Innovation Portfolio Management (IPM) Practitioner Foundations

Innovation Portfolio Management (IPM) Practitioner Foundations Thursday, April 3rd, 2014 PALISADE Regional Risk Conference Crowne Plaza Amsterdam Nieuwezijds Voorburgwal 5 Amsterdam, The Netherlands Managing innovation initiatives requires making structured decisions in environments of uncertainty and complexity. Palisade DecisionTools offers an integrated platform for quantifying risks / opportunities, gaining analytical insights, optimizing strategy, and guiding decisions. Speaker Background Scott Mongeau Analytics Manager Risk Services smongeau@deloitte.nl Experience Education • Deloitte Nederland Manager Analytics • PhD (ABD) Nyenrode • MBA • Nyenrode Erasmus Lecturer, Business Decision Making RSM • SARK7 • MA Financial Management Owner / Principal Consultant Erasmus RSM • Genentech Inc. • Certificate Finance Manager / Financial Analyst / University of California Berkeley Enterprise Architect • Atradius Web Analytics Manager • CFSI CIO • Consulting Programmer +31 68 201 9225 2 • Grad Degree Info Sys Mgmt Royal Melbourne Institute of Tech (RMIT) • MA Communications University of Texas • B Phil Miami University of Ohio Welkom in Amsterdam! Birth of modern capital markets – Dutch East India Co. (VOC) (1602) • • • • • Corporation Globalization Genesis of modern stock exchange Derivatives (futures & options) Perpetuities http://blog.sunan-ampel.ac.id/auliyaridwan/ Instruments to share risk – Corporation as an ‘entity’ – Capital markets as ‘assessors of risk’ – Wisdom of crowds vs. speculation Dutch Tulip mania – First well-recorded market bubble – Lessons in valuation – Lessons in folly and delusion – Markets are not always right, not always efficient! Slide 3 http://en.wikipedia.org/wiki/File:Flora%27s_M alle-wagen_van_Hendrik_Pot_1640.jpg Learning Objectives CONTEXT • Explain IPM foundation in terms of several contributing disciplines 4 PRACTICE • IPM as analytics challenge • Palisade as hands-on analytics tool EXAMPLE • Several practical cases to demonstrate key principles IPM CONTEXT 5 Innovation Management Powerful solutions for innovation management through state-of-the-art approaches integrating people, processes, and technology perspectives. Innovation Portfolio Management Strategy Decision Management Risk Management Investment Portfolio Management 7 Valuation Project Finance Project Portfolio Management Project Portfolio Management COMPETENCIES GOALS • • • • • • • • • 8 Strategy / Pipeline Governance Finance Risk Resourcing / Coordination Prioritize right projects & programs Build contingencies into overall portfolio Maintain response flexibility Focus on efficiencies Valuation Value Market Arbitrage Period 0 1 2 9 Risk Time Discount Factor 1.0 1 1.05  .952 1  .907 1.052 Allocation Equilibrium Cash Present Flow Value  170,000  170,000  100,000  95,238  320,000  290,249 NPV  Total  $25,011 Investment Portfolio Management OPTIMAL combinations of risks in a PORTFOLIO, given… • Market measure of risk (cost of capital) • Expected return (risk appetite) • Instrument volatility relative to ‘the market’ 10 Project Finance Slide 11 Nagji, Bansi and Tuff, Geoff. (May 2012). Managing Your Innovation Portfolio. Harvard Business Review. Strategy 12 Strategy Barriers Suppliers Rivalry Substitutes 13 Buyers Risk Management Some Likelihood Unlikely Likelihood Likely RISKS New Market Entrants Disruptive & Competing Technologies MITIGATION FINANCIAL Competitors Flood Market • Use external financing • Hedging • Pre-negotiated contracts TECHICAL Lack of Industry Coordination Alt Solution More Attractive Low Difficulty Attracting Financing Lack of Available Technology Noticable Impact Subsidies Disappear Competitor with Killer Innovation Serious • Option to abandon • Option to expand • Consider IP acquisition • Partnerships for synergy STRATEGIC • Concerted market, industry, and competitive monitoring • Flexible commercialization Decision Management Slide 15 Innovation Portfolio Management Strategy Decision Management Risk Management Investment Portfolio Management 16 Valuation ANALYTICS Project Finance Project Portfolio Management Analytics 17 Analytics in Context Data-driven solutions to address innovation portfolio and risk management issues. Innovation Portfolio Management Strategy Decision Management Risk Management Investment Portfolio Management 19 Valuation ANALYTICS Project Finance Project Portfolio Management What to do? SOPHISTICATION What are trends? What happened? PREDICTIVE DESCRIPTIVE VALUE 20 PRESCRIPTIVE Analytics as a process Making smarter decisions PRESCRIPTIVE PREDICTIVE 21 DESCRIPTIVE 2013 Scott Mongeau Decision Process 1. Problem Definition – Make profit enacting group & corporate strategy 2. Objectives Clarified – Capture market within X investment threshold 3. Alternatives Outlined – Lower/higher investment – Other projects / combine projects 4. Decompose / Model (Quantification) – Quantification (i.e. ranking / valuation) 5. Sensitivity Analysis (What if?) – What if scenarios? – Simulation / ranges 6. Follow-Up (Repeat) – Changed objectives, alternatives, preferences? Slide 23 Tools: Innovation Decision Process Slide 24 Palisade DecisionTools Suite 25 Practitioner Tools A full suite of tools to perform risk and decision analysis in order to optimize uncertain outcomes. Innovation Decision Process Slide 27 Managing Uncertainty Categorizing uncertainties Analytics Tool-driven Decision Process Analytics Suite: Palisade TOOLKIT… PALISADE DTS • Simulation • • • • • • Sensitivity analysis Optimization Correlation Econometrics Decision Trees Real Options • @Risk • PrecisionTree • NeuralTools • StatTools • Evolver • TopRank • RISKOptimizer EXAMPLE USES • Supply chain optimization: vendor mgmt. • • • • • Slide 29 Market price uncertainty: fuel costs Cost control: service offering efficiency NPV: uncertainty in new initiatives Risk Management: profitability analysis Optimization: floor configuration, services Traditional Valuation Approach Outcome is based on the single value for each defined assumption Volume Price/Mix Point Estimates Discounted Cash Flow Analysis via Hurdle Rate Cost A&P NPV Project Metrics Payback IRR 30 Simulation: Monte-Carlo Analysis • Probability distributions for all major variables • Multiple outcome simulations run (1000’s of X) • Aggregate probabilities and sensitivities emerge Slide 31 Simulation Approach to Valuation Outcome is a range of possible values generated from applying simulation techniques to key assumptions using business developed probabilities Value Ranges Range of Possible Outcomes for NPV, Payback and IRR Monte Carlo Simulation Volume Price/Mix Discounted Cash Flow Analysis Key Drivers Cost A&P Project Metrics Risks & Opportunities Analysis role becomes more value added through increased collaboration and communication with project team on key drivers and risks & opportunities 32 Variability / Volatility Slide 33 Sensitivity Analysis & Optimization • • • • Dynamic NPV analysis Probability distributions for all major variables Multiple outcome simulations run (1000’s of times) Aggregate probabilities and sensitivities emerge Slide 34 Examples 35 Trends in perspective Analytics is a rapidly evolving space. We maintain a focus on bringing new developments to bear to optimize value-creating decisions. Simulation: Scenarios •Investment – Estimated cost – Product development cost •Production – Capital expense – Overhead – Total expenses •Economic conditions – Inflation – Currency exchange – Unemployment Slide 37 •Commodity cost scenarios •Market Simulation –Estimated # Customers –Competitors –Cost per installation •Sales –Sales price –Sale volume Case 1: Integrated Operational Cost / Revenue Analysis • SEE: Mongeau, S. 2010. Cellulosic Bioethanol Plant Simulator: Managing Uncertainty in Complex Business Environments. 2010 Palisade EMEA Conference • Iterative model development working with area experts Slide 38 ViBeS: Virtual Bioethanol-plant Simulator • NPV • Revenues • Expenses (OPEX) Slide 39 • Scale • Transport • PPE (CAPEX) • Depreciation • Financing • Econometrics Case 2: Optimization and Scenario Ranking • Monte Carlo Simulation • Optimization analysis • Scenario ranking Slide 40 Process Optimization 1 Feedstock 2 Pretreatment 3 Enzymes 4 Fermentation 5 Ethanol Variable incremental Production Costs Cellulose costs: ▪ Availability ▪ Growth ▪ Gathering ▪ Transport, etc. Reduces following OH: ▪ Acid (esters)? ▪ Steam explosion? ▪ Hot water flow? Proprietary: Proprietary: ▪ Process/treatment ▪ Process/treatment ▪ Set of enzymes ▪ Set of yeasts ▪ Product? ▪ Product? R&D: Optimization Focus Subject to Monte Carlo sensitivity/ scenario analysis Slide 41 Final costs: ▪ Mixing? ▪ Testing? ▪ Filling? ▪ Transport? Case 3: StatTools – Commodity Price Analysis Huisman, Ronald. Erasmus School of Economics “Measuring price risk in the short run” Huisman, Ronald. (2009) “An Introduction to Models for the Energy Markets” Case 4: @Risk – Market Behavior Simulation • Market competition and consumer behavior simulation – Market size – Usage per customer – Chance of competitor entering market • NPV distribution result • Monte Carlo analysis • Results in distributions concerning market size and potential profits Slide 43 Case 5: Market Size Valuation Slide 44 Market Competition Simulation / Analysis • Estimates Required – – – – – – – – – Product pricing (profits) Expenses (costs) Market size Market growth rate Point of entry # of competitors Possible new entrants Relevant macroeconomic effects Estimate ratio of investment to market capture (using example data ideally) • What is achieved – – – – – Optimization / efficiency Estimates average profitability and riskiness of new products Gives confidence probability of capturing / holding certain market size Projected revenues (NPV projection with confidence levels) Sensitivity analysis (Tornado Graphs) concerning impactful factors effecting NPV – Scenario analysis with optimal scenario profiles Example: Tornado Graph – Profit Sensitivities and Competitive Effects Slide 46 Case 6: Identifying NPV Key Drivers Example: Histogram - Identifying Non-Normal NPV Distribution • Right skew • Large mean and less spread equates to lower risk of returns • Spread around mean: SD of NPV $410 million Case 7: Decision Tree Analysis Slide 49 Option value determined by… Slide 50 Real Option Analysis (ROA) Process Slide 51 Example: Real Option Analysis - Binomial Tree Options • Assumes one of two outcomes occur in each period: upside or downside • Corrects discount rate imprecision of decision tree (equidistant periods) • Values options by forming “twin” portfolio from which outcomes can be discounted • Black-Scholes option pricing formula can be used as check as volatility shifts over time ** “Corporate Finance: Ch. 22 – Real Options”, Brealey, Myers and Allen. P. 606. Slide 52 Example: Biofuel Plant Binomial Tree • Suggests highly structured rational decision paths • Can be re-run as time and volatility (risk) evolves • Embeds management decision making points and values Slide 53 Example: Drug Development Decision Tree * • Incorporates all outcomes of future project stages and outlines management’s decisions in each event • Net present value (NPV) of each possible “end state” is calculated using the standard discounted cash flow (DCF) model • Starting at end and working backwards, management chooses the highest NPV alternative at each decision point • Process clarifies whether or not it makes sense to abandon, retrial or proceed should any of the trials fail * “How to create value with Real Options based innovation Slide management”: http://www.juergendaum.com/news/12_28_2001.htm 54 A. Define Decisions & Probabilities Slide 55 B. Quantify Final Outcomes Slide 56 C. Regress to Most Rational Choice Slide 57 Palisade Precision Tree Implementation Slide 58 Example: Biofuel Plant Tree Analysis 1. 2. 3. 4. 5. Add management decision points, investments required, and probabilities NPV valuation of each node in scenarios (DCF) Work backwards to probabilistic ‘inherent value’ of management option to expand/contract at each step Choose for highest NPV value at each decision point Revise as probabilities, decisions, and values as time progresses Case 8: Integrated Simulation & Decision Making 1 Feedstock, Ethanol & Oil Price Analysis Investment Simulation Revenue w/ Competition Simulation Costing Analysis Slide 60 2 Monte Carlo Simulation ▪ Oil price scenarios ▪ Investment costs ▪ Revenues (Ethanol & Byproduct, Carbon Credit) ▪ Feedstock variable costs ▪ Energy costs, yeast & processing costs ▪ R&D costs ▪ License income ▪ Market competition 3 4 Decision Tree Analysis NPV & σ CEtOH NPV Model Example: R&D Project Optimization Slide 61 Managing Uncertainty Uncertainties Categorized 1. Target process(es) to employ • Associated costs? 2. Product strategy • Associated revenues? 3. Revenue forecasting • Competition, economic factors? 4. Process cost analysis • Productivity variability? 5. IPM planning / decision making • What decisions, made when? Slide 62 Analytics Process Defined 1. NPV analysis – Three processes – Product strategies 2. Volatility simulation – Monte-Carlo simulation 3. Decision Tree Analysis – Use range of NPV end-points – Add volatility (probability) – Add key decision points Integrated ‘Uncertainty Valuation’ Process • Base Framework • Discounted Cash Flow (DCF) analysis via Net Present value (NPV) • Allows for ‘like-to-like’ comparison of variant scenarios • Cost of Capital: hybrid industry/market derivation and aggregate volatility assessment • Variability Analysis • Monte Carlo allows for sensitivity analysis, structural optimization, and quantification of volatility (risk/opportunity) – chiefly concerned with readily quantifiable financial and physical variables • Assists in pinpointing key risks/opportunities and suggests strategies for mitigating, offloading, selling, insuring, hedging, or retaining said risks (with upside exposure) • Decision Tree / Real Options Analysis • Chiefly concerned with classification of gross uncertainties (i.e. large, nebulous scenarios) • Segments financial variables in MC model and allows for structured high-level management conversations at the Decision Tree Level (NPV values connected a tree end-points) • Final value of aggregate opportunity quantified back to regressed present point • Allows for ongoing managerial ‘options based’ decision making (continual maintenance of ‘tree’) Slide 63 Conclusion 64 Innovation Management Powerful solutions for innovation management through state-of-the-art approaches integrating people, processes, and technology perspectives. Innovation Portfolio Management Strategy Decision Management Risk Management Investment Portfolio Management 66 Valuation ANALYTICS Project Finance Project Portfolio Management Tool-driven Decision Process Slide 67 Managing Uncertainty Uncertainties Categorized 1. Target process(es) to employ • Associated costs? 2. Product strategy • Associated revenues? 3. Revenue forecasting • Competition, economic factors? 4. Process cost analysis • Productivity variability? 5. R&D planning / decision making • What decisions, made when? Analytics Process Defined 1. NPV analysis – Three processes – Product strategies 2. Volatility simulation – Monte-Carlo simulation 3. Real Options Analysis – Use range of NPV end-points – Add volatility (probability) – Add key decision points APPENDIX References 70 References: Decision Management • Blenko, M. W., Mankins, M. C., & Rogers, P. (2010, June). The decision-driven organization. Harvard Business Review, June 2010, p 54 – 62. • Hammond, J. S., Keeney, R. L., and Raiffa, H. (1999). Smart Choices: A Practical guide to Making Better Decisions. Boston: Harvard Business School Press. • An, L. (2011). "Modeling human decisions in coupled human and natural systems: Review of agent-based models." Ecological Modelling. • An, L. (2011). "Modeling human decisions in coupled human and natural systems: Review of agent-based models." Ecological Modelling. • Barney, J. (1999). "How a Firm's Capabilities Affect Boundary Decisions." Sloan Management Review 40(3): 9. • Blenko, M. W., M. C. Mankins, et al. (2010). "The Decision-Driven Organization." Harvard Business Review. • Chouinard, Y., J. Ellison, et al. (2011). "The Big Idea: The Sustainable Economy." Harvard Business review 89(10): 11. • Grote, G. (2009). Management of Uncertainty: Theory and Applications in the Design of Systems and Organizations. London, Springer. • Monch, L., P. Lendermann, et al. (2011). "A survey of challenges in modelling and decisionmaking for discrete event logistics systems." Computers In Industry 62(6): 557-567. • Zook, C. and J. Allen (2011). "The Great Repeatable Business Model." Harvard Business Review 89(10). Slide 71 References: Project Finance • Bodmer, E. (2010, October). Project modeling in excel. Program at Amsterdam Institute of Finance from October 27 – 29, 2010. Amsterdam, Netherlands. • De Servigny, A. and Jobst, N. (2007). The handbook of structured finance. ebook: McGraw-Hill. • Esty, B. C. (2004). Modern project finance: A casebook. Boston: John Wiley & Sons, Inc. • Fabozzi, F. J., Davis, H. A., and Choudhry, M. (2006). Introduction to structured finance. New Jersey: John Wiley & Sons, Inc. • Fabozzi, F. J., Kothari, V. (2008). Introduction to securitization. New Jersey: John Wiley & Sons, Inc. • Finnerty, J. D. (2007). Project financing: Asset-based financial engineering. New Jersey: John Wiley & Sons, Inc. • Gatti, S. (2008). Project finance in theory and practice. London: Elsevier. • HBS Website. HBS project finance portal. Last retrieved March 2011 from http://www.people.hbs.edu/besty/projfinportal/ • Major Projects Association Website. Major projects. Last retrieved March 2011 from www.majorprojects.org • Tan, W. (2007). Principles of project and infrastructure finance. London: Taylor & Francis Group. • Yescombe, E. R. (2002). Principles of project finance. Amsterdam: Academic Press. Slide 72 Additional Examples 73 Model 1: R&D Decision Making – Risk adjusted NPV for uncertain, multi-stage program w/ sensitivity analysis • Method – Set of triangular random variables processed through Monte Carlo simulation • What is achieved – Most likely cost of multi-stage R&D program (NPV) based on range of possible costs, range of possible timelines and associated probabilities with associated confidence level – Regression tornado graph showing relative sensitivities of major factors (i.e.: how NPV effected by standard deviation changes in key variables): thus shows where most fruitful / sensitive value stages are in terms of achieving higher NPV and reducing costs • Data / variables required – Cost (investment) for each stage, time required for each stage, final revenues, WACC; (for each variable: best, worst and most likely scenarios with probability for each) – GANTT project breakdown – Basic understanding of probabilities of success, time, etc. Model 2: R&D Decision Making – Optimal decision making path given range of directions / decisions 1 • Method – Decision tree analysis (real options / derivatives analysis) • What is achieved – Breakdown of optimal NPV based on range of possible decisions – Understanding of most rational (in terms of NPV) decision given choice to proceed or abandon an initiative with uncertain final outcome • Data / variables required – Understanding of key management decisions to be made given range of possible decision paths – Investments (costs) associated, probabilities of success, and profits from each decision Model 3: R&D Decision Making – Optimal decision making path given range of directions / decisions 2 • Method – Binomial tree analysis (real options / derivatives analysis) • What is achieved – Current NPV incorporating value of option to expand or abandon – More structured / detailed breakdown than Decision Tree (yes/no decisions only and equal time spans) – Breakdown of optimal NPV based on range of possible decisions: optimal decision path – Understanding of most rational (in terms of NPV) decision given choice to proceed or abandon an initiative with uncertain final outcome • Data / variables required – Understanding of key decisions to be made given range of possible decision paths – Decision points with yes/no, values, probabilities of success – Investments (costs) associated, probabilities of success, and profits from each decision Model 4: R&D Decision Making – Project portfolio optimization (above and beyond NPV-driven criteria) • Method – Analytic Hierarchy Process and Optimization • What is achieved – Determines relative importance of set of project objectives in a portfolio context – Resource usage (i.e.: cost, man hours) required for each project – Optimal scoring of projects within portfolio based on total benefit and bearing in mind resource constraints – Understanding of how changing input parameters effects total benefit achievable • Data / variables required – Relevant objectives for each project in portfolio – Weighting scores for each objective attached to each project (i.e.: NPV, Market Growth, Likelihood of Technical Success, Likelihood of Regulatory Approval) – Cost, work hours required, NPV Example: R&D Project Optimization Slide 78 Model 5: R&D Decision Making – Modeling New Product Profitability • Method – Triangular random variable, regression analysis, sensitivity analysis, simulation • What is achieved – Estimation of profitability and ‘riskiness’ of new product – Incorporates uncertainties / ranges such as development cost, development timeline, sales life, market size, market share, price, and variable cost • Data / variables required – Ranges for: development cost, development timeline, sales life, market size, market share, price, and variable cost Model 6: Cost Analysis – Resolving Cost of Producing Product • Method – Sampling, regression analysis and optimization • What is achieved – Based on sampled (incomplete, generalized and/or global) component cost information, determine optimized total cost of product production – Given incomplete information on costing, regression analysis alows for targeted product costing with statistical confidence level • Data / variables required – Data on cost components of product – Sample data on cost components (can also be based on similar / related processes) Model 7: Product Pricing – New Product Profitability Simulation • Method – Simulation based on uncertain market parameters • What is achieved – – – – – Estimates average profitability and riskiness of new products Gives confidence probability of holding certain market size Projected revenues NPV projection with confidence levels Sensitivity analysis (Tornado Graphs) concerning most impactful factors effecting NPV – Scenario analysis with optimal scenario profiles • Data / variables required – Number of potential customers – Growth rates for market (with confidence level) – Entry point of competition and variable effect on market share (with probability) When is managerial flexibility highest? Slide 82 Option value determined by… Slide 83 Real Option Analysis Process Slide 84 Contact 85 Contact Details Contact Details Scott Mongeau Advanced Analytics The standard of excellence as a one-stop-shop for full service advanced analytics solutions. Manager Analytics Deloitte Risk Services, Netherlands smongeau@deloitte.nl +31 (0)6 125 802 83 87