be no questioning or modification of existing theories-in-use. The case studies indicated that the change of theories-in-use often originated at the individual level, for example through testing of alternative ways of realizing a product design. The fact that he or she did so was essentially due to three different factors expressed by interviewed design technicians: 1. Personal individual initiatives, 2. Means of guidance encouraging initiatives in a specific area, 3. Direct input from customers. The first factor was, in fact, identified as a main source of several important changes that had recently occurred in the case study companies. The most striking example was that of the development of direct communication channels between design technicians in supplier and customer firms. This phenomenon was mentioned in section 7.5 and will be further analysed below in the section dealing with inter-company intra-functional learning. Concerning means of guidance, these played an important role for challenging the status quo. Management's attitudes to errors were observed to be particularly important for promoting double-loop learning. In both of the case study companies employees felt that they had the right to commit errors on condition that these were not hidden away and their causes analysed. For this reason, employees were not afraid to challenge existing practices. Moreover, in COF, 268 for example, formal project reviews were limited to new and/or more complex6 design studies in order to minimize formalities and bureaucratic procedures, and improve creativity. Finally, input from customers was found to determine what kind of learning took place in different situations. This will be further discussed in the following sections. 10.2.2 Intra-company Inter-functional Learning This learning situation was identified in project work where there was an important need for coordinating people belonging to different functions, having different professional experience and languages. The objective was to develop shared knowledge and understanding through inter-functional learning. In this case one of the main obstacles blocking organizational learning, i.e. functional isolation, was eliminated from the beginning. From the analysis of communication theories in chapter four it became clear that when different professions are concerned, simply communicating will not improve coordination. A common cognitive ground must be created. The specific learning situation that will be examined here is that concerning a design technician in one of the case study companies located in the process engineering office (c.f. section 7.4). This co-location meant a confrontation between different individual problem-solving cycles and, for the design technician, a participation by the design technician in collective process engineering work. For the process engineering technicians it meant an opportunity to exchange information with a representative from a profession that to a large extent determined their work. This kind of confrontation between two different professional cultures makes it possible for the participants to become acquainted with each other's priorities, routines, problems, and constraints and to integrate this into individual mental models and transfer the latter to other colleagues in the original function. What is particularly interesting in this example is that the transfer mechanism for organizational learning between product engineering and process engineering is the individuals from the respective functions working together in joint problemsolving. This transfer took place in the following way. Through joint problem-solving with process engineering technicians, the design technician developed a different way of thinking about design problems. Using the learning vocabulary, his or her cognitive frameworks for thinking about design changed and the process became a case of double-loop learning. Consequently, the newly acquired way of working would remain after the reintegration of the design technician in the design office. There, he or she would enter into the process of intra-functional learning between design technicians discussed previously, and the new experience of direct 'in-design' of manufacturability would be transferred to design colleagues. Through this new experience the design technician discovered that he or she 6 The threshhold of complexity was identified by the product development manager through an integration of product and process technology aspects, patentability opportunities, and commercial risk. 269 could easily integrate manufacturing considerations in the design, and in consequence his or her way of executing operational design work changed. In relation to Kim's (1993) model, the design technician had developed a new mental model guiding his or her work. Concerning what makes the difference between double-loop and single-loop learning, nothing guarantees that actors confronted in joint problem-solving will change their routines or frameworks. The learning can be limited to one of understanding the profession of a colleague without changing one's own way of working. Data from the last series of interviews suggested that two factors could block the double-loop learning in this context. The first concerned the personality of the learner in question. Not all individuals are open to changing their mental models in spite of the motivating factor of receiving better attention from colleagues7 . The second factor concerned the means of guidance in place. If no explicit managerial actions were taken in order to bring different professions together, no more than the necessary coordination for reassuring work tasks in the established manner (existing organizational routines) would take place. This confirmed the idea that specific means of guidance might make the difference between single and double-loop learning. Globally, however, the two case studies indicate that inter-functional learning in product design has a good chance of succeeding because it conforms to the natural working practice in product development. As the example above illustrated, this natural learning was unblocked through a managerial action of temporarily delocating different project participants. 10.2.3 Inter-company Intra-functional Learning This learning situation emerged from the observations of how design engineers and technicians in the case study companies sometimes worked with their counterparts in customer firms. The same way of working could theoretically also take place with interfacing suppliers, but this was developed to a much lesser extent in the studied companies. Based on very open and frequent information exchange, new knowledge and insights were developed through inter-company learning between actors having the same profession. This learning situation was not present in all customer-supplier relationships. If the way of communicating with customers at the design level is looked at on a continuum, it ranged in the studied companies from no customer interaction at all (in direct arm's length relationships), via early involvement and sporadic participation, to a broad-band and continuous information exchange. The label permanently open communication channels was chosen to name this last communication mode. It was observed occasionally in the case study firms, more precisely it was observed in indirect and direct expert supply, both in the relationships to carmakers and system suppliers. In fact, 7 The psychological factors behind such behaviour are certainly numerous and of critical importance but they are situated beyond the scope of the present research. 270 permanently open communication channels were identified as a condition for inter-company intra-functional learning. Permanently open communication channels means that design engineers in supplier firms are in continuous contact with their counterparts in the carmaker firm at any stage in the development process. The observed functioning was the following. Typically, a design engineer concentrating on a product solution called a colleague at the other side of the interface to discuss an operational problem in real-time. This could happen in all phases of the development process, also in the pre-planning stage. The communication tools were simple telephone conversations, visits, and in some cases also video conferences and electronic interchange of blueprints and specifications. Personal relationships between the actors were of crucial importance. At the engineering level, the carmaker's organisation was in fact quite easy to survey for the supplier thanks to a driven project organisation. This allowed for such relations to emerge. Moreover, there were few problems of understanding, as there can be in larger project groups, because the professional languages were similar - the interlocutors were all design engineers and technicians. COR, for example, was connected to the engine design department in the carmaker companies, and more particularly to the department for cooling systems and thermostats where COR had its specific interlocutors. COR technicians could easily discuss with team members in other projects related to their functional group. As such interconnections tend to become more and more important, the communication mode can be assumed to evolve towards a generalization of the permanently open communication channel mode. The properties of permanently open communication channels can be summarized as: • High frequency of information exchange; • Non project-tied information exchange; • Non sequence-tied information exchange; • Openness and frankness; problems are discussed in order to find solutions together and there is a margin of negotiation within the triangle of quality, cost, and lead time; After this identification and definition of the permanently open communication channel concept, the related learning situation will be focused upon. The development of open communication channels was in itself a process of double-loop learning. At the individual level, design technicians reacted to the gap between received answers and requests or questions concerning operational design problems (c.f. section 7.5). This gap made their work difficult and fragmented, caused delays, and meant it was not always optimized in terms of functionality. At the same time, customers demanded better solutions 271 more rapidly, and at a lower final cost. Mostly due to time pressure, design technicians began to short-cut the normal routine and contacted customer design technicians directly beginning with people they already knew from occasional joint work on project platforms. This was a typical example of individual initiatives leading to a double loop learning and, consequently, to a modification of existing theories-in-use. The direct contacts eliminated several of the previous problems. More complete information could be obtained, an answer to a request could arrive almost in real-time, and background and side information could be more easily obtained. The result was shorter development lead-time and more innovative solutions, or solutions with better functional performance and/or system fit satisfying customers to a larger extent. This resulted in a positive customer input which then speeded up the process of changing the existing mental model (i.e. that mainly project managers should have direct customer contacts). When project managers and engineering managers discovered an improvement concerning the parameters mentioned above, their analysis of how development work was conducted transferred the new practice to the managerial level. As a function of its positive impact, this way of functioning was then planned to be gradually implemented as a new routine and the organizational theory-in-use changed. As discussed in section 8.2, people need successful examples in order to change their mental models. Both field observations and recent literature (Ellison, et al 1995) indicate that companies are paying more and more attention to the management of a project portfolio with transfer of experience between projects. Because technologies and designs developed in one project are reutilized or transferred to other projects, each new product development project will have technological and organizational linkages or interdependencies with other past and ongoing projects (Nobeoka & Cusumano, 1994). This was exactly what was observed when the interaction with customers in the framework of systemic product development work was studied in the present research. Moreover, thanks to an accumulation of experience, it was observed that these linkages could be extended to forthcoming projects. Permanently open communication channels facilitated this transfer of experience -this inter-company learning- and materialized the links between projects. Nobeoka & Cusumano (1994) make a distinction between two types of linkages between multiple projects: inter-product line linkages between different product lines, and evolutional linkages between past and present projects. The linkage to future projects through the means of learning identified in the present study could be included in an extension of the second type. The situations observed in the case studies integrated these two types of transfer due to the fact that the design studies were much more simple (in the majority of the cases they concerned the development of one discreet component) than those described by Nobeoka & Cusumano (1994) who studied large manufacturers. In the cases, the design technicians all had 272 several design studies for products in different product lines going on simultaneously (normally around five projects). When reflecting on a specific problem in one of these projects, the design technician would typically work as described in the discussion of individual design work in section 7.2, i.e. through an integrated doing and reflecting cycle. He or she would use accumulated knowledge and know-how (acquired through past learning), consult colleagues both inside and outside the design function, and integrate with customers, i.e. practise a systemic way of working. If permanently open communication channels were established to a customer, functional problems that were discussed in relation to one component in a particular project could have immediate impacts on the development of a similar component in another project at another phase. Thus, the transfer of knowledge concerned both different product lines and different evolutional phases of projects. Such rapid design transfer between parallel projects is not without interest. Nobeoka & Cusumano (1994) found that it contributed to a reduction (minus 33%) of the number of engineering hours required for developing a product compared to a sequential experience transfer. Where, then, does learning recede? The answer is in the transfer of the commonly developed insights taking form during integrated problem solving discussions between design technicians in supplier and customer firms. When these insights are transferred between more and less advanced projects and new projects can benefit from them, learning has taken place. At the basic level there is nothing that distinguishes this learning from the one analysed before; there is a collective implementation, observation, assessment and planning cycle realized between design technicians in two different companies. This learning becomes organizational when the insights are spread to other design staff in both the supplier and customer firm. As these will have reciprocal contacts with each other, their interactive design work will be affected by this transfer. Concerning internal transfer of acquired experience to other projects, a part of it was ensured through the fact that each design technician was responsible for several parallel projects. Here, the multi-task individual was the means for transfer. However, the transfer would be incomplete if experience was not shared between design colleagues. In the case observations this happened in two ways. Either the transfer relied upon internal communication and information transmission through informal and formal communication channels, organizational routines (such as procedures for product development work), and support structures in terms of design data bases and mock-ups -in the same way as in intra-firm learning-, or it took place through the intermediary of customers. The latter could happen in the following way. A design solution elaborated in common between a design technician in a studied supplier company and his or her colleague in a customer firm was sometimes more quickly transferred in the customer's organization than in the supplier's. Another customer design technician, informed from within his or her organization of the new solution, would then present it and discuss it with 273 his or her contact person in the supplier company that would be another person, colleague to the one with whom the solution was originally developed. Looking at learning in this systemic situation, where development work is continuously undertaken with the key customers, also changes the sense of the notion of organization. More than being specific to each individual company, the design organization can be seen as being intra-firm and formed around a part of a vehicle or a functional system. This organization will influence its members' learning in the sense that customer inputs will be permanently present. The learning affecting supplier design staff that participate in such an organization will be very customer-oriented, something that results in a strong alignment to customer needs and an improved anticipation of customer manoeuvres. In this organization, a shared knowledge base between design staff in the supplier and customer firm will be developed through a successive and continuous integration of (1) direct inter-company experience and insights related to different problem solving cycles, and (2) the collective learning taking place between design technicians in the individual companies. Figure 23 illustrates this process of inter-company learning in product development. SUPPLIER CARMAKER Component design Product planning Problem solving 1* Product engineering Process engineering Running Production / Minor modifications Design Project 1 Design technician S1 technician C1 Problem solving 2* Design technician S2 Design technician S3 Design technician C2 Project 2 Design Problem solving 3* Project 3 Problem solving 4* technician C3 Project 4 Design technician C4 Transfer of learning Transfer of learning through formal and informal group work, through formal and informal group work, organizational routines and support structures organizational routines and support structures * Transfer through supplier - customer problem solving cycles Shared knowledge base Transfer of experience 274 Figure 23. Inter-company intra-functional learning in product development. The permanently open communication channels also take the reflection close to the genesis of innovation. As has been argued in the analysis of individual learning, new knowledge always begins with the individual: "A middle manager's intuitive sense of market trends becomes the catalyst for an important new product concept. A shop floor worker draws on years of experience to come up with a new process innovation. In each case, an individual's personal knowledge is transformed into organisational knowledge valuable to the company as a whole" (Nonaka, 1991, p. 97)8 . In the case of designing car components, the above described information interchange through permanently open communication channels is the very breeding ground for incremental improvements of functional solutions as well as for more fundamental innovations. As argued, it allows for anticipation of, and close alignment to future formalized customer needs. Finally, when working on black box parts in a set-based approach as discussed in chapter nine, suppliers must get accustomed to managing customer behaviour that can seem contradictory at first glance. Project managers or design technicians must understand that the freedom given at the beginning of the project will be gradually reduced following the evolution of all the other components that are functionally interrelated to the component in question. For example it might happen that a customer asks for a design change that then is later cancelled. The supplier must learn (develop a collective mental model) not to see the work on this change as a waste, but as an experience to add to the collective knowledge base in product development. In fact, the constraint of working on solutions that ultimately will be left without continuation could be transformed into a competitive advantage by looking at every design effort as an opportunity to learn (learn a new solution, learn that a specific solution doesn't work in a specific context - and why, learn to better anticipate customers' manoeuvres, learn to ask the relevant questions at the right moment, etc.). When relating the concept of permanently open communication channels to the set-based approach it seemed clear that the former was a condition for the latter. The intensive information exchange that took place in the first part of the development process -the definition of the functional concept- was found to being realized largely through direct contacts between design technicians in customer and supplier companies interacting through open communication channels. 8 It can be recalled that this knowledge includes facts, principles, experience-based insights, working procedures, research findings, and ideas (Glaser et al, 1983) 275 10.3 LEARNING IN PRODUCT DEVELOPMENT - SUMMARY CONCLUSIONS AND The analysis of the case studies with respect to learning aspects in the product development process has confirmed the ideas of Hawkins (1994) and Hughes & Chafin (1996) that learning is a flow process that needs to be stimulated and unblocked, and that individuals and organizations learn naturally. By seeing learning as a transversal and continuous process, the perspective moves away both from the idea that a learning organization is something out there to be acquired, and the idea that all learning is a good thing whatever one is learning. The first point emphasizes that besides looking for what is missing in a company in terms of know-how and knowledge, the objective is to focus on what is blocking learning and what would help unblock the natural learning process (Hawkins, 1994). The second point emphasizes the fact that the value of the learning depends on where the learning is taking the company - learning is a means to an end e.g. change in physical organization, change in communication patterns, new ways of working, extended or new capabilities, etc. 276 10.3.1 Summarizing the Research Findings on Learning The research findings developed in section 10.2 are summarized in the framework of questions identified in section 10.1.2, see table 16. Question in the framework for analysing learning What are the transfer mechanisms between individual and organizational learning? Observed features - Direct communication between individuals. - Collective problem solving. - Individuals themselves altering functions or shifting between externally and internally oriented activities. - Direct customer input. What in practice makes the difference between whether single or doubleloop learning occurs? - Personal individual initiatives and as a consequence a climate through guiding visions and performance measurements that encourages initiatives and doesn't make what already exists a rule. - The right to make errors on condition that they are brought up to the surface and their causes analysed. - The environmental reaction to a proposed modification or to the result that a new way of doing things leads to. - Absence of keeping track of the procedure of reaching a solution. What blocks learning in different phases and stages - Project management and customers are not focusing on the how question when in the learning process? assessing proposed design solutions. - Long lead-time between making a proposal and feed-back from the customer. - Distortion of messages due to numerous interfaces. Table 16. Summary of research findings on learning. 10.3.1.1 Transfer Mechanisms Among the transfer mechanisms, direct communication between individuals was a central element. This indicates that it could be useful to look at communication relations as learning relations; where there is communication there is automatically an opportunity for learning. This idea rejoins Hamel's (1991) work on strategic alliances where he argues that alliances are successful for companies that see them as learning opportunities. Hamel (1991) shows that one can approach a collaboration in different ways at the strategic level, i.e. as a learning opportunity or not. Through different means of guidance, this strategic intent in relation to an activity will be mirrored in the individual behaviour of single actors at the operational level. If general management and product development management in supplier firms develop means of guidance that lead operational people to see all interactions with both internal and external collaborators as learning opportunities this could create a significant competitive advantage. 277 However, as concluded from the literature review in section 4.2.3, if professional languages are different, communication is not enough for creating common knowledge. The example of intra-functional learning between product and process engineering technicians showed that it was through collective work on shared problems that learning took place. A design technician who developed a new way of work as a consequence of such a learning process would then be able to transfer it to his or her design colleagues. In this case an individual alternating between different functions would be the transfer mechanism. As Wheelwright & Clark (1992) argue, learning occurs and must be pursued within the context of the development team. In inter-company learning, new insights could also be transferred between design technicians through direct customer input. When individual learning took place through direct interaction between design technicians in supplier and customer firms within the frame of a specific project, another customer design technician, informed of the new solution from within his or her own organization, would present it and discuss it with his or her contact person in the supplier company. The former would, thus, be informed through the customer input playing the role of transfer mechanism. 10.3.1.2 What Makes the Difference between whether Single or Double-Loop Learning Occurs? Personal individual initiatives were found to be decisive for the testing of new alternative solutions outside existing routines and frameworks. One of the basic premises for double loop learning is that a different way of doing things can be tested and evaluated. The attitude towards errors committed by employees is therefore of crucial importance. If a new work procedure, developed on the basis of an individual initiative, is a failure, it was found that an emphasis on analysis and correction in order to avoid repeating similar pitfalls on the part of management was beneficial for an organizational climate open to new initiatives. Another way of promoting creativity and thus double-loop learning was to minimize formalism to situations where it is productive, i.e. new and complex design projects. Concerning the right to formulate problems discussed in chapter eight, the opportunities to question existing routines and the possibility of finding new solutions increases with the number of different actors that can make their voice heard in an organization. Taking into account what are central problems for people belonging to non dominant disciplines can lead to the development of new theories-in-use as in the example of inter-functional learning between product and process engineering technicians. 278 As the dominant guiding vision in the studied companies was customer satisfaction, the environmental reaction to modifications also played a determining role in making the difference between single and double-loop learning. A positive or negative reaction from customers to a different way of doing things will lead to actions in the supplier company speeding up the process of changing mental models. In the studied companies, all these features contributed to the development of a climate that encouraged initiatives and did not make what already existed a rule, making it possible for double-loop learning to occur. 10.3.1.3 Factors Blocking Learning In order to unblock learning mechanisms, the study indicates that it would be useful to develop a better awareness of the art of design work both on the part of the operational technicians themselves and their managers. This reinforces Wheelwright & Clark's (1992) argument that the search for insight about the root causes of problems must be guided by a model of how the development process ought to work. Field observations in the two case study companies indicated that design technicians, engineers and managers should pay particular attention when satisfactory solutions are quickly and easily found. In this case there was a risk that both the process leading to the solution and the solution itself would be quickly forgotten. However, keeping track of the design process leading to different solutions should not be too formalized and time consuming. It could be enough to file a sketch and some notes instead of throwing them away or piling them up without classifying. These observations have an important implication for the management of day-to-day activities from a learning perspective. As the traceability of a line of action is equally important when the environmental reaction to a design proposition is positive as when it is negative, a system of record keeping and assessment of each step in each activity must be ensured - for example through an enlargement of design data bases. Computer-aided systems for information transfer and storage will play an important role as collective memories in this context. These observations are supported by recent research on managing learning processes, e.g. Huber (1996) argues that ongoing increases in 'friendliness' and capability of computer systems are expanding the role of computers as components of organizational memory. A certain negligence of the how, both during in internal and external evaluation of product solutions, was also identified as an obstacle to learning, reinforcing the problems outlined above. A lack of such a focus did not encourage design staff to record and reflect upon the 279 process of how the organization developed its products. The central role of the following factors, previously analysed in the empirical chapters, in the learning process in product development emphasizes the importance of such a focus: • Expose prototypes and mock-ups of both successful and unsuccessful design solutions. In chapter nine physical artefacts were identified as important bearers of design information, • Locate design technicians together, the most frequently quoted driver for new ideas was real-time discussions with colleagues, • Facilitate and encourage collective work but let it emerge naturally instead of imposing group work. Pairs of very small groups (up to five participants) seemed to be more productive than larger groups in the studied settings. Two additional blockages to learning were explicitly addressed by design technicians in the two case study companies during interviews. The first concerned a long lead-time between making a proposal and customer reaction. Due to the fact that design technicians worked on numerous projects at the same time, their tasks were quite fragmented and a long feedback time meant a risk that a specific problem would fall into oblivion. The second was related to problems with feedback on product propositions and distorted messages were identified as blocking learning. More direct contacts with customers on the part of deign technicians was identified as a way of reducing these problems. 10.3.2 Benchmarking the Research Findings on Learning After the analysis of individual and organizational learning undertaken in this chapter it is interesting to benchmark the case study findings with the model of Nevis et al (1995) of organizational learning (table 15). In table 17 the learning orientations are exposed in the first column, the second contains the specifications from Nevis et al (1995), and the third exposes the case study findings in relation to each learning orientation. Learning Orientations Specifications 280 Research findings 1. Knowledge source: Internal - External Preference for developing knowledge internally vs. preference for acquiring knowledge developed externally. The studied companies generally preferred to develop knowledge internally but based this on strong interaction with external actors; customers, interfacing suppliers, research laboratories, etc. 2. Product-Process Focus: What? - How? Emphasis on accumulation of knowledge about what products/services are vs. how organizations develop, make, and deliver their products/services. The first orientation was the traditional, one embedded in the studied supplier companies' cultures, the second emerged as a new and additional emphasis. 3. Documentation mode: Personal - Public Knowledge is something individuals possess vs. publicly available knowhow. Knowledge is possessed by individuals but must be made publicly available. 4. Dissemination Mode: Formal - Informal Formal, organization-wide methods of sharing learning vs. informal methods, such as role modelling and casual daily interaction. Hybrid approach where informal dissemination methods e.g. casual daily interaction are facilitated by developing support means and structures as a function of natural working practice. 5. Learning Focus: Incremental-Transformative Incremental or corrective learning vs. transformative or radical learning. Integrating the two in interlocking cycles where one is a condition and result of the other. 6. Value chain Focus: DesignDeliver Emphasis on learning investments in engineering/production vs. sales/service activities. Not applicable due to the focus of the study. 7. Skill Development Focus: Individual - Group Development of individuals' skills vs. team or group skills. Individual skill development is necessary for development of organizational capabilities. Organizations cannot learn without their members learning. Table 17. Specification and analysis of seven learning orientations (after Nevis et al, 1995). The research findings indicate that the model developed by Nevis et al (1995) should not be seen as binary. The items were all relevant to the learning problem observed in the field (except for point six due to the focus on learning in engineering), but learning was perceived as a much more integrative process. For example, the case study companies tried to extend the learning focus from product technology - the what in product development- to the work process of realizing the product - the how in product development. Because the product is the result of the process and the product's performance determines if the process is effective or not, both are interrelated. Neglecting the how question in both internal and external evaluation of product solutions was identified as blocking learning. Concerning documentation and dissemination, knowledge is basically retained by individuals. An effort to make it public is therefore crucial for learning to take place. Informal dissemination modes seemed to be the most relevant for this, but they must be facilitated through formal means and structures. Finally, more than focusing learning on incremental improvements or 281 transformative change, these two features were found to be integrated parts of the organizational learning process. Incremental learning improves an actual situation until its performance limits are attained and, thus, informs operative people and managers that there is need for transformational change. After a transformational change, incremental learning comes in to improve the practice until performance saturation, and so on. The analysis of learning also illustrates the interconnections to the previously discussed phenomena, i.e. guiding visions, and design support structures. This is also emphasized in the literature, e.g. Wheelwright & Clark (1992) identify principles that the organization uses to guide decision-making and development activities -an item that could be classified as pertaining to the first phenomenon- and procedures, process and structure -items that could be classified as pertaining to the second phenomenon- as critical for learning. 282 11 CORE CAPABILITIES In chapter four, section 4.3.1 the concept of core capabilities was identified and analysed as a driving force for outsourcing on the part of carmakers, and, therefore, one of the causes for increased product development responsibility for suppliers. Capabilities were identified as being built up of basic resources combined in unique ways in each different company. Based on this, Grant (1991) defined a capability as the capacity for a team of resources to perform some task or activity. Core capabilities were then defined as capabilities that differentiate a company from its competitors and that are connected to real customer needs. In order to provide competitive advantage they should be durable, and difficult to identify, understand and transfer for competitors. Moreover, a company should possess clear ownership and control over its core capabilities. This literature review led to the conclusion that the core competence strategy might have a snowball effect in the entire supply chain - direct suppliers would try to identify their own core capabilities in order to develop specialized expertise within a specific field of technology. The field study confirmed this hypothesis. In the studied companies, explicit strategies for developing expertise in product development through the identification and building of core capabilities were found. Hence, a strategy of core capabilities in the industrial process of car manufacturing, seems to be leading to a renaissance of the component suppliers in terms of their engineering capacity. Data from the field study show that this strategy has been beneficial for those suppliers that immediately decided to play the game of partnership and that have learnt how to take advantage of it. This was the case of the two case study companies and as illustrated in appendix five, their turnover has developed very positively during a period marked by the integration of lean principles (1990-95). Specific capabilities related to product development are the very essence of an expert supplier strategy (c.f. section 6.3). The data confirm that top management are aware of the necessity of identifying and developing the specific skills and resources that will allow for a more clear profiling of their companies in the relationships to customers (c.f. table 14). A central question that interviewed managers came back to several times was that of the role that their supplier company played for the customer, and that of what picture customers would have of them i.e. what the studied supplier company meant for the customers. What interests managers in all industries is to develop their company's turnover and profitability. In all of the studied companies the means for achieving this passed essentially through two action strategies explained by the interviewed managers: achieve as high a 283 customer satisfaction as possible; and stay ahead of competition by a positioning in the front line of new product technology development. The first action strategy relied on alignment and adaptation to customers' demands and manoeuvres passing through a capacity to listen. The second relied on anticipation of evolving customer needs, and an ability to exploit new markets and businesses, or generate new ideas and incorporate them into the product offer. The resource-based theory and the core capability concept play an important role in this context. However, what was lacking for General Mangers and Product Development Managers in the studied companies was a better picture of the relationship between the recognition of the concept and the formulation of strategic goals for identifying and building core capabilities, and the operational reality of daily product development work. Therefore, the analysis of the core capability concept will concentrate on the emergence of capabilities in order to develop a framework of how capabilities are identified, developed, and nourished. The data analysis positioned core capabilities as a consequence of the management of the product development process, where means of guidance play an important role for orienting action strategies, and where learning is a vehicle for integrating discreet capabilities inside the firm and ensuring a dynamic evolution of the relationships to customers. This more concrete perspective on core capabilities -seen as a tangible outcome of the management of the product development process- together with the definition from Stalk et al 1992 that core capabilities are connected to real customer needs, help to demystify the concept and develop a framework for how, concretely core capabilities can be built. 11.1 CORE CAPABILITIES - AN EMPIRICAL FRAMEWORK How core capabilities can emerge will be illustrated through an example coming from one of the case study companies. In accordance with the theoretical framework outlined before and above, the capability building begins with a customer need, in the present case that of reducing the number of direct suppliers. The example considers a reconfiguration of a functional system where the studied supplier supplied the core component, i.e. the component that determined the system's performance perceived by the final customer. The customer relationship was one of direct expert supply with a strong presence of tier position competition between the different complementary expert suppliers contributing to the system. Thus, the process that will be described here originated in a reaction to the tier dilemma, identified in chapter six as the main difficulty to manage in direct expert supply. In product innovation driven by the tier-dilemma suppliers need to widen their offers, i.e. provide a more complete function, without falling into the trap of becoming a pre-assembly unit at the expense of engineering capability. The formulation of guiding visions played an important role for how 284 this problem was dealt with. The guiding vision in COR 'Extension of product functions, i.e. development of mini systems should be done as a result of innovative design and not only as a result of integrating assembly' can be compared to the following guiding vision registered in one of the interview companies 'The objective is to assemble parts in order to supply functions'. It is obvious that in the latter case there is a risk of becoming more a preassembly unit and less a development intensive expert supplier. For the company in which the example originates, the way of reducing this problem was to focus on an extended function from an engineering perspective more than from an assembly one, and integrate the design of neighbour components to the core product. Ideally, if assembly needs to be developed it should be a result of innovative design. However, achieving an extension of functional solutions is far from easy. As Karlson (1994) argues, only a product that fits into the organisation can be designed (p. 131-135). In a context where extension of product functions is essential, the ability to rapidly undertake organizational modifications, and also modifications in manufacturing equipment, becomes a significant competitive advantage for an expert supplier competing in tier position. The core capabilities that the supplier needed to build to compete in this context would therefore be related to such an organizational dynamic. Emergence of Core Capabilities - an example from the case study of COR The example of a successful extension of functional solutions concerns a technical component that traditionally was assembled into a plastic cabinet. Both the supplier of the technical component and the supplier of the plastic cabinet were considered as experts by the carmaker customer. Nevertheless, the latter wanted to reduce the couple to only one interlocutor. The most natural evolution would be to choose the plastic manufacturer, as the technical component traditionally had to be ready before being fitted into the cabinet. However, the other supplier responded by including the plastic cabinet in the design study, and succeeded in designing a cheaper overall solution with higher performance through this widening of the engineering activity. The means of guidance framing this effort were a guiding vision of developing mini systems, and performance measurements enhancing the value of inter-functional problem solving. Concerning the design, both the technical component and the plastic cabinet were completely reconfigured. Instead of what happened previously, when the carmaker presented the technical component to the plastic manufacturer to find a solution, the supplier of the technical component succeeded in making the plastic cabinet an integral part of the function and not only an interface element. 285 Besides employing their traditional design know-how, the studied company had to make an important investment in process and material technology scanning in order to ensure the possibility of manufacturing the new product at a reasonable cost. This meant that process development and manufacturing constraints became central problems in this development project, something that contributed to a modification of existing theories-in-use both for individual and collective operative design work. As a result of the global line of action described above, the supplier of the technical component -the studied supplier company- remained a direct expert. Naturally, it will be necessary to develop an assembly capacity for the integrated function, but this will happen as a result of innovative product design and is not only taken on as a necessary evil to stay in direct contact with the carmaker at any price. Besides providing a concrete framework for analysing the emergence of core capabilities, this example illustrates two other points that are useful to explain before coming back to the core capability concept: • A solution to the classical trade-off between simultaneous customer requirements of superior functional performance and lower cost (Abernathy & Utterback, 1978) is proposed by the example. It is true that the price that the customer had to pay for the integrated component was higher than the price paid for the two separate components used before. This was only true in terms of direct cost, however. When the entire cost picture was taken into account, the trading cost for the plastic manufacturer disappeared, as did the competitiveness one (c.f. section 4.2). Total cost was therefore likely to decrease and so was the industrial risk of being dependent on a large number of suppliers, and the risk of quality problems that increases with the number of interfacing components. At the same time the product's technical performance was improved. • The example also illustrates that inherent in the restructuring process of the supply chain are mechanisms that push the development of new components towards integration of technical functions and thus towards exploiting technology synergy. Technology synergy was, thus, identified in the study as a consequence of the restructuring of the automotive production chain where different mechanisms such as customers' outsourcing strategies and horizontal competition between complementary expert suppliers push the development of new components towards integration of functions. As discussed in section 4.3.2 this synergy often results in component systems with superior performance compared to earlier solutions as a consequence of the fact that an association of R&D activities within different 286 organizations will promote the process of producing exploitable innovations. This was exactly what happened in the example analysed above. Medium sized expert suppliers are directly concerned with this evolution and it means for them a new managerial challenge: that of maintaining systematically developed core capabilities in product design by integrating neighbour technologies into the know-how of the firm. 11.2 HOW CORE CAPABILITIES EMERGE - AN INTERPRETATION How, then, is this example related to the resource-based approach to strategic management and the development of core capabilities? In section 4.3.1 it was argued, based on Arrègle (1995), that a company can develop an advantage by combining basic resources in an innovative or original manner. The basic resources exist at the level of an individual or a group, for example within a function of the company. Bowen et al (1994) identified four interdependent dimensions of capabilities: knowledge and skills embodied in employees, managerial systems, physical systems, and values. The first three were identified as a combination of basic resources, while the last one could be identified as the 'glue' of attitudes, behaviour and norms that held the others together. Through this definition, capabilities were identified as a systemic notion. Capabilities were also identified as dynamic features, Arrègle (1995) argues that they take form from the interaction of technology, collective learning and organizational evolution. In the example, a combination of existing resources in product design and new process technology development resulted in a combined resource of integrating production constraints in individual and collective operational design work that was to become central for the core capability that would emerge. This combined resource interacted with skills in material technology and process technology scanning, physical systems such as the design data base and the internal and external information systems including EDI, and management systems like weekly design reviews, functional rotation between product and process engineering, and written procedures for the product development process. All this was held together through a shared framework, or value, which was materialized through the guiding vision of developing mini-systems as a function of innovative design. The value dimension represented the individual and collective theories-in-use identified as central for organizational learning. Therefore, their influence is also central for capability building. As discussed in chapter nine, they influence how people look at problems and perceive priorities, what are the dominant and non-dominant disciplines in a company, and, thus, who will have the right to formulate problems and determine what are good solutions. Then, the dynamic feature of technological development, in this case essentially in process technology for plastic manufacturing, made it possible to realize a more complex plastic 287 cabinet than traditionally without excessive cost. A collective inter-functional learning process also intervened around the guiding vision of customer satisfaction that resulted in a mobilisation of previously unconnected expertise in product and process technology and the solution of conflicts of interest and priority that inevitably occur due to functional loyalty. This was a question of organizational double-loop learning because the positive outcome of this experience resulted in a new shared framework of integrating product and process technology scanning. A process of organizational evolution brought with it a larger amount of close interaction with the customer in the design of components through the specific communication structure of permanently open communication channels discussed previously. Finally, if the emerging capability matched customer needs and also provided a differentiation in relation to competitors, it could be qualified as a core capability. All these items taken together illustrate the formation process of a core capability in system design. The tangible outcome was the successful market introduction of the product described above. This core capability then led to the development of new and/or modified systems and skills and new and/or modified resources. For example, a new resource in the form of an integrated product and process scanning activity was developed, and a new shared skill embedded in employees of functional thinking was created. This meant that there was a feedback from the new emerging capability to existing resources, systems, skills and values. As proposed by Arrègle (1995) this combination resulted in a development and refinement of existing resources, or even in the creation of new ones. The new emerging core capability also provided a strategic feed-back in terms of development and/or modification of strategic objectives. In the case study company, the success of the effort of extending functional solutions resulted in a generalization of the objective of developing mini-systems and a reflection concerning the structural support for such activities was launched. Finally, it provided a business process feedback developing and/or modifying the dynamic processes of technological development, collective learning, and organizational evolution. The importance of such a feedback was discussed in the case of the learning process in chapter ten where it was argued that the environmental reaction to the outcomes of learning processes influences the original learning. In section 10.2.1 it was concluded that if the feedback from the environment is positive one will try to know how the solution was elaborated in order to repeat it and make it a new best practice. If the answer is negative one will also try to know how the solution was elaborated in order to avoid making the same mistake and unlearn an undesired line of action. This process occurred in the case of the product developed in the above example. The systemic nature of capabilities, with multiple interconnections between different factors was illustrated by the fact that the feedback influenced learning processes, which then will 288 modify existing theories-in-use, which, in turn, will lead to a modification of the values that were identified earlier as triggers for capability building. The process is illustrated in figure 24, which presents a proposal of a model of the emergence of capabilities based on the above analysis. CORE CAPABILITIES Match with customer needs Differentiation from competitors CAPABILITIES Technological Development Business process feed-back Collective Learning Organisational Evolution Management Systems Physical Systems New systems/ skills/values Modified systems/ skills/values Values Skills Combined Resources Central for Core Capabilities New resources Modified resources Strategic feed-back Strategic Objectives translated through Means of Guidance Basic Resources Figure 24. How core capabilities emerge - a model integrating research findings with the theoretical frameworks of Bowen et al (1994) and Arrègle (1995). The model integrates the different identified elements from the literature review on core capabilities through a detailed analysis of capability development in one of the case study companies. It also shows that core capabilities make use of the three different elements previously identified and analysed as categories related to operative design work in the modelling of the product development process proposed in the present research: • •Means of guidance intervene as a decisive factor, that make critical strategic issues explicit to all collaborators in a company, and, thus, for developing relevant work processes and activities corresponding to these strategic issues. In the example analysed 289 above, the guiding vision placed the development project in the context of total company capabilities. Through the inter-functional focus, the emerging capability also provided for progress on the agendas of all the relevant functions as the company moved in new strategic directions. Bowen et al (1994) identify this as a critical issue of core capabilities. • The importance of support structures in the building of core capabilities was emphasized through the central role that management systems and physical systems were found to play. • Finally, learning intervened at two levels in the emergence of core capabilities. Firstly, the platform of existing skills, comparable to the common knowledge base discussed in section 10.2.3, determined the starting point from which capabilities could be built. Secondly, a dynamic collective learning process was identified as a necessary condition for constantly refining, developing, and modifying existing skills, management systems, and physical systems. This process, that also contributes to the dynamic evolution of the organization, will result in the development of capabilities that make it possible for the company to progress in the front-line of product technology, organizational performance, and customer satisfaction. The identified capability of integrated design in the case study company was a systemic notion, i.e. created by a complex network of firm resource factors comprising numerous direct and indirect links between a large number of factors. It depended on the organisational dimension and applied components belonging to different functions of the company. In this way it mirrored the definition of system resources from Black & Boal (1994). The network notion is important to understand correctly. By network, Black & Boal (1994) mean "the configuration of factors, as well as, their relationships with each other that result in a particular firm resource" (P. 134). Including both factors and their relationships is an important distinction in their definition. Capabilities thus represent the knowledge that the enterprise has created and possesses as organisation. This double aspect of factors and their relationships and, thus, the particularity of the capabilities that emerge in different individual companies was clearly confirmed by the analysis of the process in the case study company. The analysis undertaken above gives a content to several conceptual statements in the core capability literature such as the argument that the key to capability building is the identification and linking together of essential business processes to serve customer needs (Stalk et al, 1992). The field study confirmed that it is the understanding of the inter-relationships and systemic nature of different organizational components such as processes, systems, behaviour, values, etc. that distinguish successful capability-based competitors. The above analysis of core capability building in an expert supplier firm also confirms the correctness of the idea that in the case of black box and supplier proprietary parts, suppliers possess significantly more knowledge about the design and manufacture of components than 290 does the carmaker (Dyer & Ouchi, 1993; Lamming, 1993). Thus, carmakers become increasingly dependent on the supplier. This reversed dependency is important for suppliers to exploit. When recalling the problems of confidentiality analysed in section 6.4.4, the research findings indicate that well-managed core capabilities could act as a means for making customers more development dependent on a supplier. Finally, the core capability in question also presents the characteristics of a hard-to-imitate organizational capability that distinguishes a company from its competitors in the eyes of customers and allows for the development of more solid competitive advantages (Stalk et al, 1992). For example, it fulfilled Barney's & Griffin's (1992) four point framework for needed characteristics of core capabilities: 1. The capability fitted the company's strategy and this strategy responded to the external environment, 2. The capability was rare due to the rareness within competing firms of the resources that built it up, 3. The capability was difficult to imitate due to restricted information about how to recreate the specific combination of resources that gave a synergistic result, and about the cost of doing so, 4. Finally, the entire organization was oriented towards the utilization of the capability; it was commonly recognized as useful and all concerned actors could appropriate it. Besides these criteria, the criterion of satisfaction of customer needs through capabilities must be strongly emphasized. In other words, a core capability must directly fit the external environment and when this happens it will provide a feed-back that will guide future strategic decisions. By consequently anchoring capability development in strategic objectives and permanently benchmarking emerging capabilities to customer needs, managers can avoid the problem discussed by Black & Boal (1994) of identifying strategic important system resources only after they have been unintentionally destroyed. 11.3 CORE CAPABILITIES - SUMMARY AND CONCLUSIONS This chapter has specified the notion of core capabilities through an integration of theoretical concepts and empirical observations. The development of core capabilities plays an important role in the car component sector as the entire automobile industry is going through a process of redefinition of the roles of different actors based on an outsourcing logic. 291 If the concept of core capabilities was found to be easily understood by General Managers, its application and realization were perceived as more difficult. The objective of the analysis was therefore to propose a model for the emergence of core capabilities. Parallel to basic resources, existing at the level of an individual or a group within a specific function in a company, clear strategic intentions and objectives based on alignment to and anticipation of customer needs were identified as the building blocks of capability development. The strategic intentions and objectives relied on relevant and consequent means of guidance in order to be efficiently disseminated in the organization. The interaction between basic resources and guiding visions results in combined resources that can be either of skill, management system, or physical system type. The combined resources already in place will naturally influence the emerging capability. In other words, support structures such as information systems, interfunctional coordination, and formal and informal communication structures actively contribute to the formation of capabilities. The more intangible dimension of values played, as in the case of organizational learning, a central role in the formation of capabilities. Collective learning was also found to be a necessary condition for the development of core capabilities. This learning had, however, to be oriented towards specific objectives which means that no coherent capabilities will emerge if means of guidance are missing at the basis of the capability building process. Guiding visions indicate what core capabilities should be built in order to satisfy the business strategy. They influence the cognitive side of development work through a contribution to the development of shared mental models of how to approach product development work. Performance indicators influence the operational side of development work and, as emphasized in chapter eight, they must be adapted to the strategic objectives of core competence development. Besides collective learning, two other dynamic business processes identified in the literature technological development and organizational evolution- were confirmed as being important for the development of core capabilities. Technological development, where a single company of the kind studied here normally only has a marginal influence, must be followed up on through intensive collaboration with customers and, if possible for competitive reasons, with interfacing suppliers. Moreover, as argued in chapter nine, technology and patent scanning activities must be organized in a relevant way in order to facilitate the integration of technical information in operational design work. Organizational evolution such as development of team work, and changing roles of different actors provides another dynamic input to the development of core capabilities. The example illustrated how a greater number of direct contacts between design technicians in supplier and customer firms could promote the development of new functional solutions. 292 Capabilities emerging through the process identified in the present research must be matched to customer needs and benchmarked against the capabilities of competitors. If customers provide a positive feedback concerning the outcomes of a specific capability, and the latter seems to differentiate a supplier from its competitors, the capability in question can be qualified as a core capability. Once a core capability has been identified, different feedback mechanisms will intervene. Four different types of feedback were identified: 1. Feedback modifying or creating new basic resources, 2. Feedback modifying or creating new systems, skills, and values, 3. Business process feedback influencing the dynamic processes of technological development, collective learning and organizational evolution, 4. Strategic feedback directly influencing the elaboration of new or modified strategies and strategic objectives. As Stalk et al (1992) argue, capabilities are collective and cross-functional - a small part of many people's jobs, not a large part of a few. Because a capability is 'everywhere and nowhere', no executive or manager controls it entirely. The present research indicates, however, that the process of building capabilities must be framed and guided by different means of guidance, support structures and systems, and an active management of the learning processes. For example, Bowen et al (1994) emphasize that when making their design and development decisions, team members must have two objectives in mind: the product that will be delivered to a particular set of customers, and the company capabilities to which they are contributing. This double output of the product development process must be kept in view in guiding visions. Generally speaking, the analysis of the emergence of capabilities indicated that building strategic capabilities should be the primary agenda for top management. The systemic character of capabilities and their anchoring in the operational reality of a company means that management ought to think of business processes as the primary object of strategy (Stalk et al, 1992). In relation to the development of core capabilities, the performance goals that a company is trying to reach must continuously be set higher. If one focuses on a given practice at a given moment the current practice should be the object for continuous improvement activities. However, the performance interval for a given practice is normally limited. In order to go beyond this interval, a reengineering of the activity must take place. The data were rich in such examples: the installation of project teams and project managers in COR and COF, the integration of research and innovation in the design department and the allocation of specific resources to this particularly observed in COF, the integration of assembly activities in both of the case study companies, and the extension of functional solutions of a firm's products in 293 COR. The dynamic character of core capabilities will contribute to this continuous redefinition of different practices and routines. In the modelling of the product development process the analysis of the core capability concept has confirmed the position of core capabilities as a consequence of a well-managed product development process. The process of capability building integrates the previously identified and analysed elements means of guidance, design support structures, and learning. 294 12 SUMMARY, PRESENTATION OF THE INTEGRATED MODEL, AND GENERAL CONCLUSIONS This final chapter recalls the research purpose and the framework of research questions grounded in the literature review, summarizes the discussions and the results from the empirical chapters, relates them to the research questions, and presents the integrated model of the product development process grounded in the empirical research. Furthermore, it addresses the limitations of the research, and proposes areas for future research. 12.1 THE RESEARCH PURPOSE The purpose of the present research was to explore how supplier firms in the automobile industry, of a specific category identified as expert suppliers, are experiencing changes related to the more active part that they are supposed to take in the product development process within the context of lean production. More precisely, the objective was to describe the organizational structure of the product development process from a perspective from inside such supplier firms. To fill an identified gap in the analysis and conceptualization of product development and supply management in the auto industry, a research perspective focusing on the suppliers' situations and on the understanding of the new product development map by looking into organizations at an operational level was adopted. A basic assumption motivating the choice of the industrial sector was that the development of lean production principles is beginning to stabilize in the car industry, something that makes it possible to study factual changes and organizational adaptation to this new context. Furthermore it was assumed that the increased design responsibility concerns a broader range of suppliers besides system suppliers, above all those that are rich in technology and have a long experience in direct supply - so called expert suppliers. This organizational change will lead to a more and more integrated way of managing the product development process. The research took place in two steps. The first step focused on the lean production context trying to answer three global research questions: 1. How does the emergence of new industrial principles take place in expert supplier firms? 2. What is the place and the role of expert suppliers in the automotive production chain? 3. What lean production techniques are used, to what extent, and are managers satisfied with their application? 295 In order to answer these questions, four themes for data collection were identified: • The place and the role of the firm in the supply chain, • The drivers for change, • The change and development tendencies in the product development process, • The use and perception of lean techniques. The second step focused on the organization and strategies for product development in expert supplier firms trying to answer the remaining three overall research questions: 4. How in practice is integrated component development realized? 5. How are work processes designed to support integration? 6. Are there common characteristics between companies forming a 'high performance' organizational structure for lean component development? These questions were addressed through data collection under three themes: • The evolution of product technology and product functions; • The organizational structures and work processes; • The coordination activities and communication structure. Through the literature review in chapter three and four, the research questions were further specified in the form of additional questionings and focusing of the topics in question, c.f. section 4.6. The research questions were to a large extent 'what' questions of an exploratory kind or 'how' questions. The objective of the research was to explore and describe, to focus on meanings, to look at the totality of and the interconnections between different studied situations, and to develop ideas mainly from induction of data. A qualitative research perspective based on a phenomenological philosophy of research was therefore adopted. In the first step, dealing with the lean product development context, the data collection method was semi-structured interviews. In the second step, concentrated on the product development process, it was indepth case studies and follow-up interviews.. 12.2 RESULTS: THE LEAN PRODUCT DEVELOPMENT CONTEXT The results of the study will be presented corresponding to the two conceptual steps, i.e. the lean product development context and the organization and strategies for product development in expert supplier firms. In the first step, the results will be summarized following the section 296 headings in chapter six. Finally, some general comments on the results concerning the lean product development context will be made. This section starts by answering research question two, because the customer relationships emerged as the most important feature when managers started describing the difference between traditional and new "lean" product development. Research question one and three will then be answered through a successive resume of the different problem areas related to the lean product development context. 12.2.1 The Place and Role of an Expert Supplier in the Supply Chain The main result was that the reality of the relational structures was much more complicated in the studied companies than the tier model indicates. Not only must supply be distinguished in terms of components on the one hand and development intelligence on the other, but one and the same supplier was found to have a variety of customer relationships to manage. A typology of three different situations was identified from a transversal comparison of interview data from the eight studied companies: • Direct arm's length supply, • Indirect expert supply (triangulation), and • Direct expert supply. The objective to remain in the first tier, shared by all of the studied companies, was immediately related to the supply of development intelligence. Thus, even though the triangulation situation corresponded to an indirect supply of components, it was considered as a satisfactory situation. The three different situations, that all had their specific raison d'être in the product development strategy, focused the attention on a range of different phenomena influencing the product development process: • Triangulation was the most stable situation in the tier structure. The reason was that an outsourcing logic had already been applied. • The direct expert situation presented a specific difficulty in terms of a phenomenon labelled the tier dilemma. This means that complementary expert suppliers are in constant competition to remain in direct contact with the carmaker. This pushes suppliers to widen their product offering and develop mini systems. If this isn't done through innovative design it could comprise a risk with taking on assembly at the expense of engineering. 297 • Both triangulation and direct expert supply raise new demands on communication in product design. There was a need for frequent and rich information exchange allowing for a strong alignment to, and anticipation of, customer needs. The analysis of customer relationships in the supplier chain indicates that before making any conclusions concerning the situation of a specific supplier company, different situations must be studied in more depth. As the core element of a first tier strategy was direct supply of product development intelligence, a first tier situation with mostly arm's length relations might be much less profitable for developing product development expertise than having a triangulation relationship to the majority of customers. 12.2.2 Driving Forces for Change The three classical parameters of quality, cost, and lead-time were identified as the main driving forces for change that the studied suppliers had experienced over the last five to ten years. Concerning just-in-time, managers were satisfied with the measures undertaken for developing just-in-time and the productivity improvements that these actions had led to. The major problem was to balance JIT delivery with JIT manufacturing. The results here indicate that the imperatives of lean production must be treated with some moderation. In some situations, JIT manufacturing was simply not possible to defend economically. As a consequence, it is possible to conclude that managers did not blindly apply lean principles but preceded their application with careful analyses of each particular situation. Concerning quality management, the main managerial preoccupation was that of responding to different quality standards. Being certified according to an industrial standard, i.e. a set of rules for the entire organization of the supplier company, was a must for all of the studied companies. In relation to the issues of how buyers and suppliers are linked together, and what integrated management systems can be used to build trust and goal congruence between companies, the research findings indicate that quality standards constitute the main formal link between customers and suppliers. Instead of cross equity and inter-company career paths practised in Japan, French and European carmakers seem to rely on this contractual reassurance of a certain organization and managerial structure. Cost as a driving force for change in the strategies and organization of expert supplier firms was translated through the pressure for cost reductions on components. Within the contractual framework that tied suppliers to their customers, productivity targets translated into yearly price reductions on components were constantly present. The pressure was found to be strongest in direct arm's length relationships where there were few possibilities of negotiating 298 with the customer, for example in order to propose technical modifications of a component in order to reduce the price and maintain the supplier's margin. An interesting result in this context was that there was a trade-off between demanding price reductions and refusing modifications necessary to achieve these price reductions without reducing the supplier's margin. As this behaviour also has a negative affect on innovativeness, it adds another disadvantage to this relationship configuration in relation to the development of engineering design capabilities. 12.2.3 Change and Development Tendencies in the Product Development Process The dominating tendency where things were still far from being stabilized was the increased product development responsibility for suppliers. The continuation of the restructuring of the automotive supply chain in terms of a reduction of the number of direct suppliers (direct component supply and direct intelligence supply), a transfer of design activities to suppliers, and an evolution of integrated structures was confirmed in the study. Besides system suppliers, managers anticipated that only expert suppliers would remain in direct contact for three main reasons: • The technical functionality of the component, i.e. when the component represents the core technology in a system, • The material properties of the component. As the use of new materials often lead to a reconfiguration both of a product's functional performance and the related production process, and as the use of new materials is an important source for innovations, carmakers seemed to wish to preserve direct design links to suppliers responsible for components where material and/or process technology often changes, • The interfacing qualities of a component. The modification of interface components such as electric cables, fastening devices, and high precision engine components, lead to important chain reaction in a functional system. Without direct development relations to suppliers of such components the risk of high coordination costs increases for the carmaker. If a supplier wishes to maintain or reinforce its expert position it will be interested in focusing on components corresponding to one or several of these categories. Another anticipated change was that of changing organizational boundaries in the product development organization. When expert suppliers become more and more responsible for the engineering of a vehicle, the product development function will need to be integrated with other interfacing suppliers, system suppliers, and carmakers on a project basis. Five action strategies in product development management for responding to these change tendencies were shared by all of the interviewed General Mangers. These were: (1) promoting 299 individual and organizational learning with the objective of (2) identifying and concentrating on core capabilities; (3) improving the capacity of integrating change with the objective of speeding up learning and unlearning cycles; (4) increasing customer orientation with the objective of improving alignment and anticipation in relation to customers; and (5) developing a culture of continuous improvement in order to formalize best practices and avoid the repetition of errors (a condition for developing learning). The assessment of change and development tendencies finally revealed that a real shift in the nature of buyer-supplier relationships was beginning to occur. A greater number of long-term and permanent relations were being developed, and the new type of relationships had obliged supplier firms to restructure their organizations. Substantial problems remained, however, in the fields of pricing and confidentiality in new product development projects. 12.2.4 Use and Perception of Lean Techniques The research assessed three different areas of application of lean techniques identified in the literature review. In the design organization project management and inter-functional cooperation was applied in all of the studied companies. The project organization was generalized in all of the studied companies in accordance with the demands in the quality standards. However, it was applied with moderation, only as a function of real customer needs. Once again this showed that the companies applied lean techniques in a way that was adapted to their needs. The main advantages of design for manufacturing, i.e. reduced development lead-time and improved product quality were confirmed through the interviews with General Managers and product Development Managers. Concerning the production arrangement, specific strategies for flexibility were found in two areas. Task enlargement and task enrichment leading to flexible competencies of the workforce were developed in all of the studied suppliers. The second strategy concerned flexible production flow groups in order to balance production flows and reduce inventory. Strategies for worker empowerment were also generally observed. Internal career paths favoured the integration of experienced production staff as project managers in both of the case study companies. This indicated that suppliers were making an effort to promote design for manufacturing. Continuous improvement activities were considered an important indicator of performance both for managers employees and technicians. An explicit objective with weekly project reviews was to identify problems, propose and discuss solutions, and generalize these experiences to parallel and forthcoming projects. Concerning feed-back structures, it was made very clear that in order to progress in terms of employee commitment and continuous improvement, listening to employees is not enough. Relevant feed-back concerning the managerial reactions to propositions must be 300 ensured. If, for example, demands or propositions remained unanswered because they had little relation to current priorities in terms of customer satisfaction, this fact had to be explained to employees. Problems with lean techniques, finally, were concentrated in two areas. The first, price pressure, has already been discussed. The second concerned confidentiality during new product development. Internal confidentiality between parallel projects conducted for different carmakers was managed through a structural separation in different project groups. The main problem with external confidentiality was when carmakers asked a supplier to propose a product solution and then exposed this solution to competitors that could develop them as detail-controlled parts, and, thus, were able to propose a much lower price as they would have no development cost to support. In the absence of the 'creative tension' present in some Japanese cases, suppliers were found to need to become more severe towards customers practising this behaviour. In order to stress that this behaviour is unacceptable, suppliers could refuse to participate in other development projects with the customer in question. This of course, was only possible if the supplier disposed of a broad customer portfolio. In the long term, the only solution to this problem that emerged through the study would be to develop and refine the supplier's expert profile making customers' development dependent and more sensitive to the reactions of suppliers. 12.2.5 The Lean Product Development Context - Concluding Remarks In the light of the data in the present research the first three research questions concerning the emergence of new industrial principles in expert supplier firms can be answered in the following way: • The changing relationship to customers has been the trigger for developing new product development strategies and a different product development organization. Within the framework of a typology of different customer relationships that one and the same expert supplier normally has to manage, the main driving forces for change have been lead-time reduction, quality improvement, and price reduction of components. • Main lean techniques as identified in the literature review were to a large extent satisfactorily applied, both at the managerial and operational levels in the studied supplier firms. The most appreciated aspect was the possibility of engaging in technical negotiations directly with customers. Moreover, lean techniques and the transformation towards lean principles in general had not been implemented 'by force' or without considering the particularities of different firms. The fears that lean production would be imposed without further reflection on its suitability seem ungrounded in the light of the present study. 301 • Quality standards play an important role as a formal governance structure for buyersupplier relationships. If a formal substitute for the Japanese keiretsu system were to be designed it would be thought the contractual quality standards. • The main difficulties that the studied expert suppliers faced were related to price cuts (without possibility, for different reasons, of rethinking the component or its production process), external confidentiality (design studies are diffused to competitors), and to an additional driving force for change -increased product development responsibility- which is currently dominating the evolution of strategies and organization in expert supplier firms. 12.3 RESULTS: THE PRODUCT DEVELOPMENT PROCESS This section will firstly summarize the findings concerning the categories developed in the modelling of the product development process. Secondly, special emphasis will be put on the relationship between these categories leading to the proposition of an integrated model, based on the structure of the paradigm model (Strauss & Corbin, 1990, c.f. figure 14). Even though the categories in terms of causal conditions, context, intervening conditions, and consequences emerged through the analysis of operational design work, they will be presented here in the order defined by the paradigm model, i.e., beginning with the causal condition means of guidance. Action/interaction strategies intervene at every different phase in order to handle and manage different phenomena and will be treated as transversal in the model. Section 12.3.1 - 12.3.3 answers the research question four concerning how in practice integrated component development is realized, and the design of work processes. Section 12.3.4 answers research question five by summarizing how knowledge creation takes place, how different actors in the buyer supplier interface communicate and what factors facilitate and block learning. Section 12.3.5-12.3.6, finally, dealt with the question concerning a 'high performance' organizational structure. As stated in section 4.6 this question can not be fully answered. However, one part of a high performance supplier is its ability to develop core capabilities. The findings concerning core capability building therefore provide a partial answer to this question. Moreover, core capabilities are the outcome of the management of the product development process in an expert supplier firm. The integrative model of the product development process presented in section 12.3.6 therefore provides indications of how to develop high performance within the problem area of the research. 12.3.1 Causal Condition - Means of Guidance 302 Causal conditions are features that give rise to a phenomenon. Different means of guidance were identified as mediating factors between different external driving forces -influencing and forming the product development organization and strategies in the case study supplier companies-, and the operational reality of product development. The means of guidance could be separated into two categories; guiding visions and performance measurements. The notion guiding visions was adopted from existing literature and further developed through the research. The main result was that means of guidance play a central role for influencing the cognitive side of development work and implicitly guide the numerous design decisions that operational people have to make in everyday activities. Strategic objectives such as customer orientation, promotion of individual and organizational learning, and development of core capabilities all needed clear frameworks for being developed. Guiding visions were identified as the tangible outcome of conceptual leadership because they influence and modify cognitive frameworks for how to look at product development through a leadership preceding practice. The objective with guiding visions was to orient development work in different desired situations; in the case study companies, for example, towards developing products that reduce the global cost of a component solution through innovative functional solutions, or developing mini systems as a function of innovative design. The research provided a framework for how to deploy guiding visions based on an extensive communication of strategic objectives and their funding, balanced with a bottom-up return of information concerning operational issues influenced by changing strategies. Moreover, it was found to be essential to support guiding visions with concrete interventions in the organization and to quickly point out successful outcomes of the application of guiding visions. Concerning performance measurements, these were found to play a central role for guiding product development work in the sense identified by guiding visions, and had therefore to be coherent with guiding visions. In fact, if there was a conflict between guiding visions and performance measurements, operational people were found to be more likely to follow the means of guidance that were closest to the evaluation of their performance and expressed in the most familiar language of each individual actor. In relation to existing theory, the research made it possible to establish guiding visions as the tangible outcome of conceptual leadership and the mediating factor between strategic objectives and operational reality. Moreover, if guiding visions are intended to inspire 303 ownership across disciplines and functions they ought to be articulated within the common cognitive ground that exist between these actors. Finally, as guiding visions were found to be developed largely based on customers' expectations, it must be further specified how driving forces from customers are received and who receives them. Relevant feed-back of all kinds of market and technical information must reach the designers of the guiding visions. The development of guiding visions will therefore be dependent on different support structures in the product development context. 12.3.2 The Core Category - Operational Design Four levels of operational design work were identified: individual work, group work, project work, and systemic work. A basic result concerned the conceptualization of operational product development work at the individual level. The product development literature can be classified into two groups, one focusing on the doing in design work, the other on the reflecting side. It was found that rather than being contradictory, these two perspectives complement each other. A model for individual learning was found to integrate them through a description of a cyclical process of conceptual assessment and elaboration of design solutions, and an operational implementation and observation of the same. Compared to the observed reality, this model had a shortcoming, however; it made an artificial distinction between thinking and doing. As operational activities are permanently guided by an inner picture of what the result will be, operational implementation and observation will also be affected by conceptual reflections. This gave a more concrete expression to the phenomenon of tacit knowledge identified in the literature review as a central characteristic of product development work. When starting a new design project, for example, it was the tacit knowledge of a design technician gained by previous experience, and the present means of guidance influencing the way of looking at product development work, that determined what line of action to apply. Concerning group work, a central result was that formal group work in the form of meetings was not always a very productive use of time. One reason for this seemed to be that a followup of different discussed issues did not exist in a broad-banded and interactive sense. Equally important was that the objectives of the meeting rarely went beyond discussing issues principally identified by management. In order to involve employees more fully in such formal group work activities, the discussion needed to be balanced between management and participant-driven issues, and focus on causes and solutions and their degree of satisfaction. Informal group work was found to be a central feature of operative design work and emerged whenever it was needed, but not independently of structural support structures for information dissemination and storage. It was found that an additional reflection-in-practice -for example 304 through co-locating design technicians in a common work space- multiplied the chances of finding relevant solutions to a design problem at hand. The third level concerned project work where the central issue was inter-functional interaction between different professional groups participating in the design work. In relation to the notion of tacit knowledge, the results confirmed that alternating work tasks for example in product and process engineering is the only manner of acquiring complete knowledge in a field. Moreover, this was found to improve commitment to inter-functional considerations and achievement of a balance of status between different professions (something that later was found to be important for developing double-loop learning). The reduction of problems related to different understandings, false priorities and lack of interfunctional understanding that task enlargement and job rotation were found to lead to, implied that it could be interesting also at the technician and engineering level in spite of some objections in terms of the need for experts rather than generalists. Concerning the role of project managers it was found that since project managers were also operating design engineers in the limited sized structures that were studied, this could have advantages in spite of the stress that it caused. Through the integrated problem-solving that frequently occurred between project managers and one or several of their attached design technicians, a good balance between technical, economic, and marketing parameters was reached. The fact that project managers shared operational design increased their credibility in the product development organization. Finally, few problems with inter-functional working were met in the case studies. In fact, the project organization was found to correspond very well to the natural working practice of spontaneous interaction in development work. The central question was not how to go about working together, but how to adapt the organization to the natural inter-functional work and how to promote it. Guiding visions played an important role in this context. If management did not explicitly adopt a cross-functional focus and deploy coherent performance measurements, cross functionality could easily be hindered. The final level in the model concerned systemic work. Here, two main results emerged from the research. The first was the establishment of the importance of allocating individuals to specific innovative development projects. Such individuals were important for pulling design capabilities forward, and for integrating technologies applied in different product families. This happened through an interlocking cycle between drawing on existing design knowledge, developing innovative solutions, and integrating these into the basic design studies. This emphasized the systemic interdependence between these activities. 305 The second result concerned integration with customers. It was found that a permanent information exchange between customers and suppliers was present through relationships at two levels; the project level, and the individual level. In the first case, the internal diffusion of the exchanged information was a function of the internal project management structure in place in the supplier company. If the interfaces that a piece of information had to pass through were too numerous, there was a risk of distortion of the initial message leading to a loss of implicit elements in the original messages and a problem of not returning completely relevant information. Frequent exchange of very broad-banded information at the individual design technician level was found to reduce these problems. In the framework of the set-based approach in black box engineering such direct customer contacts were found to be indispensable for succeeding in the process of mutual adaptation and evolution of interfacing components. Systemic work was found to be both a driver and a facilitator for coordination and integration. Moreover, it was through systemic work integrating customers, design technicians, project managers, and dedicated design engineers that core capabilities took shape, product technology evolved , and inter- and intra-firm learning were connected. The negative side of systemic work -too many and too dispersed activities- had to be recognized and managed through adequate support structures and new forms of organization of the product development process. 12.3.3 Context - Design Support Structures Product specifications, computer-aided systems for information transmission, and technology scanning emerged as three important support structures pertaining to the context in which operational design was carried out. Concerning specifications, the research provided some interesting insights in the case of black box engineering. It was found that specifying was to be seen as a process taking place at the systemic level of operational development work - it needed input from customers, from all project participants inside the supplier firm, and from experts such as consulting engineers. In this way it became a vehicle for integration and creativity in which design technicians had to play a much more active role than traditionally. An important result leading to the proposition of specific action strategies was that it was often difficult for design technicians to determine when a specific design project pertained to the category of black box parts. This indicated that suppliers must become more proactive in their relationships to customers and actively search for the information that they needed. The research data also showed that the more active a 306 supplier was in looking for customer information, the greater the chance of profiling the company as an expert supplier. A particularly interesting result was that different parallel solutions developed in each black box project were continuously presented to and assessed together with the customer. More than focusing on dimensional definitions and an early freezing of specifications, this process enabled the customer and supplier to test out different alternatives and, thus, promote product innovation. A similar line of action has been conceptualized in the literature as the set-based approach. In the case study companies, tendencies of working in this way were observed where suppliers had personalized contacts at the operational design level in the framework of direct or indirect expert supply. These findings modify the generally accepted idea that design changes must be minimized at any price. As long as no dimensional definitions have been made, modifications of functional parameters are not too disturbing. The price to pay for the potential of improving innovativeness through a set-based approach is, however, extensive prototyping in order to test out different solutions. The analysis of the set-based approach -through an integration of the research findings and the existing literature- actually led to the proposition of dividing the design process into two parts; the functional concept and the dimensional definition. If dimensions are defined only after intensive joint problem-solving with customers and interfacing suppliers, lead time and redundant work could be reduced. The central result emerging from these observations was a contradictory practice compared to the general perception of 'lean' product development. Suppliers are indeed involved early in the development process, but hard specifications tend to be fixed as late as possible to promote innovative solutions. Thus, instead of talking about early design involvement, the case studies suggest a permanent design involvement and a late component procurement. The core problem in simultaneous engineering was found to be the handling of the related information flow. Computer-aided systems for information transfer therefore emerged as a second support structure of particular interest. A clear trend in both of the case study companies was to focus on the development of such systems. After the use of information systems for technical data (CAD), the possibility of transmitting qualitative information and enabling interactivity were central preoccupations both at the management and operational levels. An important finding was that such support systems would facilitate the creation and functioning of micro-organizations that the natural ground line in product development was very often found to require. However, before undertaking development of new information systems, the internal project structure and the configuration of the customer-supplier interfaces of a company must be assessed and maybe modified. An information system must not only be legitimized in an 307 organization, but the latter must also be able to make use of the system. The development of information systems could also be an opportunity for integrating dominant and non-dominant disciplines or different functions such as design and marketing in order to support the development of common cognitive grounds. Concerning the third identified design support structure, i.e. technology scanning, the results indicated that this activity should be extended to benchmarking with users of new process technology both within and outside the car industry, that particular attention must be paid to the dissemination of obtained information, and that the productivity of scanning activities increased if those responsible for scanning also made an effort to bring operational R&D staff closer to the source of the technical information. The dissemination of obtained information was found to be facilitated and made more reliable through the development of an internal interactive information system allowing for a real-time reaction to information and making it possible to store information, related analyses, and decisions in a convenient format. It was finally concluded that a common characteristic of the identified design support structures was that of an organizational memory. Holding acquisitions in terms of design capabilities available to all potential users, and making it possible to recall and reuse them were central objectives in all of the three analysed support structures - specifications, computer-aided information transmission systems, and technology scanning. The main problem with existing support structures was that they lacked appropriate dissemination functions thus hindering the active participation of all the concerned actors, for example relative to a specific product design problem in a particular customer relationship, and that they were too fragmented. True support structures for learning should be transversal and active. In summary, research question four concerning how in practice integrated component development is realized can be answered in the following way: • Integrated component development starts with the customer relationship -direct expert supply or triangulation- and the guiding visions that are developed at the management level in order to respond to customer needs and develop the supplier firm. • A comprehensive analysis of operational product development work following the four level model identified several key points for integrated component development. The understanding of the individual and collective intra-functional reflecting-in-practice, that was found to characterize product development work, demonstrated the need for direct operational contacts with design colleagues in customer firms and in interfacing supplier firms. • Project management is one step towards reducing the number of interfaces that a piece of design information has to pass before reaching its final destination, but ultimately it seemed like true integration could be ensured only if frequent exchange of very broad-band 308 information exchange (open communication channels) between suppliers and customers at the individual or collective design technician level was made possible. In such a context the problems of participation, satisfaction, and transfer of experience identified in more formal project or group structures were eliminated. • Management of the specification process was identified as crucial for providing efficient support for integrated component development. The role of specifications change in the integrated context; suppliers working in black box project must learn to work with several parallel sets of possible solutions. • Information technology, finally, was found to play an important role for realizing integrated component development. If exchange of technical data between customers and suppliers could be supplemented with groupware type software, design engineers and technicians thought that innovativeness and development lead-times could be largely improved. 12.3.4 Intervening Condition - Learning in Product Development Learning emerged as a transversal issue in all of the different levels in operational design. Three distinctive learning situations were identified through the case studies: intra-company intrafunctional learning corresponding to the individual and group levels; intra-company interfunctional learning corresponding to the project level; and inter-company intra-functional learning corresponding to the systemic level. The central role of learning as an integrative process in product development emerged through the case studies and not through the literature analysis. Therefore, an additional brief literature review was necessary in order to develop a framework for analysing the learning issue. The literature analysis allowed for a clarification of individual and organizational learning, the latter identified by interviewed managers as an explicit strategic goal. Central concepts for approaching learning were mental models, or theories-in-use, and single and double-loop learning. The former were identified as deeply held internal images of how the world works which have a powerful impact on what people do because they also affect what people see. Mental models are individual but also collective if shared images exist within an organization. Double-loop learning modifies mental models while single-loop learning only corrects errors within an existing system of rules. Individual learning becomes organizational when changing individual routines influence organizational ones and when new individual cognitive insights influence shared images of how to look at the world (or more concretely, of how to look at product development). The central results from the research, answering research question five, were the following: • Individual learning takes place through a questioning of the outcomes of design activities and an 'if-what' reflection on possible solutions. The incentives for learning decrease with 309 the facility and rapidity with which a satisfactory solution is found - the problem being that this facility is hazardous depending on singular circumstances. The assessment of design solutions conducted by project managers and customers should therefore not only focus on the solution itself - the what, but on the process of reaching the solution - the how. • The breeding ground for double-loop learning was found to be the process through which an individual realizes a particular design solution. If the individual applies alternative methods outside the institutionalized routines of the organization and if the internal and environmental feed-back is positive, this will lead to a change in individual, and through spreading of the example, collective theories-in-use. • Besides personal initiatives related to the creativity of individuals, two factors were identified as inviting design engineers and technicians to try out alternative lines of action. The first was 'means of guidance' encouraging initiatives in a specific area, the second was direct inputs from customers. This result emphasized the central role of 'means of guidance' for the outcome of operational design, and the importance of integrating operational design staff with customers. • Joint problem solving between development project participants belonging to different functions was found to be an important vehicle for learning. An example showed that through the co-location of design technicians and process engineering technicians, the former developed a different mental model of how to look at design tasks. An individual double-loop learning occurred that ultimately could be transmitted to the organizational level through intra-functional learning processes between design technicians. In order for this inter-functional learning and its consequences to take place, there was a need for explicit managerial intervention such as spatial reorganisation and/or creation of project groups. • A specific communication mode between design technicians in supplier and customer firms, labelled 'permanently open communication channels', was identified as a facilitator for inter-company intra-functional learning. This high frequency, non-project-tied, nonsequence-tied information exchange was found to be a vehicle for transferring experiences and it made material the links between different parallel development projects. This concerned both different product lines and different phases of evolution in projects. • The analysis of the learning processes made it possible to better understand the set-based approach discussed previously. Permanently open communication channels were in fact found to be a condition for developing this approach. The intensive information exchange taking place in the first step (the determination of the functional concept) was identified as taking place through permanently open communication channels allowing for the exploration of a larger number of solutions than traditionally was the case in the studied companies. 310 The different results related to learning show that the product development process and its outcomes will be influenced by what learning takes place -single-loop or double-loop- and at what level it takes place; inter- or intra company, and inter- or intra-functional. This was the reason for which learning was identified as an intervening condition to the core category. On a more detailed level, learning was in its turn found to be dependent on means of guidance and design support structures. Guiding visions provide an external input to learning, a complementary approach to the reflection-in-practice, namely that of an influence on and a modification of mental models that precedes practice. The objective with guiding visions was in fact to develop consistent shared theories-in-use in a company. The successful realization of guiding visions were dependent on a common cognitive ground that in its turn depended on development inter-functional learning. Finally, some results of general interest for managing learning processes were obtained from the research. They can be classified into three groups: • The transfer mechanisms between individual and organizational learning. Direct communication between individuals, collective problem solving, and individuals altering functions were three important factors that could be encouraged through a strategic intent at the management level of seeing all communication links and joint activities as learning opportunities. • Factors making the difference between single and double -loop learning. Individual initiatives were found to be important for developing double-loop learning. Different means of guidance should therefore encourage employees to challenge the status quo. An organization's attitude towards errors, and the environmental (both internal and external) reaction to proposed modifications were other important issues. • Blockages to learning. The absence of keeping track of work procedures, a focus on the 'what' while ignoring the 'how' during assessment of product solutions, a long lead-time between making a proposition or asking a question and the feed-back, and a distortion of messages due to numerous interfaces were four important identified factors blocking learning. These results could immediately be translated into action strategies for improving the learning processes in product development. 12.3.5 Consequence - Core Capabilities Core capabilities were found to play an important role in the underpinning logic of the restructuring of the automotive supply chain. The literature review positioned a strategy of focusing on core capabilities as a driving force for outsourcing on the part of carmakers and 311 thereby for the development of product development activities in supplier firms. At the same time, as the very concept of a core capability strategy is a driving force for changing product development organization and strategy, the building of specific development capabilities in individual supplier firms become a consequence of the product development process and the way it is managed. They also constitute an important competitive advantage. How capabilities emerge was a central preoccupation for managers in the case study companies. Often, they only realized a posteriori that they had developed a specific capability, for example, as the result of an achievement in terms of a new product solution. Once identified it would be possible to nourish and further develop it, but the problem was that they would have little understanding of how to consciously build core capabilities. The research therefore focused on the process through which capabilities emerge. The central result was the development of a model for the emergence of core capabilities. This model integrated the previously identified and analysed components in the model of the product development process. At the basis of the capability building process are basic resources common to all companies in a specific sector. Means of guidance then intervene as a decisive factor making strategic issues explicit and orienting the development of related work processes in terms of combined resources of a skill, management system, or physical system type. Among these resources are found different support structures and the ones already in place will influence the emerging capability. Finally, a dynamic collective learning process was identified as a necessary condition for constantly refining, developing, and modifying existing skills, management systems, and physical systems. This leads to an organizational evolution that influences the emerging capability through a systemic interaction of basic resources, means of guidance, existing values, combined resources, and learning processes. Technological development external to a specific company will also influence the development of capabilities. For example, new achievements in process technology can make it possible to realize more complex designs and thereby propose innovative components, or allow for more efficient communication and thereby more rapid learning leading to the emergence of a specific capability. Finally, if an emerging capability matches customer needs and differentiates the supplier from its competitors it can be qualified as a core capability. A new emerging core capability will then modify or even create new basic resources, strategic objectives, and system skills and values. Moreover a feed-back influencing technological development, collective learning, and organizational evolution will intervene. An integration of the findings on learning and core capabilities indicates that learning should not only be problem centred (the intra-functional intra-company learning was to a large extent 312 problem centred) but oriented towards the creation of new theories-in-use and capabilities. In managing learning this should in fact be stated as an objective, an end, where the learning is a means - a vehicle for creating something new and/or different. Thus, learning should not be managed only as a continuous improvement process (a kaizen process) but as a process leading to more fundamental change and development of new solutions (a re-engineering process). 12.3.6 An Integrated Model for the Product Development Process Figure 25 presents the integrated model developed from the research with a summary of the most significant results and the related action strategies for managing the phenomena in question. The qualitative research undertaken in this project results in this model that can be seen as a source of hypotheses of what could be a 'high performance' organization in automotive expert supplier firms. The model emphasizes the links between the different phenomena that have been studied. The results of the present research demonstrate that the nature, content, and inherent coherency of guiding visions and performance measurements influence both individual and collective problem solving. Learning theories provide a powerful framework for understanding the links between means of guidance and the actual output of the design process in terms of design capabilities. Central in this framework is the concept of mental models or theories-in-use that are inherent in each individual and influence both the doing and reflecting side of design work making the latter a reflection-in-practice. Even though mental models are inherent in each individual, they are continuously adjusted and modified through collective implementation, observation, assessment, and planning at the group, project, and systemic levels of operational design work. It is therefore possible to recognize the existence of collective mental models. The problem solving and learning processes are not independent of support structures in terms of physical systems, management systems, and a shared knowledge base retained in different types of memory devices playing the role of a context in which learning processes intervene and design work is carried out. Of specific interest in the context of support structures was the need for creating ad hoc organizations, i.e. temporary micro-organizations for resolving a specific design problem. A need for such support structures was identified in several situations, for example in technology scanning, in informal group work in operational design, and in inter-functional and inter-firm learning. In all cases there was a need for informing, exchanging with, and deciding with those actors actually concerned that could be other functional participants (in marketing, process 313 engineering, manufacturing, research...) or people working in customer companies or interfacing supplier companies. 314 CAUSAL CONDITION - MEANS OF GUIDANCE Contribution from the Present Research Action/Interaction Strategies Establishment of guiding visions as the tangible outcome of conceptual leadership and the mediating factor between strategic objectives and operational reality. Communicate strategic goals and their reason for being, combined with a cyclical process of top-down and bottomup information transmission and feed-back. A framework for deploying guiding visions. Support guiding visions through concrete interventions in the organization. Close alignment of guiding visions and performance indicators. Contribution from the Present Research Integration of the doing and reflecting side of development work with a recognition of the tacit dimension. Assessment of the productivity of formal group work. Systemic work Facilitate natural integration through organizational arrangements. Project work Informal group work emerges whenever needed corresponding to a natural ground line of integration in operational work. Project manager's credibility increases if they are also operative design engineers. Design problem-solving can be facilitated and innovative solutions found if design technicians work together internally with engineers responsible for innovative projects and with their customer colleagues. Action/Interaction Strategies CORE CATEGORY OPERATIONAL DESIGN Group work Promote task enlargement and job rotation. Individual work Identify people susceptible to exercising technological leadership and integrating their long and medium-term activities with day-to-day developments. intra-functional inter-functional involving customers, interfacing suppliers and lower tier suppliers Facilitate direct communication between technicians in supplier and customer firms. CONTEXT - DESIGN SUPPORT STRUCTURES Contribution from the Present Research Action/Interaction Strategies Specifying is a process taking place at the systemic level of design work where design technicians need to take an active part. Become more pro-active in the relationships to customers and actively search for necessary information in design projects. Black box engineering can be seen as a two-step process based on permanent design involvement and late component procurement. Support permanent design involvement through a distinction (through guiding visions) between the functional concept and the dimensional definition. Computer-based systems for information transfer allowing for interactivity would facilitate the creation of temporary micro- organizations for problem solving. A basic role for support structures should be that of an organizational memory. INTERVENING CONDITION - LEARNING IN PRODUCT DESIGN Contribution from the Present Research Action/Interaction Strategies The incentives for learning decrease with the facility of obtaining satisfying solutions - problems occur if this facility is hazardous. The assessment of design solutions should focus also on the 'how' besides the 'what' in order to encourage learning in problem solving. Double-loop learning depends on individual initiatives to change and positive environmental feed-back to the change. Equally important was joint inter-functional problem solving. Develop means of guidance encouraging initiatives, arranging for greater customer input to reach design technicians, and structures for inter-functional work. A high frequency, non-project tied, non-sequence-tied information exchange -labelled open communication channels- revealed the links between parallel projects and was a vehicle for transferring experiences. Develop open communication channels in order to develop a set-based approach to product development. 315 CONSEQUENCE - CORE CAPABILITIES Contribution from the Present Research The process of core capability development integrates a large number of factors in the product development process. Basic resources common to companies in a common sector will be the basic building blocks, means of guidance will determine what priorities will be made, support structures will operate in favour of certain design choices, and learning processes will add a dynamic input to the process. Action/Interaction Strategies When understanding that core capabilities are the consequences of the structure, organization, and management of the product development process and the benchmarking of their companies' offerings to customer needs, managers will master the creation and renewal, and nourishing of their companies' product development capabilities. Figure 25. An integrated model for the product development process. 316 The support structures that design engineers and technicians found to be the most relevant, or were asking for as tools for facilitating their work and improve its performance, were based on a possibility of reaching a large number of internal and/or external collaborators and inviting those who had a concern or interest in a question to jointly work on it and discuss it. An important result of the research that was also transversal in the different developed concepts was the need for more direct contacts at the operational technician level between suppliers and customers. The maintenance of open communication channels was found to be a crucial element for speeding up development lead-time, reducing the risk of misunderstandings in information transmission, and therefore promoting organizational learning both within and between companies. This way of communicating that takes place between design engineers and technicians in customer and supplier firms is only one part of the coordination process in parts development and procurement. To fully take advantage of the closer relationships between suppliers and customers, the integration mustn't be limited to the design function only. However, achieving a wide operational synergy between the entire production process within the supplier firm and the one in the carmaker firm is a much more difficult task. As it was clearly established that an important amount of the knowledge that a person possesses is tacit, i.e. only learned through action, it becomes clear that for example a design engineer, a process development technician, and a production manager not only have different priorities and ideas of how to best solve a problem, but also different professional languages. Hence, besides focusing on interaction with customers in the design process, general managers must also focus on how to facilitate the transfer of tacit knowledge, i.e. the inter-functional learning during the entire production process. One way of doing this is to develop support structures enabling different participants in the product development process to learn different tasks 'from inside'. Informal communication together with rigour in formalizing also puts special demands on operating design engineers and technicians: they must be capable of developing favourable relationships with colleagues in customer and partner companies, and they must be able to use information technology to document the design work and the innovative solutions that are informally discussed. Through the close and continuous integration with customers and the need of optimizing the global product offer, the design function finds itself in a particular interface situation. From traditionally having been a relatively isolated function, the design department is beginning to play a role of a double interface: that of ensuring the functional product quality towards the customer, and that of ensuring the quality of realisation, i.e. a product easy to manufacture. Managerial focus must therefore also be on the 'new' organization around the customersupplier interface. In product development this means focusing on communication modes, information transfer, inter-company learning and collective memories that go beyond company 317 limits. Findings imply that the term organization should be used with caution. In daily managerial language it is often synonymous with a company. In the context of this research such an equation between company and organization would distort reality. The interface organization -that to a large extent is created spontaneously- must be managed by coordinating several companies' managerial systems (tangible means and procedures as well as intangible mental maps and values). Something that becomes evident from the case studies is the intimate link between organizational learning and core capabilities. In fact, learning emerged as a link between the coordination problem in operational product development (vividly recognized in the literature) and the core capability problem in product development strategy (equally vividly recognized in the literature). 12.4 LIMITATIONS OF THE RESEARCH AND AREAS FOR FUTURE RESEARCH This study has applied a qualitative research perspective from inside organizations. The research questions were of an exploratory and descriptive kind and the objective was to develop a broad understanding of the phenomena under study. The results and propositions issued from this methodology show its strengths but also its weaknesses. The strengths lie in the consideration of the context in which different findings are situated and in the possibility of mapping complex causal relationships between phenomena. Using the paradigm model allowed for a certain structuring of the different concepts that emerged from the study and lead to the proposition of an integrated model for the product development process in automotive expert supplier firms. However, this research ends with a number of proposed relationships and interconnections reflecting the product development process in a small number of companies. The study is therefore exploratory in its nature and the phenomena and interrelations that are proposed need a complementary validation from studies in a larger number of companies. An additional difficulty concerns the reporting of findings in this type of qualitative research. According to the methodology developed by Strauss & Corbin (1990) the empirical chapters are presented more as a 'story' integrating the findings from the three data collection phases in the research (i.e. the initial interviews with General Managers and Product Development Managers, the two case studies, and the four final follow-up interviews with Product Development Managers) than in an analytical way separating different findings. 318 The areas for future research can be divided into two groups. The first concerns a quantitative validation of different findings, above all in the lean product development context. If the proposed typology of customer relationships can be confirmed through a quantitative survey, it would constitute an important development of the existing tier model. The hypothetical findings concerning quality standards as inter-firm governance structures, the extent of application of lean techniques, and the satisfaction with the application of the latter would also be relevant to test quantitatively. The second area for future research would focus on strategies and organization of the product development process and a deepening and refinement of the proposed model. Two areas of particular interest seem to be information systems as support structures, and the testing of the model of the emergence of core capabilities. Concerning information systems, the rapid evolution of technology and the generalization of electronic mail systems and interactive message systems can be supposed to further modify the access, use, and management of information. If, for example, virtual prototyping becomes a reality in day-to-day design, the set-based approach might be quickly generalized. A focus on the horizontal linkages between suppliers should be particularly interesting to develop. Concerning the management of core capabilities, there seems to be a clear need for much better mastery of the development of capabilities. The proposed model might be a helpful tool for doing this, but it needs direct empirical testing in order to be definitely established. Finally, cross company studies in the production chain would provide a complementary perspective on the studied issues. It would be particularly interesting to focus on the horizontal relationships between interfacing suppliers that it was anticipated would become more and more important, but where the 'tier dilemma' momentarily hindered a more open collaboration. 319 POSTSCRIPT The research confirmed that lean techniques were well established in the auto industry which means that the proposed model of the product development process could be used as a benchmark providing interesting guidelines of how to manage the product development process in industries beginning to transfer to lean. Furthermore, the research confirmed that partnership relations do exist in at least two kinds of customer relationship; direct and indirect expert supply. It can be concluded that a further development of partnership relationships not only might reduce the main problems in lean product development, i.e. price pressure and external confidentiality, but also might improve the efficiency and innovativeness in operational design work through a larger amount of direct customer inputs at an operational level. The study has also shown that suppliers of medium size can develop an expertise that make them indispensable to the final customer, at least in terms of direct intelligence links. All automotive supplier firms, including system suppliers, have gone through a period of turbulent changes and tremendous pressure over the last ten years. Nothing indicates that this pressure in terms of quality, cost and lead time will decrease. In this operational reality, medium-sized suppliers must move away from the often negative discourse of the 'poor' supplier and begin to play a more active role of a 'demanding' supplier. For example, suppliers mustn't be afraid of explicitly stating the requests that the company finds it necessary to work in an efficient way and of pointing at gaps between the discourse and behavioural reality of customers. My hope is that the present work can help managers to better face some of the problems that are related to the strategy and organization of lean product development. 320 BIBLIOGRAPHY Abernathy, W.J. & Utterback, J.M. (1978), "Patterns of Industrial Innovation", Technology Review, vol. 80, no. 7, p. 40-47. Abo, T. (ed.), (1994), Hybrid Factory, The Japanese Production System in the United States, New York: Oxford University Press. Adler, P. S. & Cole, R. E. (1993) "Designed for Learning: A Tale of Two Auto Plants", Sloan Management Review, vol. 34, no. 3, p. 85-94. Adler, P. S. & Cole, R. E. (1994), "NUMMI vs. Uddevalla, Rejoinder", Sloan Management Review, vol. 35, no. 2, p. 45-49. Allen ,T.J. (1977), Managing the Flow of Technology, Cambridge, MA.: MIT Press. Paper back ed. 3rd printing 1986. Andersen Consulting, (1994), Worldwide Manufacturing Competitiveness Study, The Second Lean Enterprise Report, Andersen Consulting UK. Anderson, P. & Tushman, M. (1990), "Technological Discontinuities and Dominant Designs: A Cyclical Model of Organization Change", Administrative Science Quarterly, vol. 35, no. 4, p. 604-633. Anthony, R. N. (1965), Planning and Control Systems: A Framework for Analysis, Cambridge, MA.: Harvard University Press. APRODI, (1985), (Association for Industrial Promotion and Development), Les PMI en France aujourd'hui (SMEs in France Today, in French), Fondation Nationale pour l'enseignement de la gestion des entreprises. Argyris, C. & Schön, D.A. (1978), Organizational Learning, Reading, MA: Addison-Wesley. Argyris, C. (1994) "Good Communication that Blocks Learning", Harvard Business Review, vol. 72, no. 4, p. 77-85. Arrègle, J-L. (1995), "Le savoir et l'approche Resource Based : une ressource et une compétence", (Knowledge and the Ressource Based Approach: One Resource and One Competence, in French), Revue Française de Gestion, no 105, p. 84-94. Asanuma, B, (1985a), "The Organisation of Parts Supply in the Japanese Automotive Industry", Japanese Economic Studies, no 15, p. 32-53. Asanuma, B, (1985b), "The contractual Framework for Parts Supply in the Japanese Automotive Industry", Japanese Economic Studies, no 15, p. 54-78. Automobile Management (1996), "Only the Aggressive will thrive", p.26-28, "Tough but fair", p. 30-31, "A million dollar savings program", p.32, Automobile Management, Germany, June 1996. Bamberger, J. (1978), Développement et Croissance des entreprises, (Development and Growth of Companies, in French), Cahiers de recherche, Rennes: Institut de Gestion de Rennes, I.A.E. Barney, J. B. & Griffin, R. W. (1992), The Management of Organizations: Strategy, Structure, Behaviour, Boston, MA: Houghton Mifflin Co. BCG (1991), The Competitive Challenge Facing the European Automotive Components Industry, Brussels: Boston Consulting Group for the European Commission. BCG (1993), The Evolving Competitive Challenge Facing the European Automotive Components Industry, Brussels: Boston Consulting Group for the European Commission. Bensaou, M. & Venkatraman, N. (1995), Vertical Relationships & the Role of Information Technology: An Empirical Study of US & Japanese Supplier Relationships in the Auto Industry, Working Paper 95/20/SM/TM, Fontainebleau: INSEAD. 321 Bergadaa, M. & Nyeck, S, (1992), "Induction et déduction dans la recherche marketing" (Induction and Deduction in Marketing Research, in French), Recherche et Applications en Marketing, vol. 7, no. 3, p. 2344. Berggren, C. (1990), Det nya bilarbetet (The New Automotive Work, in Swedish), Arkiv Avhandlingsserie no 32, Lund: Studentlitteratur. Berggren, C., (1992), The Volvo Experience, Alternatives to Lean Production in the Swedish Auto Industry, New York: ILR Press. Berggren, C. (1994:1), "NUMMI vs. Uddevalla", Sloan Management Review, vol. 35, no. 2, p. 37-45. Berggren, C. (1994:2) Volvo Uddevalla, A Dead Horse or a Car Dealer's Dream? Actes du GERPISA no 9, Evry: Université d'Evry - Val d'Essonne, p. 129-143. Bertodo, R. (1991), "The Role of Suppliers in Implementing a Strategic Vision", Long Range Planning, no 3, p. 40-48. Bettis, R.A, Bradley, S.P. & Hamel, G (1992), "Outsourcing and Industrial Decline", Academy of Management Executive, vol. 6, no. 1, p. 7-22. Birchall, D., Chanaron, J.J. & Söderquist, K. (1996), "Managing Innovation in SMEs: A Comparison of Companies in the UK, France and Portugal", International Journal of Technology Management, vol. 12, no. 3, p. 291-305. Birchall, D., Frost, Y., Robinson, H. & Söderquist, K. (1994), Introducing Team-Based Work Practices, Final report to the European Foundation for the Improvement of Living and Working Conditions, Henleyon-Thames: The Henley Research Centre. Black, J.A. & Boal, K.B (1994), "Strategic Resources: Traits, Configurations and Paths to Sustainable Competitive Advantage", Strategic Management Journal, vol. 15, Special Issue, Summer, p. 131-148. Blanc, G, Dussauge, P. & Quélin, B. (1991), "Stratégies concurrentielles et différentiation" (Competitive Strategy and Differentiation, in French), Gérer et Comprendre, septembre, p. 75-86. Blomgren, H. & Karlsson, B., (1993), Assembly Skills, Process Engineering, and Engineering Design: An Example from the Volvo Uddevalla Plant, Working Paper, Stockholm: Royal Institute of Technology, Institute of Industrial Economics and Management. Borys, B. & Jemison, D.B. (1989), "Hybrid Arrangements as Strategic Alliances: Theoretical Issues in Organisational Combinations", Academy of Management Review, vol. 14, no 2 pp. 234-249. Bouchut, Y. (1990), "Organiser et gérer la production" (Organizing and Managing Production, in French). In: Jacot, J.H. (ed) Du Fordisme au Toyotisme (From Fordism to Toyotism, in French), Etudes et Recherches, Commissariat Général du Plan, no 7-8, Paris: La Documentation Française, p.169-212. Bowen, H.K., Clark, K.B., Holloway, C.A., Wheelwright, S.C (1994), The Perpetual Enterprise Machine, New York: Oxford University Press. Boyer, R. & Freyssenet, M. (1995), "The Emergence of New Industrial Models", Actes du Gerpisa, no. 15, Evry: Université d'Evry - Val d'Essonne, p. 75-142. Brocquet, C. (1995), "L'avenir des équipementiers automobiles" (The Future for the Automotive Suppliers, in French), Etudes Economiques, Special Edition: Equip'Auto 95. Bryman, A. (1988), Quantity and Quality in Social Research, London: Unwin Hyman. Bucciarelli, L. L. (1988), "An Ethnographic Perspective on Engineering Design", Design Studies, vol. 9, no. 3, p.159-167. Camuffo, A. & Volpato, G. (1993), "Making Manufacturing Lean in the Italian Automobile Industry: The Case of Fiat", in The Fiat & VW Trajectories, GERPISA, First International Colloquium June 17-18-19 1993. Paris: Univerité d'Evry Val d'Essonne. Carr, C. (1993), "Global National and Resource-Based Strategies: An Examination of strategic Choice and Performance in the Vehicle Component Industry", Strategic Management Journal, vol. 14, no. 7, p. 551568. Carter, D. (1992), Concurrent Engineering, Reading, MA.: Addison Wesley. 322 Chanaron, J.J, (1991), Innovation technologique et développement économique (Technological Innovation and Economic Development, in French), Grenoble: Université Pierre Mendès-France. Chanaron, J.J., Gorgeu, A., Laigle, L., Lautier, M., Lecler, Y. & Mathieu, R. (1993), "Constructeurs / fournisseurs : Spécificités et dynamique d'évolution des modes relationels" (Assemblers/Suppliers: Specificities and dynamics of the evolution of relationship forms, in French). In Constructeurs/Fournisseurs, GERPISA, First International Colloquium June 17-18-19 1993. Paris: Univerité d'Evry Val d'Essonne. Chanaron, J.J. & Lung, Y. (1995), Economie de l'automobile (Economy of the Automobile, in French), Paris: La Découverte. Chandler, A. (1962), Strategy and Structure, Cambridge, MA: MIT Press. Charron, E., Freyssenet, M., (1994), "L'usine d'Uddevalla dans la trajectoire de Volvo" (The Uddevalla Plant in the Volvo Trajectory), Actes du GERPISA, no. 9, Evry: Université d'Evry - Val d'Essonne, p.11-55. Charue, F. & Midler, C. (1994), "La robotisation des tôleries automobiles : apprentissage organisationnel et maîtrise des technologies nouvelles" (Robitization in Automotive Stamping Plants: Organizational Learning and Mastery of New Technologies, in French), Revue Française de Gestion, no 97, p. 84-91. Clark, K. (1989), "Project Scope and Project Performance: The Effects of Parts Strategy and Supplier Involvement on Product Development", Management Science, vol. 35, no. 10, p. 1247-1263. Clark, K. & Fujimoto, T. (1991), Product Development Performance. Strategy, Organisation and Management in the World Auto Industry, Boston, MA: Harvard Business School Press. Conner, K. (1991), "A Historical Comparison of Resource-Based Theory and Five Schools of Thought within Industrial Organization Economics: Do we have a New Theory of the Firm?", Journal of Management, vol. 17, no 1 , p. 121-154. Contractor, F.J & Lorange, P (1988), "Why Should Firms Cooperate? The Strategy and Economics Basis for Cooperative Ventures" in Contractor, F.J & Lorange, P (eds.) Cooperative Strategies in International Business, Lexington, MA: Lexington Books. CRESAL-CNRS (1996), Le système automobile en région Rhône-Alpes (The Automobile System in the Rhône-Alpes Region, in French), Saint-Etienne: CRESAL-CNRS. Cusumano, M. A. (1985), The Japanese Automobile Industry, Technology and Management at Nissan and Toyota, Cambridge, MA: Harvard University Press. Cusumano, M. A., Takeishi, A. (1991), "Supplier Relations and Management: A Survey of Japanese, Japanese-Transplant and U.S. Auto Plants", Strategic Management Journal, no. 8, p. 563-585. Cyert, R. M., & March, J. G. (1963), A Behavioural Theory of the Firm, New Jersey: Prentice Hall. Danielsson, A. (1983), Företagsekonomi - en översikt (Managerial Economics - An Overview, in Swedish), 3rd ed. Lund: Studentlitteratur. Danielsson, A. (1986), "Ledarskapes villkor" (The Conditions of Leadership), in Danielsson A. (ed.) Samtal om ledarskap ledning och ledare (Discussions of Leadership, Management and Managers, in Swedish), Stockholm: Svenska Dagbladets Förlag, p. 12-43. De Banville, E. & Chanaron, J.J. (1985), Le système automobile français : de la sous-traitance au partenariat ? Eléménts de problématique, (The French Automotive System: from Sub Contracting to Partnership? Elements of a Problematic, in French), CPE Etude, no. 56. De Banville, E. & Chanaron, J.J. (1990), "Les relations d'approvisionnement", (Supply Relationships, in French). In: Jacot, J H (ed) Du Fordisme au Toyotisme (From Fordism to Toyotism, in French), Etudes et Recherches, Commissariat Général du Plan, no 7-8, Paris: La Documentation Française, p.77-123. De Banville, E. & Chanaron, J.J. (1991), Vers un système automobile Européen, (Towards a European Automobile System, in French), Paris: Economica. De Banville, E., Barnerias, J., Deranlot, M., Chanaron, J.J. (1997), La sous traitance automobile en RhôneAlpes, (Automobile Supply and Sub-contracting in the Rhône-Alpes Region, in French), Charbonnièresles-Bains: Région Rhône-Alpes. Denzin, N. (1970), The Research Act, Chicago: Aldine. 323 Deranlot, M. (1993), L'industrie automobile française dans son contexte européen (The French Automobile Industry in its European Context, in French), independant consultant report, Paris: DATAR. Derrouch, B, (1983), La gestion des risques spécifiques à l'entreprise de taille humaine et son influence sur les stratégies financières et de développement, (Management of Specific Risks for the Company of Human Size and its Influence on Financial and Growth Strategies, in French), unpublished doctoral thesis, Grenoble: University of Social Science, dept. "Administration des entreprises". Dertouzos, M.L., Lester, R.K. & Solow, R.M., (1989), Made in America: Regaining the Productive Edge, Cambridge, MA.: MIT Press. Dey, I. (1993), Qualitative Data Analysis. A User-Friendly Guide for Social Scientists, London: Routledge. DTI & SMMT (1994), A Review of The Relationships between Vehicle Manufacturers and Suppliers, February. Duimering, P.R, Safayeni, F. & Purdy, L. (1993), "Integrated Manufacturing: Redesign the Organization before Implementing Flexible Technology", Sloan Management Review, vol. 34, no. 4, p. 47-56. Durand, T. (1995), "Concurrent Engineering and Interfunctional Project Groups", International Journal of Technology Management, vol. 10, no. 1, p. 67-78. Duymedjian, R. (1996), "De la contingence des normes : les effets inattendus de l'ISO 9000 dans une entreprise experte", (Contingency of Standards: Unexpected Effects of ISO 9000 in an Expert Firm, in French), Revue d'économie industrielle, no. 75, p. 95-111. Dyer, J. & Ouchi, W. (1993), "Japanese-Style Partnerships: Giving Companies a Competitive Edge", Sloan Management Review, vol. 35, no. 1, p. 51-63. Easterby-Smith, M., Thorpe, R., Lowe, A., (1991), Management Research, An Introduction, London: Sage Publications. Ellegård, K., (1989) Akrobatik i tides väv. En dokumentation av projekteringen av Volvos bilfabrik i Uddevalla (A Documentation of the Planning for Volvo's Car Plant in Uddevalla), Göteborg: Department of Human and Economic Geography, School of Economics and Legal Science, University of Gothenburg. Ellegård, K., Engström, T., Johansson, B., Nilsson, L., Medbo, L., (1992), Reflektiv produktion, industriell verksamhet i förändring (Reflective Production, Industrial Activity in Change), Göteborg: AB Volvo Media. Ellegård, K., Engström, T., Johansson, B., Johansson, M., Johnsson, D., Medbo, L., (1994), "Reflective Production in the Final Assembly of Motor Vehicles: An Emerging Swedish Challenge", Actes du GERPISA no. 9, Evry: Université d'Evry - Val d'Essonne, p. 109-128. Ellison, D.J, Clark, K.B, Fujimoto, T, Hyun, Y (1995), Product Development Performance in the Auto Industry: 1990s Update, working paper 95-066, Harvard Business School. Evrard, Y., Pras, B., & Roux, E., (1993), Market, études et recherches en marketing (Market, Studies and Research in Marketing, in French), Paris: Nathan. FIEV (1994), Chiffres Clefs 1994 (Key Figures 1994), Paris: The French Vehicle Equipment Industry Federation. Fiol, C M & Lyles, M. A. (1985), "Organisational Learning", Academy of Management Review, vol. 10 no. 4, p 803-813. Ford, D. (1980), "The Development of Buyer-Seller Relationships in Industrial Markets", European Journal of Marketing, no. 5/6, p. 339-353. Forrester, J.W. (1971), World Dynamics, Cambridge, Mass.: Productivity Press. Freyssenet, M. (1993), "La trajectoire de Renault de 1974 à 1989" (The Renault Trajectory from 1974 to 1989, in French), in: The PSA and Renault Trajecories, GERPISA, First International Colloquium June 1718-19 1993. Paris: Univerité d'Evry Val d'Essonne. Freyssenet, M. (1994), "Volvo-Uddevalla, Analyseur du Fordisme et du Toyotisme" (Volvo-Uddevalla, Analysor of Fordism and Toyotism, in French), Actes du GERPISA, no 9, Evry: Université d'Evry - Val d'Essonne, p.161-183. 324 Garrette, B. (1991), Les alliances entre firmes concurrentes : configurations et déterminants du management stratégique (Strategic Alliances Between Competitors: Configuration and Strategic Management Determinants, in French), Unpublished Doctoral Thesis, Jouy-en-Josas: Ecole des Hautes Etudes Commerciales Paris. Garvin, D.A. (1993), "Building a Learning Organization", Harvard Business Review, vol. 71, no. 4, p. 78-91. Gerlach, M. (1987) "Business Alliances and the Strategy of the Japanese Firm". In: Carrol, G & Vogel, D (eds) Organisational Approaches to Strategy, Berkely: University of California, p. 127-143. GERPISA, (1994:1), The Gerpisa letter no 86, Evry: Université d'Evry - Val d'Essonne Gianfaldoni, P. (1994), Innovations de forme organisationnelle dans l'industrie automobile : apprentissage organisationnel d'une firme imitatrice (Organisatiional Innovations in the Automobile Industry: Organisational Learning in an Imitating Firm, in French), Working Paper, GERPISA workshop 4 November, Evry:Université d'Evry - Val d'Essonne. Glaser, E. M, Abelson, H. H. & Garrison, K. N. (1983), Putting Knowledge to Use, San-Francisco, CA: Jossey-Bass Godener, A. (1996), La survenu des seuils organisationnels dans les petites et moyennes entreprises industrielles en croissance (The Arrival of Thresholds in Industrial Growing SMEs), Unpublished Doctoral Thesis, Grenoble: Université Pierre Mendès-France, Grenoble II, Ecole Supérieure des Affaires. Gong, B.H. (1993), "Cooperative Relationships Between Assemblers and Suppliers in the Automobile Industry: A Comparative Study of Japan and Korea", International Economic Discussion Paper, no. 65, Nagoya: School of Economics, Nagoya University. Goodman P. S. & Darr, E.D. (1996), "Exchanging Best Practices Through Computer-Aided Systems", Academy of Management Executive, vol. 10, no. 2, p. 7-19. Göranzon, B. (1993), The Practical Intellect: Computers and Skills, Berlin, Heidelberg: Springer Verlag. Grant, R. (1991), "The Resource Based Theory of Competitive advantage: Implications for Strategy Formulation", California Management Review, vol. 33, no. 3, p. 114-135. Håkansson, H. (1982), International Marketing and Purchasing of Industrial Goods: An Interaction Approach, Chichester: Wiley. Håkansson, H. (1987), Product Development in Networks in Industrial Technological Development: A Network Approach, London: Croom Helm. Hamel, G (1990), Competitive Collaboration: Learning, Power and Dependence in International Strategic Alliances, Unpublished Doctoral Thesis, Michigan: The University of Michigan. Hamel, G. (1991), "Competition for Competence and Inter-Partner Learning within International Strategic Alliances", Strategic Management Journal, Special Issue - Summer, p. 83-103. Hanna, D.P. (1988), Designing Organizations for High Performance, Cambridge, MA, New York: AddisonWesley. Hawkins, P. (1994), "Organisational Learning: Taking Stock and Facing the Challenge" Management Learning, Vol. 25, No 1, p. 71-82. Hayes, R. H. & Abernathy, W (1980), "Managing our Way to Economic Decline", Harvard Business Review, vol. 58, no. 4, p. 67-77. Hayes, R. H., Wheelwright, S & Clark, K. B. (1988), Dynamic Manufacturing, New York: The Free Press. Hedberg, B. (1981), "How Organizations Learn and Unlearn" In: Nyström, P.C. & Starbuch, W.H. (eds.), Handbook of Organizational Design, London: Oxford University Press, p. 3-27. Helper, S. (1991), "How Much has Really Changed Between US Automakers and their Suppliers?", Sloan Management Review, vol. 32, no. 4, p. 15-28. Helper, S. (1994), "Three Steps Forward, Two Steps Back in Automotive Supplier Relations", GERPISA, Second International Meeting, Communications Industrial Organisation, Evry: Université d'Evry - Val d'Essonne. 325 Herold, D.M (1979), "The Effectiveness of Work Groups", in Kerr, S. (ed.), Organizational Behaviour, New York: John Wiley & Sons. Hines, P. (1995) "Network Sourcing: A Hybrid Approach", International Journal of Purchasing and Materials Management, vol. 31, no. 2, p. 18-24. Hirata, H. (1993), Autour du modèle japonais : Automatisation, nouvelles formes d'organisation et de relations de travail (On the Japanese Model: Automation, New Forms of Organisation and Work Relations, in French), Paris: L'Harmattan. Huber, G., (1991), "Organisational Learning: The Contributing Processes and Literature", Organization Science, vol. 2, no. 2, p. 88-115. Huber, G., (1996), "Organizational Learning: A Guide for Executives", International Journal of Technology Management, Special Issue on Unlearning and Learning for Technological Innovation, vol. 11, no. 7/8, p. 821-832. Hughes, D.G. & Chafin, D.C. (1996), "Turning New Product Development into a Continuous Learning Process", Journal of Product Innovation Management, vol. 13, no. 2, p. 89-104. Hyun, Y. (1994), "Buyer-Supplier Relationships in the European Automobile Component Industry", Long Range Planning, no. 2, p. 66-75. Jakobiak, F. (1991), Pratique de la veille technologique (Practicing Technology Scanning, in French), Paris: Les Editions d'Organisation. Jannik, A., (1991), " Humanvetenskapens kardinalproblem - oenigheten om begreppens innebörd", (The Cardinal Problem of Social Science - The Disunity of the Meaning of Concepts, in Swedish), in Cordelias Tystnad: om reflexion och kunskapsteori (Cordelia's Silence: About Thinking and Theory of Knowledge, in Swedish), Stockholm: Carlssons. Johansson, J. & Mattsson, L-G., (1987), "Interorganisational relations in Industrial Systems, A Network Approach Compared With the Transaction Cost Approach", International Studies in Management and Organisation, p.34-48. Juran, J.M. (1981), Product Quality - a Prescription for the West, EOQC Conference, Paris. Karlson, B. (1991) Konstruktionsändringar - fel eller...? (Design Rework, errors or...?, in Swedish), Unpublished Licentiate Dissertation, Stockholm: Royal Institute of Technology, Institute of Industrial Economics and Management. Karlson, B. (1994), Product Design, Towards a New Conceptualization of the Design Process, Unpublished Doctoral Thesis, Stockholm: Royal Institute of Technology, Department of Industrial Economics and Management. Karlsson, C. & Åhlström, P. (1996), "The Difficult Path to Lean Product Development", Journal of Product Innovation Management, vol. 13, no. 4, p. 283-295. Kaulio, M. A. (1996), "Specifiacations as Mediating Objects: On the (Tactical) Use of Specifications in Technical Product Development Work", Proceedings from the 3rd EIASM International Product Development Management Conference, INSEAD, Fontainebleau, Part 2, p. 491-500. Kawamura, T. (1994), "Characteristics of the Japanese Production Model", in Abo, T. (ed.), Hybrid Factory, The Japanese Production System in the United States, New York: Oxford University Press, p. 2657. Kern, H. & Schumann, M. (1986) Das Ende der Arbeitsteilung (The End of the Division of Labour, in German), München: Verlag C.H. Beck. Kim, D.H. (1993), "The Link Between Individual and Organizational Learning", Sloan Management Review, vol. 35, no. 1, p. 37-50. Kirk, J. & Miller, M. (1986), Reliability and Validity in Qualitative Research, Qualitative Research Methods Series 1, London: Sage Publications Klein, J.A. (1991), "A Reexamination of Autonomy in Light of New Manufacturing Practices", Human Relations, vol. 44, p. 21-39. 326 Kogut, B. (1988), "Joint Ventures: Theoretical and Empirical Perspectives", Strategic Management Journal, no 4, p. 319-332. Kogut, B. & Zander, U. (1992), "Knowledge of the Firm, Combinative Capabilities and the Replication of Technology, Organisation Science, vol. 3, no. 3, p 383-397. Kotler, P. (1989), Marketing Management, 6th. edition, Englewood Cliffs, NJ: Prentice Hall. Krafcik, J. (1988), "Triumph for the Lean Production System", Sloan Management Review, vol. 30, no. 1, p. 41-52. Lacity, M.C., Willcocks, L.P. & Feeny, D.F (1995), "IT Outsourcing: Maximize Flexibility and Control", Harvard Business Review, vol. 73, no 3, p. 86-93. Laigle, L, (1994), Les grands étapes de délégation des activités auprès des fournisseurs : de la soustraitance classique au co-développement (The Big Steps in Delegating Activities to Suppliers: From Classical Subcontracting to Joint Development, in French), GERPISA Working Paper 4/11/94. Lamming, R. (1993), Beyond Partnership, Strategies for Innovation and Lean Supply, Hertfordshire: Prentice Hall International. Lecler, Y, (1993), La Référence Japonaise (The Japanese Reference, in French), Limonest: L'interdiciplinaire. Likert, R. (1961), New Patterns of Management, New York: McGraw-Hill. Loubet, J.L. (1993), "Les changements de trajectoire du Groupe PSA" (The Change in Trajectories of the PSA Group, in French), in The PSA and Renault Trajectories, GERPISA, First International Colloquium June 17-18-19 1993. Paris: Univerité d'Evry Val d'Essonne. Mahoney, J. T. & Pandian, J.R. (1992), "The Resource-Based View within the Conversation of Strategic Management", Strategic Management Journal, vol. 13, no. 5, p. 363-380. May, T. (1993), Social Research: Issues, Methods and Processes, Buckingham: Open University Press. Mair, A. (1994), Hondas Global Local Corporation, Basingstoke: MacMillan. McMillan, J. (1990), "Managing Suppliers: Incentive Systems in Japanese and U.S. Industry", California Management Review, Vol. 34 no. 4, p. 38-55. Midler, C. (1993), L'auto qui n'existait pas. Management des projets et transformation d'entreprise, (The Car that didn't Exist. Project Management and Transformation of a Company, in French), Paris: Inter Editions. Miles, M. B. & Huberman, A. M., (1994), Qualitative Data Analyses, An Expanded Sourcebook, Thousand Oaks, CA: Sage Publications. Minzberg, H. (1979), The Structuring of Organisations: A Synthesis of the Research, Englewood Cliffs NJ: Prentice Hall. Moenaert, R.K. & Caeldries, F. (1996), "Architectural Redesign, Interpersonal Communication, and Learning in R&D", Journal of Product Innovation Management, vol. 13, no. 4, p. 296-310. Moisdon, J.C. & Weil, B. (1992), Groupes transversaux et coordination technique dans la conception d'un nouveau véhicule (Transveral Groups and Technical Coordination in the Design of a New Vehicle, in French), Working Paper no. 3, Paris: Ecole des Mines. Monden, Y. (1983), Toyota Production System, Atlanta: Institute of Industrial Engineers. Morgan, G. (1986), Images of Organizations, Bevely-Hills, CA: Sage Publications. Nevis, E.C., DiBella, A.J. & Gould, J.M., (1995), "Understanding Organisations as Learning Systems", Sloan Management Review, vol. 36, no. 2, p. 73-85. Nobeoka, K. & Cusumano, M.A. (1994), Multi-Project Strategy, Design Transfer, and Project Performance: A Survey of Automobile Development Projects in the U.S. and Japan, IMVP working paper May, Massachusetts Institute of Technology. Nonaka, I. (1991), "The Knowledge Creating Company", Harvard Business Review, vol. 69, no. 6, p. 96-104. 327 Nonaka, I. (1994), "A Dynamic Theory of Organizational Knowledge Creation", Organization Science, vol. 5, no. 1, p. 14-37. Nonaka, I. & Takeuchi, H. (1995), The Knowledge Creating Company, New York, NY: Oxford University Press. Odhnoff, J. (1987), "Den vässade kunskapen" (The Sharpened Knowledge, in Swedish), in Odhnoff, J. & von Otter, C. (eds.) Arbetets rationaliteter (The Rationality of Work), Stockholm: Arbetslivscentrum (The Swedish Centre for Working Life), p. 131-142. Odhnoff, J. (1993), On Management of Technical Change, Working Paper, Department of Industrial Economics and Management, Stockholm: Royal Institute of Technology. Pahl, G. & Beitz, W. (1988), Engineering Design: A Systemic Approach, Berlin: Springer Verlag. Polanyi, M., (1966), The Tacit Dimension, New York: Doubleday & Co. Porter, M. E. (1980), Competitive Strategy, Techniques for Analysing Industries and Competitors, New York: The Free Press. Porter, M. E. (1985), Competitive Advantage, New York: The Free Press. Porter, M. E. & Fuller, M. B. (1986), "Coalitions and Global Strategy". In: Porter, M. E. (ed.), Competition in Global Industries, Cambridge, Massachusetts: Harvard University Press, p. 315-344. Prahalad, C. K. & Hamel, G. (1990), "The Core Competence of the Corporation", Harvard Business Review, vol. 68, no. 3, p. 79-91. Pras, B. & Tarondeau, J-C., (1979), "Typologie de la recherche en gestion" (Typology of Management Research, in French), Enseignement et Gestion, Nouveau Serie no. 9, p. 5-11. Quinn, J.B. & Hilmer, F.G. (1994), "Strategic Outsourcing", Sloan Management Review, vol. 35, no. 4, p. 4355. Reuchlin, M., (1990), La Psychologie (Psychology, in French), Paris: PUF. Richardson, J. (1993), "Parallel Sourcing and Supplier Performance in the Japanese Automobile Industry", Strategic Management Journal, Vol. 14, p. 339-350. Rispal, M. (1993), Les modes de création et de fonctionnement d'accords de coopération (AC) transnationaux entre dirigeants de PME-PMI Européens : une analyse qualitative inductive, (Creation and Management of Transnational Cooperative Agreements between Managers of European SME: A Qualitative Inductive Analysis, in French), unpublished doctoral thesis, Bordeaux: University of Bordeaux, Department of Economics and Management. Rothwell, R. (1991), "External Networking and Innovation in Small and Medium Sized Manufacturing Firms in Europe", Technovation, vol. 11, no. 2, p. 93-112. Rothwell, R. & Zegveld, W. (1985), Reindustrialisation and Technology, Harlow: Longman Publishers. Sako, M. (1992), Prices, Quality and Trust: Inter-firm Relations in Britain and Japan, Cambridge: Cambridge University Press. Sako, M., Lamming, R. & Helper, S. (1994), "Supplier Relations in the UK Car Industry: Good News - Bad News", European Journal of Purchasing and Supply Management, vol. 1, no. 4, p. 237-248. Sandberg, Å, (1994), ""Volvoism" at the End of the Road?", Actes du GERPISA, no 9 1994, Evry: Université d'Evry - Val d'Essonne, p. 145-159. Schein, E.H. (1993), "How can Organizations Learn Faster? The Challenge of Entering the Green Room", Sloan Management Review, vol. 35, no. 2, p. 85-92. Schermerhorn, J.R., Hunt, J.G. & Osborn, R.N. (1991), Managing Organizational Behaviour, New York: John Wiley & Sons. Schön, D.A. (1983), The Reflective Practitioner: How Professionals Think in Action, Basic Books. Senge, P. (1990), The Fifth Discipline: The Art and Practice of the Learning Organization, New York: Doubleday. Shigley, J.E. & Mishke, C.R., (1989), Mechanical Engineering Design, 5th ed., New York: McGraw-Hill. 328 Shimizu, K. (1993), "Toyota, Résultat provisoire des études sur le vingt thèmes" (Toyota, Temporary Results from Studies of Twenty Items, in French), in Trajectories of Automobile Firms, GERPISA, First Intrernational Colloquium, June 17-18-19 1993. Paris: Univerité d'Evry Val d'Essonne. Shingo, S. (1981), The New Japanese Production Philosophy, Tokyo: Japan Management Association. Shrivastava, P. (1983), "A Typology of organisational Learning Systems", Journal of Management Studies, vol. 20, no 1, p. 7-28. Silverman, D., (1993), Interpreting Qualitative Data, Methods for Analysing Talk, Text and Interaction, London: Sage Publications. Simon, H. (1969), The Sciences of the Artificial, Cambridge, Mass.: MIT Press. Smith, P. G. & Reinertsen, D. G. (1991), Developing Products in Half the Time, New York: Van Nostrand Reinhold. Smitka, M.J., (1991), Competitive Ties: Subcontracting in the Japanese Automotive Industry, New York: Columbia University Press. SOFRES (1995), Valorisation de l'image qualité des produits et services Français, (Valorisation of the Quality Image of French Products and Services, in French), Summary Report, October, Ministry of Industry. Spekman, R. E. (1988), Strategic Supplier Selection: Understanding Long-Term Buyer Relationships, Business Horizons, no 4, p. 75-81. Spradley, J.P. (1979), The Ethnographic Interview, New York: Rhinehart & Winston. Stalk, G., Evans, P., & Shulman, L.E., (1992), "Competing on Capabilities: The New Rules of Corporate Strategy", Harvard Business Review, vol. 70, no. 2, p. 57-69. Steensma, H.K, (1996), "Acquiring Technological Competencies Through Inter-organizational Collaboration: An Organizational Learning Perspective", Journal of Engineering and Technology Management, vol. 12, no. 4, p. 267-286. Sterne, D. (1992), "Core Competencies: The Key to Corporate Advantage", Multinational Business, no. 3, p. 13-20. Stiles, J. (1995), A Philosophical Justification for a Realist Approach to Strategic Alliance Research, Working Paper, HWP, 9526, Henley-on-Thames: Henley Management College. Strategor, (1993), Stratégie, Structure, Décision, Identité. Politique Générale d'Entreprise (Strategy, Structure, Decision and Identity. General Strategic Management, in French), Paris: Inter Editions Strauss, A., & Corbin, J., (1990), Basics of Qualitative Research, Grounded Theory Procedures and Techniques, Newbury Park CA: Sage Publications Tesch, R. (1991), "Software for Qualitative Researchers: Analysis Needs and Programme Capabilities", in Fielding, N.G. & Lee, R.M. (eds.), Using Computers in Qualitative Research, London: Sage. Thiétart, R. A., Koenig C. (1987), "Programmes aérospatiaux : la stratégie de l'organisation mutuelle" (Aerospace Programmes: The Strategy of Mutual Organisation, in French), Revue française de gestion, no 62, pp. 42-53. Thompson, G. (1991), Markets, Hierarchies and Networks, the Coordination of Social Life, London: Sage. Thurley, K & Wood, S (1983), "Business Strategy and Industrial Relations Strategy", in: Thurley, K. & Wood, S. (eds.), Industrial Relations and Management Strategy, Cambridge: Cambridge University Press. Twiss, B. (1986), Managing Technological Innovation - 3rd edition, London: Pitman Publishing Van Maanen, J., Dabbs, J.M., Faulkner, R.R. (1982), Varieties of Qualitative Research. Beverley Hills, CA: Sage Publications. Van Maanen, J., (1983), Qualitative Methodology, Beverley Hills, CA: Sage Publications. Von Hippel, E. (1988), The Sources of Innovation, New York: Oxford University Press. 329 Ward, A., Liker, J.K., Cristiano, J.J. & Sobek II, D.K (1995), "The Second Toyota Paradox: How Delaying Decisions Can Make Better Cars Faster", Sloan Management Review, vol. 36, no. 3, p. 43-61. Wernerfelt, B (1984), "A Resource-Based View of the Firm", Strategic Management Journal, vol. 5, no. 2, p.4-12. Wheelwright, S. & Clark, K. (1992), Revolutionizing New Product Development, New York: The Free Press. Whitney, D. E. (1988) "Manufacturing by Design", Harvard Business Review, vol. 66, no 4, p. 83-91. Wild, R. (1989), Production and Operations Management, Fourth Edition, London: Cassel Educational LTD. Williamson, O. E. (1975), Markets and Hierarchies, New York: The Free Press. Williamson, O. E. (1981), "The Modern Corporation: Origins, Evolution, Attributes", Journal of Economic Literature, vol. 19, pp. 1537-1568. Williamson, O. E. (1985), The Economic Institutions of Capitalism, New York: The Free Press. * Womack, J. P. Jones, D. T. & Roos, D. (1990), The Machine That Changed the World, New York: Rawson Associates. Womack, J.P. & Jones, D.T. (1994), "From Lean Production to the Lean Enterprise", Harvard Business Review, vol. 72, no. 2, p. 93-105. Yin, R., K., (1989), Case Study Research, Design and Methods, Newbury Park, CA: Sage Publications. * Pages referred to in the text are from the Harper Perennial paper back version from 1991. 330 APPENDICES 331 APPENDIX 1 PROJECT RESUME 332 LA CONCEPTION "AU PLUS JUSTE" CHEZ LES FOURNISSEURS EXPERTS DE L'AUTOMOBILE : L'IMPACT ORGANISATIONNEL ET MANAGERIAL DU DEVELOPPEMENT INTEGRE DES COMPOSANTS ORGANISATION DU PROJET Etudiant en doctorat : Klas SODERQUIST, Ingénieur, Ecole Royale Polytechnique, Stockholm, assistant d'enseignement et de recherche et doctorant à l'Ecole Supérieure de Commerce de Grenoble. Le projet est organisé en collaboration entre le Groupe ESC Grenoble et Henley Management College - Brunel University (Grande-Bretagne). DIRECTEURS DE RECHERCHE • Professeur Jean-Jacques CHANARON, Directeur de Recherche au C.N.R.S., spécialiste de l'industrie automobile et responsable du programme doctoral au Groupe ESC Grenoble ; • Professeur David BIRCHALL, Directeur de la Recherche, Henley Management College. LE CONTEXTE ET LES OBJECTIFS DE LA RECHERCHE L'organisation de la production a connu un développement continuel durant les cent ans d'existence de l'industrie automobile. Un changement général dans la façon d'organiser la production vers le modèle dit "de la production au plus juste" (lean production), fortement inspiré par les pratiques japonaises, a actuellement lieu dans le secteur automobile au niveau mondial. Aux bureaux d'études des constructeurs, le besoin de raccourcir le temps de développement et une prise de conscience de l'impact important en aval (coûts et qualité notamment) des décisions de conception ont suscité une ouverture vers les fournisseurs. Dans les nouvelles relations donneur d'ordres - fournisseurs, ces derniers doivent prendre une plus grande responsabilité pour mieux concevoir et mieux fabriquer des composants de plus en plus complexes. Ceci concerne notamment les fournisseurs de premier rang, voire des fournisseurs de systèmes ou des spécialistes. 333 Quelques exemples de tels changements : • une réduction importante du nombre de fournisseurs de premier rang, • la sélection des fournisseurs basée sur les performances passées (excellence, implication) plutôt que sur des critères de prix uniquement, • des livraisons exigées en juste-à-temps, de synchrone à hebdomadaire, • des structures communes pour l'assurance de la qualité, l'analyse des coûts, l'échange des informations et l'amélioration de la productivité, • une intégration des composants en fonctions ou sous-ensembles, • l'implication des fournisseurs de premier rang au plus tôt dans le processus de conception, • moins de composants commandés avec des cahiers de charges complets, plus d'exigences d'innovation et de conception indépendante, Ces nouvelles méthodes de travail et ces changements organisationnels interagissent dans le nouveau modèle de production qui est en train de se mettre en place chez les constructeurs comme chez leurs fournisseurs. *** La conception des produits nouveaux (modèles de véhicules, composants) et les nouvelles relations avec les fournisseurs, dans le cadre de la production au plus juste, ont été étudiées dans plusieurs ouvrages et rapports récents. Ces travaux contribuent à une compréhension globale des nouveaux concepts de production et indiquent les principaux champs de changement managérial. La perspective large laisse pourtant de la place pour étudier les démarches concrètes et la réorganisation du travail qui interviennent quand les concepts sont introduits dans la pratique. En plus, la problématique est pour la majorité des cas abordée du point de vue des constructeurs, ce qui semble quelque peu réducteur. Le présent projet de recherche vise à étudier le développement de la nouvelle organisation de la production dans une perspective opérationnelle et du côté des fournisseurs. Un intérêt particulier sera accordé aux moyens de déploiement organisationnel des stratégies d'allégement, au développement et à l'élargissement des compétences du personnel, aux processus d'échanges d'information entre le management et les opérationnels et à l'interaction entre de nouvelles méthodes de travail et celles qui existent déjà. L'objectif global est de repérer les implications managériales, à partir de la réalité vécue dans les entreprises, pour développer une organisation capable de répondre aux nouvelles exigences dans la conception des composants. Le choix du secteur automobile comme terrain d'observation s'explique par le fait que dans ce secteur, le développement des pratiques "au plus juste" a atteint un tel niveau que les conséquences sur les fonctions opérationnelles semblent aujourd'hui commencer à se stabiliser. Ceci constitue une opportunité pour l'étude de l'adaptation organisationnelle aux changements identifiés. 334 METHODE ET QUESTIONNEMENTS Centrée sur les caractéristiques opérationnelles et organisationnelles de la conception des composants chez les fournisseurs de taille moyenne de l'industrie automobile, la recherche est organisée autour de deux volets principaux : "la production au plus juste dans l'entreprise" et "les aspects organisationnels du processus de conception" qui en découle. Comme méthode j'envisage des entretiens pilote auprès de responsables d'entreprises (chef d'entreprise, responsable du bureau d'études) et entre 3 et 4 études de cas plus approfondies des fournisseurs de premier rang (de moyenne taille) disposant de leur propre bureau d'études. L'objectif de ma recherche est de pouvoir répondre, à la fin du projet, aux questions globales telles que : 1. Quels sont les moteurs des changements actuels dans la conception de composants et comment sont ces changements vécus par les fournisseurs ? 2. Quelles techniques de la production au plus juste sont utilisées, dans quelle et avec quel degré de satisfaction ? mesure 3. Comment se réalise, en pratique, la conception intégrée des composants ? Comment sont organisés les bureaux d'études et comment s'organise la coopération inter-fonctionnelle et la coopération avec les constructeurs ? 4. Comment sont élaborées les procédures de travail pour soutenir la coordination, la coopération, la communication, l'apprentissage et les échanges d'information entre les différents acteurs du processus de développement de nouveaux composants ? 5. Y a-t-il des traits communs entre firmes qui formerait une structure organisationnelle de "haute performance" dans le contexte de la conception au plus juste? A l'issu de l'étude, un rapport de synthèse sera adressé à toute entreprise ayant participé au projet. 335 APPENDIX 2 INTERVIEW TOPIC GUIDE 336 INTERVIEW TOPIC GUIDE - Component supplier Phase 1 LEAN PRODUCTION IN THE COMPANY. Theme 1 Place and role of the firm in the supply chain. General Manager Product Development Manager 1.1 Who are the main customers? 1.2 Who are the main suppliers? 1.3 What is the tier-structure of the company? (Graphical representation) 1.4 Is the tier structure different in relation to different product groups and/or customers? 1.5 What is the role of the company in relation to different actors in the tier-structure? idem - an innovation - a specialist in product technology - a specialist in process technology - other specific competencies determining the supplier's role? 1.6 What is the strategy in the tier context? idem Industrial organisation: 1.7. Have you perceived a change in your activities during the last five years? What change? 1.9 Have you got increased design responsibility? idem 1.10 Are you giving more responsibility to your own suppliers? in what field and in what sense? idem 1.11 Could you name and describe the core competencies in your company? idem 1.13 What detailed information do you give to customers? idem - Details of process steps? - Quality management? - Cost information? - design activities? 1.14 What detailed information do you give to suppliers? 337 idem Theme 2. Drivers for change. General Manager Product Development Manager 2.1 Open question: What are the main drivers for strategic change in the management of your company that you - have perceived during the last five years? - are perceiving right now? 2.2 Open question: What are the main drivers for change in the way that you manage your department that you - have perceived during the last five years? - are perceiving right now? Theme 3. Change and development tendencies in the product development process. General Manager Product Development Manager 3.1 How has your company answered to the driving forces discussed above? idem 3.2 What has been the most difficult changes? idem 3.3 What did you do to overbridge problems? idem 3.4 To what extent can you be demanding towards your customers? 3.5 In relation to what problems can you / would you like to be most demanding? 3.6 What characterizes suppliers' strategies in product development right now? idem 3.7 What are the main priorities right now? idem 3.8 What actions are you preparing to achieve these goals? idem Theme 4. Use and perception of lean techniques. This is intended as a "benchmark" of lean techniques defined in chapter 3. More in-depth questions concerning some of these items will be asked under theme six and seven. General Manager Product Development Manager 4.1 Have there been a specific strategy in the company for: Design for manufacturing / assembly 338 idem General Manager Product Development Manager 4.2 Time management: idem - Reduction of development lead time? - Reduction of delivery lead time? - How is just-in-time managed? - What is the relation between production and delivery inventory? 4.3 Flexibility idem - In the production tool? - In competencies? 4.4 Quality idem plus - What is your strategy in terms of quality management? - Has a quality project affected the design organisation? (process, way of work) - Under what quality standards are you certified? - Could you describe your quality project? 4.5 Organisational Structure: idem - How many hierarchical levels? Flat organisation? - What kind of interfunctional cooperation? - Project organisation? - Systematic enlargement of personnel's' competencies? how? - System for learning from errors - continuous improvement - kaizen? - Feed-back structure. Could you describe what happens with a proposition from an employee? 4.6 Have you had the feeling that things have happened during the latest years that are not well anchored in the industrial reality? idem 4.7 Have good practice been thrown on to the rubbish-heap unjustly idem (change for change's own sake)? 4.8 Have assemblers seen a collaborative role for themselves in the process of giving suppliers more design responsibility? idem 4.9 Have their attitudes changed? if yes, why from your perspective? idem Phase 2 THE PRODUCT DEVELOPMENT PROCESS Theme 5. Evolution of product technology and product functions General Manager Product Development Manager 5.1 What are your recent product innovations? idem 5.2 What are the main sources of product innovation? idem 5.3 Is your product technology changing, how? idem General Manager Product Development Manager 339 5.4 Are organisational changes as discussed above influencing product technology? how? idem 5.5 When discussing the place and role in the supplier chain, the company's core competencies were identified. idem in product development (PD). What strategy for developing these core competencies and related strategic goals? 5.6 Do you have some basic ideas or priorities that guide the way the idem in PD company has to work internally achieve these goals? 5.7 What are these ideas or priorities? idem in PD 5.8 What kind of performance criteria are you working with globally in the company? idem in PD 5.9 Could you describe the process of defining them? idem in PD 5.10 What are the main performance improvement actions globally? idem but in PD Theme 6. Organisational structure and work processes. General Manager Product Development Manager 6.1 What are the organisational changes (your structure) in product development ? idem - Organisation today? - Organisation before? 6.2 What are the main objectives in the relation with customers concerning the product development process and organisation that you are working towards right now? idem 6.3 What are the main problems that your organisation meets in connection with product development? idem 6.4 Can you identify some sort of linkage/relations between these problems? 6.5 Who is concerned by these links? 340 General Manager Product Development Manager 6.6 How many products do you develop per year? 6.7 Can you provide a typology for new product development projects? - new product family - expansion of product family - minor product modifications - customer adaptations 6.8 Could you describe the product development process for a typical development project? - initiation, - the inputs, - the actors, - the sequences, - the structures, - the functions, - who are the contact persons in customer firms? - who in your development staff have customer contacts? 6.9 Could you specify what departments / functions are attached to the development team and what are not, even if they are directly concerned with product development? 6.10 What characterizes your work with the customer in the following phases? Structure and Actors - Concept generation - Product planning - Product engineering - Process engineering Project Organisation 6.11 What criteria define a project? 6.12 What is the composition of different projects? 6.13 Is there a project leader? - How is he assigned? - What are his tasks and roles? 6.14 Could you describe the sequences in the project? 6.15 What were the intentions when developing a project approach? (why did you do it = what were the expectations?) 6.16 Have you evaluated these tests or the project approach? formally / informally 6.17 What was the outcome? 6.18 Have you had or do you have an approach of "adapting the design to an innovative manufacturing process"? 341 Theme 7. Coordination activities and communication structure. General Manager Product Development Manager 7.1 What is the role of better external integration with customers in product development? 7.2 How is the coordination with your customers organized? idem - What are the main objectives? - What are the structures? - What are the communication channels? - How does this co-operation work? 7.3 What is the role of better internal integration between different functions? idem 7.4 How is internal coordination organized? idem Objectives, structures, communication channels in coordination between - engineering and manufacturing, - engineering and purchasing, - engineering and engineering, - marketing and technical staff. - How does the cooperation work? 7.5 What is the role of better external integration with suppliers in product development? idem 7.6 How is the coordination with your own suppliers organized? idem - What are the main objectives? - What are the structures? - What are the communication channels? - How does this co-operation work? 7.7 Are you satisfied with existing processes for integration at these three different levels? idem 7.8 How do they correspond to goals for better integration? idem 7.9 What do you wish to develop? idem General Manager Product Development Manager 7.10 What are the main problems in integration between manufacturing, design, customer, and supplier personnel in the product development process? - information sharing, - work practices, - collaboration (have there been isolated persons), - interest conflicts, - professional language, - commitment. 7.11 What has worked well? 342 General Manager Product Development Manager 7.12 What variables are important for smooth integration? - Internal, - External. 7.13 Have you participated in programmes for supplier development (assemblers offer assistance to suppliers in developing their capabilities)? 7.14 What were the main objectives with these activities? - ensure quality? - improve organisation? - develop control systems? - introduce work practices? - other? 7.15 Do you have personnel from your company staying for longer periods with customers? - If yes, for how long time and what determines this time? - What kind of personnel is that? - What are their main activities? - With whom do they work? 7.16 Do you have personnel from suppliers staying for longer periods in your company? If yes, for how long time and what determines this time? What kind of personnel is that? - What are their main activities? - With whom do they work? 7.17 How is learning idem between actors planned to take place? within the same function and between different functions 7.18 Do you have a learning strategy? idem 7.19 How do you think learning will take place best? idem 343 Appendix 3 CASE STUDY TOPIC GUIDE 344 Case Study Topic Guide This document is not an interview guide. It is developed for the researchers own use in order to focus observations. Evolution of product technology and product functions What is the relationship between different customer relationships and the evolution of product technology? Does triangulation inspire specific changes in the capacity of the company to improve and innovate in product technology? Organisational structure and work processes What work groups are in place? Observation of work group efficiency What is discussed? - problems, - causes, - solutions, - actions. In what way? - implication / commitment / mode What are the supports? What influences the continuous improvement structure? What is the role and situation of respectively - encouragement to make propositions? - disposability (time) for making analysis? - the way (frequency, delays, content, attitudes) in which feed-back is given on propositions? - access to information of all kinds needed in order to make relevant propositions? What are the reasons for low participation and other failures in the work organisation? Coordination activities and communication structures. Has integration with customers been privileged at the expense of integration between internal departments? For example, do design engineers and technicians have a tendency to communicate and coordinate more easily with colleagues in customer and supplier companies than with internal colleagues in process engineering and purchasing? *** 345 How does information in problem solving transfer from one project to another in the customer interface? between design people internally, and between colleagues in process engineering and purchasing internally? Coordination and communication - External pool What formal tools are used in the coordination with customers and suppliers? What informal structures are working? Communication mode continuum Coordination and communication - Internal pool What is the situation of the project manager in relation to customers and/or product type? - specialized, - specialized but polyvalent, - polyvalent: disposability decides project responsibility. What is the most advantageous in relational and technological perspectives? *** What characterizes interfunctional relations? Related to supplier linking (the building of relationships to other suppliers both in higher and lower tiers) are there conflicts between the perceived operational efficiency of a sub contractor by engineering and formal demands on demerits, managerial capacity etc. defined by purchasing? In the relation product engineering - quality are there specific problems? For example, insufficient prototype testing? What are the causes? -overload? -time on the CAD + prototype manufacturing? -no possibility to extend delays: send incomplete = not tested products? - testing resources (equipment, human resources)? Integrative questions Place and role of the firm in the supply chain - Change and development tendencies in the product development process. It can be assumed that the studied company has a variety of relationships to carmakers, for example: direct expert supply, triangulation and direct arm's length supply. What are the rationales for each of them? Can they be successfully managed jointly? How? Is one of them becoming more important or less important than the others? 346 Drivers for change - Evolution of product technology and product functions Is the tier dilemma, i.e. the competition in direct expert supply between two potential first-tier suppliers that supply related components, present? What is done to manage this phenomenon? Is there a trade-off between this competition in terms of tier position and the need to collaborate in a "harmonized" supplier base? *** Are customers demanding extended functional solutions? If yes, is assembly taken on at a larger extent than before? If yes, what is done in order to avoid that assembly activities jeopardize a focused competence in product design? (extended product continuum). Organisational structure and work processes - Coordination activities and communication structures. Is management clear in defining objectives that guides employees in their day to day activities and in operational decision making (so called guiding visions)? What are the evaluation criteria and performance measurements? Is there coherency between "official" goals (management discourse) and performance measurements and evaluation criteria? Who has the right to formulate problems? What is done in the field of management of priorities? Evolution of product technology and product functions - Organisational structure and work processes - Coordination activities and communication structures. Can core capabilities be identified, how? Are they made explicit by management? If not, how can this be done? If yes, how are they managed, that is developed, strengthened, protected, "nourished"...? What role do they play in successful development projects? Are there signs of permanent design involvement and delayed component procurement? If yes, this indicates that a situation where the partners know each other so well that tacit assumptions about the other's way of work reduces the need for formal coordination and communication. *** Competence building concept: In what sense is competence building a strategy for improving the design output? Is there an emphasis on / what is done in - know-how? 347 - knowledge? - behaviour? The objective with building competence is to achieve progress. Are measurements in place that coordinate managerial objectives, undertaken actions and obtained results? If not, how could this be done? 348 APPENDIX 4 DOCUMENTS USED IN THE CASE STUDIES 349 The following documents were studied in the case studies: Quality manuals • Procedures for the product development process, the treatment of customer orders and the design reviews • Customers' auditing instruction documents • Design reviews • Performance indicator histories • Project planning documents • Blueprints for specific products quoted as examples in the thesis • FMEA documents • Meeting reviews • Internal information notes and memoranda concerning the product development function • 350 APPENDIX 5 PRESENTATION OF THE CASE STUDY COMPANIES and LIST OF THE INTERVIEW COMPANIES 351 PRESENTATION OF THE CASE STUDY COMPANIES Company Regulation (COR) This company belongs to a foreign group that employed, in all, around 3,000 people in 1995. The group has two divisions: automotive and electronic systems. It has manufacturing units in its home country, in France (the case study company), in the United States and in Mexico. The automobile division manufactures: • Bowden cables, gear sticks, and hand breaks. • Temperature regulation systems, thermostats for cooling systems, thermostatic elements. • Servo steering systems. • Door hinges, and seat components. The studied unit manufactures temperature regulation systems and was managed as an independent company except for very general strategic guidelines that they did not have any influence at all on the company besides the one exerted by the general movements in the sector. In particular, the company had to auto finance R&D and investments. Number of Employees in 1995 The studied unit employed 75 people. Recent Evolution of Turnover (in %) A 62% increase between 1990 and 1995. Around 90% of the turnover was realized in the automobile sector. Main Automobile Customers to the Studied Unit PSA, Renault, Ford, Lucas, Magneti Marelli, Filtros Mann. Formation of the company in 1969; the business was 100% automotive thermostats. Share holding transferred to the Adwest Group in 1990, beginning of a diversification activity outside the auto industry. Longterm objective: 30% of turnover. 352 Main Suppliers Suppliers of screw cut parts, technical plastic parts, and electronics. Raw material suppliers of metal and wax. Organizational Chart Organizational Chart General Management Finance and Administration Purchasing and Commercial Marketing and Sales Quality Management Quality Assurance Research R&D and Industrialization Development Manufacturing Industrialization Project Managers and attached Design Technicians per Customer Products Supplied to the Automotive Industry The studied company manufactures thermostats, thermostatic elements and diesel engine heating systems for the automotive sector which represent almost 90% of the company's turnover in 1995. The company also has a smaller part of its turnover with non automotive customers (temperature regulation in sanitary installations and heating). COR's products for temperature regulation of water, oil and air generally use the principle of wax expansion. For the automotive industry four product ranges are manufactured: • Thermostats for engine cooling systems. • Diesel temperature regulators. Installed before the diesel pump, the regulator mixes hot and cold diesel fuel in order to obtain a constant temperature at the inlet of the pump. • Thermostatic elements for air filters. These elements move a shutter in the air filter in order to mix hot air from the engine and cold air from outside. • Fast idle devices for diesel engines. These elements register the cylinder-head temperature and transmit this information to the diesel pump. Quality Certifications 353 Renault/PSA, Ford Q1, ISO 9001. Contractual certifications with the main system supplier customers. Elements of Global Strategy, and Place and Role of the Company in the Supply Chain. SA Regulation was a first tier supplier of components and development intelligence with three carmakers: the PSA Group, Renault, and Ford. It was a second tier supplier of components and an indirect or direct supplier of development intelligence with several system suppliers, for example air filter manufacturers (Técafiltres, FRAM, Filtros Mann...), fuel filter and injection pump manufacturers (Lucas Roto Diesel), and carburettor manufacturers (Magneti Marelli). The group's strategy was to develop the division automobile and electronic systems in parallel. For COR this meant that new markets outside the car industry had consequently been developed since 1993. An important strategic orientation was to continuously make investments promoting internal growth. The automobile division should develop by designing complete systems rather than developing individual products. For COR this meant that the company must move in the frontline of technological development (product, process and know-how), and organizational evolution in automotive supply. For COR the supply of complete systems was translated through a global approach to the temperature control function: design of thermostats and the integration of their plastic or aluminium compartment including interfacing elements such as electronic captors and indicators. COR had an explicit first tier strategy which meant a presence of competition with complementary suppliers in terms of tier position. 354 Company Fastening (COF) Number of Employees in 1995 600 in France, 1600 world wide (5 implanted in Europe outside France and one in the United States). All data below refer to the studied unit. Recent Evolution of Turnover (in %) A 27% increase between 1990 and 1995. Around 70% of the turnover in the auto sector. Main Automobile Customers to the Studied Unit PSA, Renault, one of the Japanese carmakers, Plastic Omnium, Manducher. Launch of the automotive activity in 1936, upswing from 1950. Main Suppliers Raw material suppliers of plastic and metal. Organizational Charts General Organizational Chart General Management Accounting Marketing and Sales Technical Management Quality Management Finance and Administration 355 Information Systems Quality Assurance Personnel Organizational Chart R&D Technical Management R&D Plastic Projects Process Engineering Metal Projects Manufacturing Maintenance Research Project Managers and attached Design Technicians per Customer Products Supplied to the Automotive Industry The company manufactures products for all kinds of technical snap fastenings (metal and plastic) that go into a car. For example, attachment of external protection devices and shock absorbers, cables, pipes and tubes, door interiors, seats, dashboards, power trains, engine components, and so on. Process technologies include cutting, stamping, plastic moulding, and assembly. Quality Certifications Renault/PSA, ISO 9001 in process in 1996. Over ten contractual certifications with system supplier customers. Elements of Global Strategy, and Place and Role of the Company in the Supply Chain. In 1995, three quarters of total automotive turnover was realized in the first tier, while one quarter was realized in the second or third tier. Turnover was steadily decreasing in the first tier, and a fifty-fifty balance was forecast in three years. This reflected the increasing system supply of seating, shock absorbers, and dashboards. With reference to the tier structure, the place of this company was determined by two product categories: products that are assembled on the 'body in white' and products that goes into sub assemblies. The rate of first tier supply beyond 'body in white' components then depended on the degree of system supply for other components practised by each individual customer. 356 COF had not developed a specific strategy in the tier context, but it had an explicit strategy to design and produce all that goes into the profession of snap fastening systems. This basic strategy was very strongly emphasized to stay in the profession and continuously improve innovation capacity, design capacity, quality, manufacturing processes, and related services. There was a clear strategy to conserve the total mastery and the intellectual property of the manufactured products. The company did not want to diversify to be able to supply a larger part of the car, because it was not an assembly company. A core feature of the strategy was the 'expert supplier'. The company was cultivating this image in relation to its customers, and that was the reason for limiting its activities to the core profession. In fact, COF was already considered as one of the 4-5 world-wide experts in its business by car manufacturers in Germany, France, Italy, Sweden and the United States, a place that had to be defended and continuously developed to last. In these countries, the company was associated to the advanced design departments of the main carmakers locally present. List of Interview Companies (the two case study companies included) Supplier Products Interviewees Supplier 1 High precision mechanical parts General Manager, Product Development Manager Supplier 2 High precision screw cut parts General Manager Supplier 3 Temperature regulation systems (COR) General Manager, Product Development Manager Supplier 4 Fastening components (COF) General Manager, Product Development Manager Supplier 5 Technical plastic parts Product Development Manager Supplier 6 Technical plastic parts Product Development Manager Supplier 7 Seat Components Product Development Manager Supplier 8 Technical plastic parts Product Development Engineer 357