Market Segmentation and Positioning in Specialized Industrial Markets

Author(s): Peter Doyle and John Saunders

Source: Journal of Marketing, Vol. 49, No. 2 (Spring, 1985), pp. 24-32
Published by: American Marketing Association
Stable URL: http://www.jstor.org/stable/1251562
Accessed: 27-07-2015 08:18 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.

American Marketing Association is collaborating with JSTOR to digitize, preserve and extend access to Journal of Marketing.

http://www.jstor.org

This content downloaded from 141.6.11.24 on Mon, 27 Jul 2015 08:18:51 UTC
All use subject to JSTOR Terms and Conditions

Peter Doyle & John Saunders

Market

and
Segmentation
in Specialized
Positioning
Industrial Markets

Manycompanies in basic industriesare shifting to speciality products in an effort to boost growth and
profits.But such moves requireoften unfamiliarmarketingskills, especially in segmentation and positioning.The marketliteratureis surprisinglydevoid of practicalexamples of segmentation and positioning techniquesappliedto industrialmarkets.This articlepresents such a model and illustratesit in detail
for a company diversifyinginto specialitychemicals. Implicationsfor other industrialcompanies are discussed.

COMPANIES historicallyconcentratedin basic

commodity industries face problems when they

try to switch into higher growth, higher margin industrialproducts. The first section of this article discusses why such companies often lack the marketing
expertiseto manage the transitioneffectively. A simple but practicalmodel is presentedwhich can guide
the development of a marketingstrategy for a company shifting from commodities to more complex,
specialized industrialmarkets. A detailed case study
is given, showing how one basic chemical company
appliedthis model to successfully position itself in an
attractivespeciality chemical market that was radically new to it. This is one of very few real examples
in the literatureof marketsegmentationand positioning techniques applied to an industrial marketing
problem.

Background
In the United States and Western Europe, companies
in basic, old line manufacturing
industriessuch as steel,
forest
auto, metals, chemicals,
products, and mining
are being forced to implement "tough new survival
PeterDoyleis Professor
of Marketing
andJohnSaunders
in
is Lecturer
Bradford
Centre,
Bradford,
Marketing,
University
Management
England.

strategies"(Business Week 1982a, p. 46). Since the

late 1970s, profits in these industrieshave been pressured by overcapacity, fewer orders, and fierce foreign and domesticcompetition.While the focus in the
short-termhas been on retrenchment,in the longerterm
many of these companies see their futurein switching
from basic commodities into faster growing, higher
margin, "value added" speciality products (Business
Week 1982b, Financial Times 1982). But as The
Economist(1982) noted, these large chemical or steel
companies"gearedto churningout multipurpose,bulk
productsare finding that breakinginto specialist markets requiresunfamiliarskills" (p. 99).
In basic industries the key success requirements
are technological skills, scale economics, and productionknow-how (Hall 1980, Hayes and Aberathy
1980). But in the new markets success depends on
identifying and solving particularproblems. Often it
requirescreating a package of productsand services
tailoredto an individual buyer and produced after a
searchinganalysis of customers'problems. Such flexible responsivenessand skills in marketsegmentation
and positioning are rarely present in these old line
businesses.
Besides lack of experience, additionalreasons for
a lack of marketingorientationare often cited. Webster's (1981) survey of chief executive officers confirms that the technology culture of industrialcom-

24 / Journalof Marketing,
Spring1985
This content downloaded from 141.6.11.24 on Mon, 27 Jul 2015 08:18:51 UTC
All use subject to JSTOR Terms and Conditions

Journal of Marketing
Vol. 49 (Spring 1985), 24-32.

panies commonly gives them a productionratherthan

a marketing orientation. Ames (1970) showed that
marketingis more complex in industrialmarketsbecause marketingeffectiveness dependsto a greaterdegree on otherbusiness functions, especially manufacturing,R&D, inventorycontrol,andengineering.These
difficulties become more acute in specialist products
where the often unique problemsthat the customer is
tryingto solve call for greatersupplierflexibility. Organizationalinability to reorient themselves to flexible response to customer needs accounts for the failure of many giant companies to match the success of
the smaller, more marketingoriented firms in these
sectors.
Companiesundertakingthis transitionwill not find
much help in academic research. Many writers have
noted the lack of researchin modeling industrialmarkets, especially on marketsegmentationand positioning (e.g., Choffrayand Lilien 1980, Shapiroand Bonoma 1984, Webster 1978, Wind 1978). While the
literatureis replete with consumer segmentationand
positioning studies, there are only a few empirical
studiesof industrialmarkets(Choffrayand Lilien 1980,
Greenand McMennamin1974). The consequences of
this neglect are not agreedupon. Wind (1978, p. 318)
asserts that consumer segmentation techniques are
"equally applicable to industrial market situations."
Webster(1978, p. 24), on the other hand, reaches the
opposite conclusion, arguing that the complexity of
industrialmarketingcalls for new approachesthat"must
often be differentfrom those found in consumermarketing." Our findings imply a middle course: While
the underlyingconcepts are similar to consumer marketing, the approachrequires some significant modi-

likely to resist new entrantsin these markets, which

are characterizedby many diverse companies, high
fixed costs, high exit barriers, and strong strategic
commitmentson the partof the leading firms. Market
shareobjectives are importantbecause they influence
segmentationand positioningstrategies.For example,
if the business has modest share goals, then a lower
risk niche strategywill be appealing. If, on the other
hand, it is looking to obtain a major share, it needs
to attackthe most competitive segments of the market.
Determine Market Segments
In consumermarketing,a marketsegment is defined
as a groupof consumerswith homogeneousneeds. In
specialized industrialmarkets, however, this definition is often not manageriallyuseful: situationspecific
variablesare usually much more relevantthan general
customercharacteristics.In industrialmarketing,benefits sought depend less on the internalpsychology or
socioeconomic characteristicsof the buyer and more
on the externalend use of the product.The buyerneeds
differentproductsfor different purposes, and in specialized industrial markets, will often buy multiple
specialistproductsfrom differentsuppliers. For these
reasons, it is often more relevantto segment by product benefits ratherthan customer. Such analysis gives
a more effective insight into the structureof the market and the opportunitiesfor repositioning and new
entrants.

A concept of the marketsegmentationand positioning

process can be briefly presented, since the terms will
be familiarto marketers.The model is kept as simple
as possible since its main task is to communicateto
managers,who in industrialcompanies often have no
formal marketingtraining, the work to be done and
the goals of the analysis.The procedureinvolves seven
steps:

Evaluate Attractiveness of Alternative

Segments
Once the segments are identified, managementwill
need to evaluate the opportunitiesthey represent. A
productbased segmentationusually makes it easier to
estimate the size and growth rates of the segments.
More complex is the analysis of competitors' goals,
strategies,and capabilitieswhich will determinetheir
abilityto defend their segment positions. In speciality
markets, it is common to find some segments dominated by more aggressive competitors than others.
Porter(1980, chapter 7) emphasizes segment choice
as the decision of which strategic group to compete
against. Finally, the firm will have to match the operating requirementsof alternativesegments with its
own technical and marketingcapabilities.

Defining Objectives
Managementneeds to define its marketand financial
objectivesover a three- to five-year planninghorizon.
These should be quantitativeand realistic. Technology orientedcompaniesoften overestimatemarketresponsivenessto their new productsand underestimate
the ability of existing competitorsto retaliate.As Porter (1980, pp. 17-23) notes, existing competitorsare

Select Target Market(s)

The heterogeneousrequirementsof speciality markets
mean that success is rarely achieved by undifferentiatedstrategiesor overallcost leadership.Even though
the major multinationals typically operate in these
markets,it is seldom that a competitorhas the capability of achieving broad marketcoverage by product
differentiation.More focused strategiesare their gen-

fication.

A Market Segmentation and

Product Positioning Model

MarketSegmentationand Positioning/ 25
This content downloaded from 141.6.11.24 on Mon, 27 Jul 2015 08:18:51 UTC
All use subject to JSTOR Terms and Conditions

eral characteristic(for some examples see Hall 1980,

Levitt 1980, Woo and Cooper 1982). The numberof
segments that the firm will compete in will depend
upon its overall share goals, the flexibility of its manufacturingbase, and the heterogeneityof the buyer's
requirements.
Develop a Positioning Strategy
After the choice of target marketsegments, management has to decide how it will compete in these segments. Productionoriented managers often underestimate the need to develop competitive advantagesif
they are to profitablygain marketshare. Ideally, too,
this positioning should be sustainableand not easily
matchedby competitors.
Develop the Marketing Mix
Following decisions on targetmarketsand positioning
strategyin these segments comes implementationin
terms of the detailed marketingmix. Consistent decisions on productdesign, technical support, service,
channels, distribution, promotion, and pricing must
be made.
Validating the Strategy
The entrant'ssuccess or failure will depend upon its
segmentationand positioning strategy, and so it is essential to validate the previous analysis and derived
strategicplan. Follow-up researchin the marketis essential to test the reliability of the findings and the
appeal of the marketingplan.

An Application
The applicationof the model is illustratedby a true
and rathertypical case of a companydiversifyinginto
a specialitymarket.1Boliet (sales $50M, pretaxprofit
$1.5M) is a subsidiaryof a large raw materialbased
multinationalfirm headquarteredin southernEurope.
Boliet concentratedon the processing of pine gum
rosins, which were sold as a raw material to manufacturersof paper size and synthetic resins. It undertook very little upgradingof the rosins-it was essentially a raw material supplier. The decision to
integrateforwardinto specialitychemicalsthroughthe
furtherprocessingof the rosins was stimulatedby lack
of growthand the marginpressuresfrom being in the
handsof a limited numberof customers. The decision
requiredthe investmentof $10 million in a new plant
and the acquisitionof know-how througha licensing
agreement with a leading U.S. chemical company,
giving Boliet applicationrights in Western Europe.
The furtherprocessingof the commodityrosinsinto
'While for competitive reasons some of the details on margins and
competitive advantages have been disguised, such changes have been
kept to a minimum.

polymerizedrosin and stabilized rosin esters opened

up a majornew specialized marketfor Boliet. These
specialized chemicals, called "tactifiers," are a key
component bought by manufacturersof adhesives.
Tactifiersappearedan attractivemarketworth around
$60 million in 1981 and growing at over 10%per annum. With per capitaconsumptionunderhalf the U.S.
level, managementbelieved long-term potential was
high. Companiesin the marketincludedEastmanKodak, ICI, Bayer, Arizona, Grunel, and DRT, but two
U.S. companies, Hercules and Exxon, held 32- and
17%respectively. Managementbelieved that while its
gum-basedproductswould not be of as high quality
as some competitorproducts,for most applicationsthey
were adequate, and Boliet's low material cost base
would enable it to significantlyundercutthe prices of
the marketleaders.
The decision to enter this marketwas made with
little marketplanning. The Groupboard accepted the
investmentproposal on the basis of Boliet's forecast
of a 25% marketshare. Management'slack of a marketingorientationwas reflectedin the absence of analysis of targetmarkets,positioning strategies, product
specifications, and customer support requirements.
Soon afterwards,however, events increased concern
about the lack of marketingknow-how and the ambitious natureof the share objective. First, it became
apparentthat the recession meant next year's market
growthwould be limited to only 1%. Second, it was
discoveredthat both Hercules and Exxon were building new plantsin Europe.It was unclearwhetherthese
were to add to capacity or lower production costs
(previouslyboth had importedthe semifinished product from the U.S.).

The Analysis
Developing a detailed marketingplan for Boliet requireda marketingframeworkand more data.
Creating the Framework
Managementscience has had a disappointinglevel of
acceptancein industrybecause managementwill not
use techniques it does not understand(Little 1979).
This has been especially truefor segmentationand positioningtechniqueswhich have had little application
in industrialmarketing(Wind and Cardozo 1974, pp.
160-161). To overcome these barriers,a three-dayintensive briefing session was held for managementto
providethe consultantswith furtherinformationabout
the technology and its application, and for the consultantsto lead the joint developmentof a marketing
frameworkfor approachingthe problems.
Data Collection
Primarydata was then collected on the two key categories of variables of the model-competitors and

26 / Journalof Marketing,Spring1985
This content downloaded from 141.6.11.24 on Mon, 27 Jul 2015 08:18:51 UTC
All use subject to JSTOR Terms and Conditions

customers. Tactifier competitors could be distinguished by raw material base: petroleum, hydrocarbon, polyterpenes,wood, tall oil, and gum rosins. This
affected relative production costs and certain applications.2The specialist nature of the market resulted
in many brands. The 13 leading competitorstogether
produced over 70 products with measurable market
shares. Competitorswere judged to differ in strategic
commitment, strength, technical reputation, geographicalcoverage, service, and resources. But the
marketwas close to the "naturalstructure"observed
by Buzzell (1981), who found that marketsharestend
to follow the semi-log distribution with the largest
competitorsholding market shares of 33-, 19-, 12-,
and 7%.
About 120 adhesive manufacturersoperated in
WesternEurope. These purchasedtactifiers plus wax
andbase polymersand formulatedadhesivesfor a wide
range of end users, including packaging, bookbinding, woodwork, product assembly, and non-woven
goods manufacturers.After exploratory discussions,
semi-structuredinterviewswere held with a sample of
20 large and 20 other buyers of tactifiers. In each organizationthe buyer or majorspecifier agreed upon a
list of variables accounting for choice of brand and
source. The research identified six product specific
variablesaffecting choice: softening point, viscosity,
color stability, startingcolor, tack, and price, and four
companyvariables, the supplier's productrange, service support,geographicalcoverage, and general reputationfor reliability.The productcharacteristicswere
objective, quantifiabledata. The company variables
were semantic differentialsratedjointly by corporate
buyers and specifiers. The researchalso revealed that
formulatorstypically bought from several tactifiers
because none could meet all the applicationrequirements. No simple general customer segmentation
scheme was plausible since formulatorssold to multiple end users, and tactifiershad applicationsin various end uses (Figure 1).

Statistical Analysis
Factor and cluster analyses were employed to determine if the formulators'choice criteriacould be summarized into simpler dimensions and whether the
complex product applications could be grouped into
segments. Principalcomponentsanalysis was used to
determinewhetherthe standardizeddata matrixof the
six productand four companyvariablescould usefully
be reduced into fewer dimensions. Alternativefactor
rotationalmethods were explored but did not appear
to offer significantadvantagesin applicationor interpretability.Table 1 shows the four component solu2Readersinterested in the technical aspects of these materials should
see Othmer (1980).

FIGURE1
Structure of the Tactifier Market
Tactifier
Suppliers

Adhesive
Formulators

End Users

tion chosen for analysis. These accountedfor 86% of

the cumulativevariance.The loadingssuggest the first
factordescribesthe perceived strengthof the supplier,
the second the breadthof application of its product
range, and the third and fourthfactors relate to technical characteristicsdefining the product's use.
Clusteringof productsis not new in segmentation
research (Wind 1978, pp. 67-105). Wind distinguishes four types of segmentationmodels-a priori,
cluster based designs, flexible, and componential. A
priori segmentationwas ruled out because management did not know in advance the numberand types
of segments. Flexible and componentialdesigns, both
of which rely on conjointanalysis, were inappropriate
because of the large numberof productsin the market. Ward's(1963) hierarchicalclusteringroutinewas
then used on the factor scores to search for segments
among the 76 products in the tactifier market. This
clustering method was chosen initially because previous studies (Blashfield 1976, Wishart 1978) have
reportedthat this technique is consistently more accurate than others in recovering data from different
populations.In its most familiarform, Ward'smethod
is a hierarchicalclustering process that at each stage
combines the two clusters whose fusion leads to the
least increasein the errorsum of squares.The process
produces compact spherical clusters with a centroid
that can be used to representthe position of the cluster.
As with all clusteringtechniques, one of the problems was deciding how many clustersto choose. Several authors(e.g., Friedmanand Rubens 1967) have
suggested that the number of clusters should be set
when furthercombining of clusters gives a large increase in error sum of squares. Other more complex
statisticaltests have been proposed(Beale 1969, Wolfe
1971). Unfortunately,these methods are rarely sat-

MarketSegmentation
andPositioning/ 27
This content downloaded from 141.6.11.24 on Mon, 27 Jul 2015 08:18:51 UTC
All use subject to JSTOR Terms and Conditions

TABLE1
Factor Analysis of Purchasing Criteria for Tactifiers
Factors
1
2
3
Softening point
Viscosity
Color stability
Start color
Tack
Price
Range
Service
Coverage
Reputation
Variance explained (%)

-.0702
-.4790
.4494
-.5194
-.2813
.5871
.5748
.8427
.8130
.7759

.2562
.4865
-.4265
.6927
.3571
-.3471
.6222
.4076
.4082
.0949

.8051
-.5305
-.2871
.2711
-.0385
.2799
-.0338
-.1536
-.2249
.2949

-.2069
-.1528
-.0929
-.0949
.8134
.5206
.2142
-.0531
-.1208
-.1962

34.3

19.4

13.3

11.1

isfactory in practice (Saunders 1980). For management, the real criteriais utility, which depends upon
such factors as the value of the segment, their reachability, competitive activities, and management'scapacity to implementdifferentsegmentationstrategies.
The approachadoptedwas to present to management
as clearly as possible several levels within the clusteringhierarchyso thatthe decision makerscould find
the configurationthat fits most closely theirneeds and
capabilities.
The approachled to a 12-clustersolution, the centroids of which are shown in Figure 2. It is worth
noting that in these specialized markets, the number
FIGURE2
of the Tactifier Market
Segmentation
~~~~ ~ ~ ~
,

of clusterswill be large. Management,however, must

evaluate the trade-off between increasing segment
homogeneity and their diminishing dollar value and
rising costs of servicing. The options can be seen in
Figure 3, which shows that clusters A, B, C, and D
can be easily combined without a great deterioration
in the stress factor. The nine-cluster alternativewas
rejected by management, however, because the former clusterwould have too many productsto be strategically valuable. But this decision can be kept under

review.

Statistical Validation
Choffrayand Lilien (1980) suggest a three stage validationprocess:test the sensitivity of the cluster analysis to extremeobservationsor outliers, conduct tests
on the nonrandomnessof the data structure,and determinethe uniqueness of the clustering solution obtained.
To identify the outliers and determinetheireffects
on the cluster analysis, the approach suggested by
Blashfield (1976) was followed. Single linkage cluster analysiswas used to identify the most weakly connected clusters. This analysis grouped two clusters
FIGURE3

Dendogramfrom Twelve-ClusterSolution
CUMULATIVE
BRANDS

CLUSTER SIZE
CLUSTER

10
10
A

9
B

20

30

40

8
C

1-

cn

32-

3-

28 / Journalof Marketing,Spring1985
This content downloaded from 141.6.11.24 on Mon, 27 Jul 2015 08:18:51 UTC
All use subject to JSTOR Terms and Conditions

50

7
D

2
E

12
F

4
G

60

70

76
I

where they each contained a subject that was more

similarthanany otherpair of subjectsfrom otherclusters. The resulting process validated Ward's clustering method; the outliers were identified in segments
K andL. Inspectionconfirmedthatthese were distinct
segments which should be kept separatein the analysis.

A characteristicof cluster analysis is its ability to

form homogeneousgroups even from randomdata. A
simulationapproachcan be used to compare the results with those generatedfrom randomdata. To provide an estimate of the randomclusteringprocess, 10
datasets were generatedrandomlyfrom the same empirical distributionas the observed data. Each of the
sets was then factor analyzed and the scores clustered
by Ward'smethod. Table 2 shows that there is a very
large difference between the actual results and those
expected from randomly generated data. The actual
stress switching from a lower level to a higher than
expected level suggests a naturallevel of clustering
and gives furtherconfirmation to the 12-cluster solution utilized.
The final step examined the sensitivity of the segmentationanalysisto the particularclusterroutineemployed. Given the diversityof similaritymeasuresand
algorithms, there is little likelihood of identical solutions.Nevertheless,resultsappearmorerobustif they
can be generatedby differentmethods. Four common
cluster techniques were compared: Ward's method,
euclidean distance with complete linkage, euclidean

distancewith single linkage, and correlationwith single linkage. These are not described here (for a review, see Everitt 1974, Hartigan1975), but they were
chosen because they tend to produce different types
of clusters and so provide a severe test of the results.
Table 3 comparesthe results and shows that there is
a high concurrencebetween the methods. In particular, Ward's method has the highest concurrencewith
each of the other approachestested.

CompetitorAnalysis
The next stage was to evaluate the attractivenessof
the alternativesegments and assess the strength of
competition.Each segment was first profiled in terms
of applicationand end use. For example, cluster A
was broadly used for heavy adhesives, especially in
the woodworking sector. Then the size, growth,
strengthof competition, and Boliet's relative capability to operatewas estimatedfor each segment (Table 4). For instance, its capabilitywas rated by managementas good in segmentJ becauseit could produce
tactifiers with all the characteristicsof the leading
brandsand make high marginsat prices substantially
below the existing marketprice.
Competitive capabilities are also summarizedin
Table 4. These were formally evaluated by management on the basis of (1) strategiccommitmentto the
industry, (2) ability to retaliate, and (3) difficulty in
displacing. It was importantto push managementto
formallyconsider competitive strategies, since at this

TABLE2
Test for Nonrandomness of Data Structure
Stress
Number of Cluster
70
60
50
40
30
20
15

Actual
0.084
0.167
0.249
0.354
0.524
0.672
0.846

Expecteda
0.497
0.587
0.653
0.701
0.764
0.863
0.945

S.D.b
0.027
0.030
0.033
0.032
0.031
0.029
0.039

t-test
-15.29
-14.00
-12.24
-10.84
-7.74
-6.59
-2.54

14
13
12
11

0.961
0.964
0.970
1.019

0.954
0.967
0.984
0.999

0.038
0.039
0.040
0.040

0.18
-0.08
-0.35
0.50

10
9
8
7
6
5
4
3
2
1

1.162
1.285
1.700
2.152
2.248
2.332
2.558
2.641
2.955
3.131

1.013
1.033
1.052
1.083
1.121
1.193
1.239
1.375
1.763
3.866

0.041
0.040
0.039
0.041
0.038
0.040
0.050
0.052
0.061
0.087

3.63
6.30
16.62
26.07
52.73
28.47
26.38
24.35
19.54
-8.45

aExpectedresults from 10 randomly generated data sets.

bStandarddeviation of expected stress levels.

MarketSegmentation
andPositioning/ 29
This content downloaded from 141.6.11.24 on Mon, 27 Jul 2015 08:18:51 UTC
All use subject to JSTOR Terms and Conditions

..

..

.~

TABLE3
Convergence of Clustering Methods
Ward's
Euclidean
Euclidean
Method
CompleteLinkage
Single Linkage

Correlation
Single Linkage

Ward's Method
Euclidean Complete Linkage

65.8%a
(508.2)b

Euclidean Single Linkage

77.6%
52.6%
(588.8)
(424.2)
CorrelationSingle Linkage
68.4%
60.5%
67.1%
(491.7)
(390.3)
(396.3)
'Percentof consistentclassifications:numberof items that are groupedin the same clusterfor the two twelve-clustersolutions.
test for hypothesesthat pairedresultsare independent.All rejectedat 99.9%confidencelevel.
bChi-squared

stage of marketdevelopmentit is common for leading

competitors to diversify into new segments (Porter
1980, chapter 10). Table 5, which cross-classifies
competitorsagainst segments, confirms that companies currentlytend to pursue highly focused strategies. But comparing strategies was suggestive. For
example, company 5, which was regardedas highly
aggressive, was viewed as a potential competitor in
segmentsE and G because company4, which had the
same raw materialbase and similar policies to 5 in
segmentsI, J, K, and L, had alreadymoved into these
new segments.
The strategic segmentationand evaluation generated many other insights. For instance, it is leading
to the explorationof "gaps"-potential product segments not yet developed. It has also dissuaded management from risk of the "majorityfallacy," since

Features
No. of brands
Softening point
Viscosity
Color stability
Start color
Tack
Price
Range
Service
Coverage
Reputation
Attractiveness:
Size ($M)
Growth (% per
annum)
Competitive
strength
Capabilityfit
(Note: S = strong, M

Figure 2 and Table 4 show that segments away from

the "middle"of the marketwere less competitive and
made possible high marketshare brands.

Segmentationand Positioning
Selecting Target Markets
From the evaluation and further qualitative judgments,managementdeterminedthatseven of the twelve
marketsegments were possibly viable, but of these,
four were particularlyattractive(clustersC, G, J, and
K). These four representedsales of $32.8 million or
54% of the market. Given the importanceof developing a strong customer interface, it was decided to
focus on these four segments for the first 18 months
of the plan, aiming for a 30% share or 19.2% of the
total market.

TABLE 4
Characteristics and Attractiveness of Alternative Segments
Clusters
A
B
C
D
E
F
G
H
I
10
91
8
102
8
2
9.8
16
2
2
2

109
8
126
9
2
9.7
20
2
2
2

96
9
95
7.5
2
6.6
10
2.6
2.6
2

7
151
8
51
11
2
8.0
18
2
2
2

92.5
7
100
8
1
7.7
2
2
3
1

2.4
3

1.8
5

4.2
7

3.2
12

5.0
15

4.0
8

12
122
6
70
8
1
9.1

10

90
6
77
7
1
9.2

97.5
4
90
3
2
14.3
17
1.5
2
2

90
9
200
4
2
9.1

80
14
130
9
1
6.5

7.5
2.5
2.5
2

109
7
92
8
2
12.3
55
4
4
3

36
2.5
3
2

110
4
500
.5
1
14.3
21
4
4
3

10.1
10

0.8
8

4.2
4

10.4
14

8.1
13

7.4
10

10
2
2
3

30
4
4
3

medium, W = weak.)

30 / Journalof Marketing,Spring1985
This content downloaded from 141.6.11.24 on Mon, 27 Jul 2015 08:18:51 UTC
All use subject to JSTOR Terms and Conditions

TABLE5
Company Brands versus Product Segments
A

1
2

10
-

Company

4
5

6
7
8

9
10
11
12
13
Total

Product Segments and Brands

E
F
H
I
G
1

6-

3
4

10

1
-

-1
-

1
2

2
3
-

Total

7
12

1
3

12
10

1
1

3
3

12

1
6

3
4

1
7

5
8

10

3
3
3
3
7
76

(Note: The hypothesis that segments and Company are independent may be rejected with 99.9% confidence. Chi-Square corrected
for small cell size = 198.8, DF = 132.)

Conclusions

Product Positioning
After the choice of segments, a detailed positioning
strategyhad to be formulatedto seek differentialadvantages. In segments G and J, Boliet decided to offer
essentially "me-too"productswith dramaticprice reductionsmade possible throughthe substitutionof gum
for higher cost tactifiers. In C and K, price cutting
was less attractivebecause of low cost alternatives,
and the positioning thrust was in terms of improved
technical characteristicsand support.
Marketing Mix
The final step was the development of a marketing
plan and mix for each product.Productspecifications,
price, promotion, and distributionpolicies were formulated to be consistent with each product's positioning statement. Until then, managementhad given
little thought to the problems of marketing in eight
countries, but now, with a clear direction, progress
and decisions could be made rapidly.
Validation
Before implementationmanagementundertooka thorough validation of the analysis by interviewing samples of formulatorsand end users in each of the major
markets.Specifically, the researchdeterminedwhether
these clustersmade sense to buyers-did they see the
products in the clusters as potentially substitutable?
Were product and cluster differences accountable in
terms of the dimensions suggested in the analysis?
Would buyers see Boliet's products and marketing
support as offering differential advantages over the
competition?The feedback led to certain refinements
in the products'positioning and marketingmixes but
broadly validated the segmentation scheme, increasing the confidence of managementin the strategy.

While for competitivereasonsBoliet's resultsone year

after the launch cannot be presentedin detail, it can
be confirmedthat the company's strategyis having a
significant impact on the market. Management remains confident about the chosen direction, and forecasts of marketand competitivereactionshave proved
to be on target. Currentlymarket share exceeds the
predictedlevel for this stage, and while margins are
lower than forecast, total profits are close to budget.
In summary, the objective of this project was to
develop a marketingstrategyfor a basic raw material
producerswitching into a speciality chemicals market. First, agreed market share objectives had to be
formulatedin the light of a careful assessment of the
competitive environment. In this process an unrealistic top-down set of objectives was replaced by one
which was worked up from a competitive marketappraisaland which then formeda more relevantparameter for the planningand budgeting exercise.
Second, the marketsegments were defined, evaluated, and targetmarketsselected. Multivariateanalysis provideda valuable frameworkfor decomposing
the competitiveenvironmentand building up a viable
positioning strategy. Finally, the marketingmix was
formulatedto match the products' positioning statements.
Managementproved highly receptive to the market segmentationand positioningconcepts. Partlythis
was due to the opportune timing-tougher market
conditions at this time triggered a high felt need for
a marketingoriented approachto the strategic problems facing the business. Second, commitment was
generatedby the high involvement managementwas
encouragedto take in the programand the speed of
today's computer processing, which reduces the ar-

MarketSegmentation
andPositioning/ 31
This content downloaded from 141.6.11.24 on Mon, 27 Jul 2015 08:18:51 UTC
All use subject to JSTOR Terms and Conditions

tificial gestation periods between stages in the discussion. A third factor was the scientific orientation
of managers in typical industrial concerns. As Webster (1978) observed, the complexity of industrial
marketing makes moder management science tools
appealing. Managers in industrial companies appear
to be much more capable of appreciating fairly ad-

vanced statistical and computer based techniques than

their colleagues in consumer goods businesses whose
technical education is generally much weaker. For these
reasons, one might expect to see a greater application
of moder marketing techniques in the industrial area
in the future.

REFERENCES
Ames, B. Charles (1970), "Trappings vs. Substance in Industrial Marketing," Harvard Business Review, 48 (July),
43-102.
Beale, E. M. L. (1969), Cluster Analysis, London: Scientific
Control Systems.
Blashfield, R. K. (1976), "Mixture Model Tests for Cluster
Analysis: Accuracy of Four Agglomerative Hierarchy
Methods," Psychological Bulletin, 23 (no. 3), 377-388.
Business Week(1982a), "Survivalin the Basic Industries"(April
26), 46-52.
(1982b), "Hercules: Cooking Up a New Mix Restores Its Name in Chemicals" (December 13), 80.
Buzzell, Robert D. (1981), "Are There 'Natural' Market
Structures?,"Journal of Marketing, 45 (Winter), 42-51.
Choffrey, J. M. and G. L. Lilien (1980), Market Planning for
New Industrial Products, New York: Wiley.
The Economist (1982), "Speciality Chemicals Need Flair," 285
(November 6), 99.
Everitt, B. (1974), Cluster Analysis, London: Heinemann.
Financial Times (London) (1982), "Specialisation is the Key
to Success" 28, 905 (October 21), II.
Friedman, H. P. and J. Rubens (1967), "On Some Univariate
Criteria for Grouping Data," Journal of the American Statistical Association, 62 (September), 1159-1178.
Green, Paul E. and J. McMennamin (1974), "Marketing Position Analysis," in Marketing Managers' Handbook, S. H.
Britt, ed., Chicago: Dartnell.
Hall, William K. (1980), "Survive Strategies in a Hostile Environment,"Harvard Business Review, 58 (September), 7585.
Hartigan, J. A. (1975), Clustering Algorithms, New York:
Wiley.
Hayes, Robert H. and William J. Aberathy (1980), "Managing on Way to Economic Decline," Harvard Business
Review, 58 (July), 67-77.
Levitt, Theodore (1980), "Marketing Success through Differentiation of Anything," Harvard Business Review, 58 (January), 83-91.

Little, John D. C. (1979), "Decision Support Systems for

Marketing Managers," Journal of Marketing, 43 (Summer), 9-27.
Othmer, Kirk (1980), Encyclopedia of Chemical Technology,
New York: Wiley.
Porter, Michael E. (1980), Competitive Strategy, New York:
Free Press.
Saunders, J. A. (1980), "Cluster Analysis for Market Segmentation," European Journal of Marketing, 14 (no. 7),
422-435.
Shapiro, B. P. and T. V. Bonoma (1984), "How to Segment
Industrial Markets," Harvard Business Review, 62 (MayJune), 104-110.
Ward, J. H. (1963), "Hierarchical Grouping to Optimize an
Objective Function," Journal of the American Statistical
Association, 58 (March), 236-244.
Webster, Frederick E. (1978), "Management Science in Industrial Marketing," Journal of Marketing, 42 (January),
21-27.
(1981), "Top Management's Concerns about Marketing Issues for the 1980's," Journal of Marketing, 45
(Summer), 9-16.
Wind, Yoram (1978), "Issues and Advances in Segmentation
Research," Journal of Marketing Research, 15 (August),
317-337.
and Richard Cardozo (1974), "Industrial Market
Segmentation," Industrial Marketing Management, 3 (October), 153-166.
Wishart, D. (1978), CLUSTAN: User Manual, Program Library
Unit, Edinburgh University.
Wolfe, J. H. (1971), A Monte Carlo Study of the Sampling
Distribution of the Likelihood Ratio for Mixtures of MultiNominal Distributions, Naval Personnel and Training Lab.,
Technical Bulletin STB 72-2, San Diego.
Woo, Carolyn Y. and Arnold C. Cooper (1982), "The Surprising Case for Low Market Share," Harvard Business Review, 60 (November), 106-113.

32 / Journalof Marketing,Spring1985
This content downloaded from 141.6.11.24 on Mon, 27 Jul 2015 08:18:51 UTC
All use subject to JSTOR Terms and Conditions