0% found this document useful (0 votes)

178 views

Segmentation and Targeting

Uploaded by

Maitha Almemari

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

178 views

Segmentation and Targeting

Uploaded by

Maitha Almemari

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 330

Segmentation

and Targeting

1
Outline

• The segmentation-targeting-positioning (STP)

framework
• Segmentation
– The concept of market segmentation
– Managing the segmentation process
– Deriving market segments and describing the
segments
▪ Step 1: Cluster analysis
▪ Step 2: Discriminant analysis
• Targeting

2
Learning goals

• Understand what STP means and how it creates value

• Explain the concept of market segmentation
• Choose relevant segmentation bases based on
differentiability and identifiability
• Know the steps in designing a segmentation study
• Perform a cluster analysis to derive market segments
• Conduct a discriminant analysis to characterize the
market segments
• Select segments for targeting using portfolio analysis
STP - (Segmentation, Targeting, Positioning) is a
Decision Process

– To identify and select groups of potential

customers...
−Organizations, Buying Centers, Individuals

– Whose needs within-groups are similar and

whose needs between-groups are different (S)

– Who can be reached profitably (T)

– With a focused marketing program (P)

How STP Creates Value

Segmentation Targeting Positioning

Create competitive
Identify segments Select segments advantage

Marketing resources are focused to better meet customers

needs and deliver more value to them

Customers develop preference for brands that better meet

their needs and deliver more value

Customers become brand/supplier loyal, repeat purchase,

communicate favorable experiences

Brand/supplier loyalty leads to increased market share and

creates a barrier to competition

Fewer marketing resources needed over time to maintain

share due to brand or supplier loyalty

Profitability (value to the firm) increases

STP – Segmentation, Targeting, Positioning

All consumers
Product in the market
Marketing Mix

Price Target
Target marketing
market
and positioning segment(s)
Place

Promotion

Marketing strategies
of competitors

6
What is Marketing Mix – 4 P’s

7
What is Marketing Mix – 4 P’s

8
What is Marketing Mix – 7 P’s

9
How STP creates value

• More focused marketing efforts can better meet

customer needs
• Customers develop preferences for offerings that
deliver greater value and satisfaction
• Customers become loyal to the brand and the firm if
the brand/firm provides value and satisfaction
• Loyalty leads to greater market share and insulates
the firm from competition
• Profitability increases

10
Motivation for market segmentation

• “One size fits all” usually doesn’t work (all potential

customers are not created equal)

• Segment-of-one marketing is often not feasible

(costs outweigh benefits)

• Compromise: Market segmentation

11
Market segmentation

Partitioning a market that is characterized by

heterogeneity in customers’ response to the
marketing mix into more homogeneous
submarkets.

12
Differentiating segments based on differences in
customer response

Response

Segment B
B2

A2 Segment A
A1
B1

marketing
x1 x2 variable

13
Segmentation bases

General Product-specific

Observable features Behavioral characteristics

of the physical and (user status, loyalty status,
Observable social environment usage rate)
(esp. demographics) Usage situations

Values, lifestyles and Awareness

Latent psychographics, Product attributes and
personality variables benefits
Willingness to buy

14
Problems with many segmentations
• Markets can be segmented using many different variables, but
these variables may not capture differences in response to the
marketing mix;
• Product-specific segmentation bases are usually better
indicators of differences in customer response than general
segmentation bases;
• Particularly motivational variables (purchase motivations,
customer needs, benefits sought) are important for
segmentation;
• However, they are not directly observable, so they have to be
supplemented with managerially useful descriptors that
characterize the segments;

15
Identifying segment members

Response
Who’s
this?
Segment B
B2

A2 Segment A
A1
B1
Who’s
this?

marketing
x1 x2 variable
Segmentation criteria
• The essence of market segmentation:
– market response is homogeneous within segments and
heterogeneous between segments (differentiability)
– individuals can be assigned to a segment based on a meaningful
profile of segment characteristics (identifiability)
• Additional requirements:
– the size and purchasing power of relevant segments can be
determined (measurability)
– the company is able to develop a marketing mix that will appeal
to the members of a given segment (actionability)
– members of a segment can be reached with the appropriate
marketing mix (accessibility)
– segments and segment membership do not change in the short
run (stability)
Recommendation

• Use product-specific segmentation bases (esp.

purchase motivations and product benefits sought)
to derive segments (segmentation variables):
difference in response is key

• Use general segmentation bases (esp. readily

observable customer characteristics) to profile the
segments (discriminant variables): identifiability is
key
Managing the segmentation process
• Define the segmentation problem
– Objectives, resources, and constraints
• Identify data needs
– Primary vs. secondary data
– Sample definition (category users, existing customers, heavy vs. light
users, loyals vs. switchers)
– Segmentation and discriminant variables (based on available data
and/or qualitative research)
• Conduct the segmentation study and analyze the data
– Step 1: Derive the market segments (cluster analysis)
– Step 2: Describe the market segments (discriminant analysis)
• Implement the results
Step 1: Deriving market segments

• The idea is to group (potential) customers who are similar in

their response to some element of the marketing mix (e.g.,
response to different product features, including price;
response to different advertising or promotion campaigns;
response to different distribution channels)

• Choose segmentation variables that capture relevant

response differences, which can eventually be used to
position the firm’s offering to the “right” customers;

• Assume that we have data for a relevant sample of customers

on a set of segmentation variables of interest; how can we
perform a segmentation analysis?
A simple segmentation example:
Preferences of 6 consumers for 2 attributes of coffee

Observations /
Flavor Bitterness
Segmentation Variables

R1 10 10

R2 8 9

R3 5 6

R4 6 5

R5 3 3

R6 1 2
A simple segmentation example:
Preferences of 6 consumers for 2 attributes of coffee
A simple segmentation example:
Preferences of 6 consumers for 2 attributes of coffee
Actual segments in the beer market
(based on Consumer Reports)
Segmentation in the real world

• In practice, we have
– Many potential customers
– Many segmentation variables

• What to do?

Custer analysis to the rescue!

Cluster Analysis

• Basic question: How can objects (customers, brands,

stores, etc.) be grouped such that objects within the
same cluster are similar and objects in different
clusters are dissimilar?
• In segmentation, the objects of interest are
customers and similarity is assessed in terms of
relevant segmentation variables;
• Issues in cluster analysis:
– How is similarity measured?
– How are clusters formed?
– How many clusters should be distinguished?
– How should the clusters be interpreted?
How is similarity measured?
• Overall measures of similarity [not relevant here]
– Direct measures of overall similarity
– Indirect measures of overall similarity (e.g., switching data)
• Derived measures of similarity (e.g., based on
preferences for certain benefits)
– Metric data
• Correlational measures (e.g., similarity in the profile of ratings
across certain benefits)
• Distance measures (Euclidean, city-block)
– Non-metric data
• Matching coefficients (i.e., extent to which customers want
the same features in a product)
Euclidean distance

𝑑𝑅1𝑅6 = (𝑋2 − 𝑋1 )2 +(𝑌2 − 𝑌1 )2

Similarity data as input to cluster analysis

R1 R2 R3 R4 R5 R6

R1 --

R2 S21 --

R3 S31 S32 --

R4 S41 S42 S43 --

R5 S51 S53 S53 S54 --

R6 12.04 S63 S63 S64 S65 --

How are clusters formed?
• Hierarchical cluster procedures: result in a tree-like (nested) structure
that can be represented in a dendrogram;
– Agglomerative (bottom-up) methods: initially there are as many
clusters as objects and then objects are combined;
• Single linkage
• Complete linkage
• Average linkage
• Centroid method
• Ward’s method
– Divisive (top-down) methods: initially there is one large cluster that is
subsequently divided into smaller clusters;
• Non-hierarchical cluster (partitioning) procedures:
– K-means clustering: an initial partition into G groups is chosen and
objects are reassigned if the total error can be reduced; solutions for
different G are analyzed;
Agglomerative methods
• Single linkage: similarity is based on the shortest distance between
any two points in two clusters (nearest-neighbor approach); at
each step, the most similar clusters are joined;
• Complete linkage: similarity is based on the largest distance
between any two points in two clusters (farthest-neighbor
approach);
• Average linkage: similarity is based on the square root of the
average of the squared distances of all objects in two clusters;
• Centroid method: similarity is based on the distance between the
centroids of the clusters;
• Ward’s method: clusters are formed such that the increase in
within-group variability is minimized;
Clustering by similarity

R1 R2 R3 R4 R5 R6

R1 0.00 . . . . .

R2 2.24 0.00 . . . .

R3 6.40 4.24 0.00 . . .

R4 6.40 4.47 1.41 0.00 . .

R5 9.90 7.81 3.61 3.61 0.00 .

R6 12.04 9.90 5.66 5.83 2.24 0.00

Hierarchical agglomerative methods
How many clusters should be formed?

• No generally accepted stopping rule is available;

• In a hierarchical cluster solution, inspect the
dendrogram (tree graph), which shows the distance
(dissimilarity) at which two clusters are joined;
• Look for the point in the dendrogram where
combining two clusters results in a large increase in
the within-cluster heterogeneity;
• Ultimately, a cluster solution should be practically
useful; try out different solutions and choose the one
that is most interpretable and yields the most
actionable insights.
Dendrogram
How should the clusters be interpreted?

• Compute the average score of the cluster members

on the clustering variables used to compute the
similarity measure.

• Name the clusters!

• If additional variables not used during clustering are

available for each of the objects, use these variables
to further profile and differentiate the clusters.
Cluster averages for maltiness and bitterness:
Name the clusters!

Cluster 1 Cluster 2 Cluster 3

Favor 2.0 5.5 9.0

Bitterness 2.5 5.5 9.5

Special problems in cluster analysis
• Clustering variables:
– The final cluster solution depends strongly on the variables that were
included in the cluster analysis. Clustering variables have to be chosen
carefully.
– If clustering variables are very similar, this may exaggerate the
influence of the underlying common factor. If some variables are
highly correlated, it may be better to combine these variables prior to
clustering.
• Outliers: Unusual observations can greatly distort the final solution
obtained in the analysis. Check for outliers before doing the analysis.
Outliers can also be detected in the dendrogram.
• Standardizing the data: Variables with large variances have a
disproportionate influence on similarity. If the clustering variables are
measured on different scales, standardize the data (usually by variable,
but possibly by observation).
Office Star: Segmentation
• 40 respondents rated the importance of 6 attributes when
choosing an office supply store: variety of choice, (availability
of) electronics, (availability of) furniture, quality of service,
low prices, and return policy (segmentation data);
• Importance was rated on a scale from 0 (not at all important)
to 10 (extremely important);
• Data on three descriptor variables are also available: whether
or not the respondent is a professional, the respondent’s
income, and the respondent’s age (discriminant data);
• Data on these three descriptor variables are also available for
an additional 300 respondents for whom no segmentation
data were collected (classification data);
Office Star: Segmentation (input data)
Office Star: Segmentation (options)
Enginius dendrogram
3-cluster solution for Office Star data
Segment size
Population Segment 1 Segment 2 Segment 3

Size 40 18 14 8

Relative size 100% 45% 35% 20%

Segment description
Population Segment 1 Segment 2 Segment 3
Variety of choice 7.53 9.11 6.93 5.00
Electronics 4.58 6.06 2.79 4.38
Furniture 3.45 5.78 1.43 1.75
Quality of service 4.00 2.39 3.50 8.50
Low prices 5.05 3.67 8.29 2.50
Return policy 4.50 3.17 6.29 4.38

Segmentation variables that are statistically different from the rest of the population are high-
lighted in red (lower) or green (higher).
Segment description for Office Star data
Means
Means of segmentation variables by cluster and overall
10

0
Variety of choice Electronics Furniture Quality of service Low prices Return policy

Overall Cluster 1 Cluster 2 Cluster 3

Segment membership of the
first 10 respondents

Segment
Respondent 1 1
Respondent 2 2
Respondent 3 2
Respondent 4 2
Respondent 5 3
Respondent 6 2
Respondent 7 2
Respondent 8 1
Respondent 9 1
Respondent 10 2
Assignment for next class

• Read LRB Chapter 3

• Segmentation Tutorial (Enginius)
• GE McKinsey Matrix Tutorial (Enginius)
Recap: Cluster analysis

• (1) Calculate similarities (or differences) between objects

▪ (2) Derive clusters

Step 1: (R1&R2) vs. R3 vs. R4 Step 1: (R1&R2) vs. R3 vs. R4
Step 2: (R1&R2) vs. (R3&R4) Step 2: (R1&R2) vs. (R3&R4)
Step 3: (R1&R2) & (R3&R4) Step 3: (R1&R2) & (R3&R4)
48
Recap: Cluster analysis

▪ (3) Choose the number of clusters based on the dendrogram

▪ (4) Interpret the clusters

Seg 1 Seg 2 Seg 1 Seg 2 Seg 3
Tartar control 9.5 2.5 Tartar control 9.5 9.0 1.0
Whitening 1.5 10.0 Whitening 1.5 10.0 10.0
49
Step 2: Describing market segments

• In order to make the segmentation actionable, the

market segments have to be profiled (particularly if
the segmentation variables are not directly
observable);
• The segmentation study should include readily
observable variables that can be used to characterize
the segments;
• The goal is to find actionable variables that are useful
for predicting customers’ segment membership;
• Enginius uses either the binary or the generalized
logit model (to be discussed later) to accomplish this;

50
Discriminant analysis using the logit model

• Basic question: How can we predict the cluster

(segment) membership of an object (customer)
based on certain explanatory (discriminant)
variables?

• Issues in discriminant analysis:

– How can clusters (segments) be differentiated based on
many potential profiling variables, and which variables are
most effective in discriminating between the clusters
(segments)?
– How can we assess the overall quality of discrimination?

51
Predicting segment membership
• We know the cluster (segment) membership of customers
(based on the cluster analysis in the first step) and we can use
the discriminant variables to predict cluster membership;
• The largest segment is chosen as the baseline and we estimate
(G-1) sets of coefficients (where G is the total number of
segments) such that the discriminant variables optimally
differentiate the other segments from the baseline segment;
• The magnitude and significance of the coefficients tell us which
variables discriminate well between the segments;
• For purposes of classification, observations are assigned to the
segment for which the model predicts the highest membership
probability;

52
Assessing prediction accuracy

• To assess the overall quality of discrimination we can use a

hits-and-misses table (confusion matrix);
• For two groups this looks as follows:

Predicted
Group 1 Group 2
Group 1 Correct Incorrect
Actual
Group 2 Incorrect Correct

• To assess classification accuracy, we need a benchmark for

chance prediction:
❑Use the proportional chance criterion: σ 𝑝𝑖2 , where the pi are the
prior probabilities of segment membership;
53
Example of two-group discriminant analysis with two
classification variables
Discrimination Data

x1 x2
Variables / Observations Cluster Age group Level of education
(younger to older) (low to high)

1 1 1 3
2 1 1 5
3 1 2 4
4 1 5 2
5 2 2 8
6 2 4 8
7 2 5 6
8 2 6 4
9 2 7 7
10 2 8 5
54
Prediction model

Parameter estimates Segment 1

(Intercept) 30.1
x1 -3.39
x2 -3.88

p-values Segment 1
(Intercept) 0.577
x1 0.661
x2 0.614

[ Segment 2 is used as the baseline ]

55
Discriminant variable means by segment

Population Segment 1 Segment 2

X1 (Age) 4.10 2.25 5.33

X2 (Education) 5.20 3.50 6.33

Note: If there are few differences between clusters based on the discriminant
variables, the discriminant analysis will not yield useful results!

56
Assessing the quality of discrimination

Counts Predicted 1 Predicted 2 Total

Segment 1 4 0 4
Segment 2 0 6 6
Total 4 6 10

Proportions Predicted 1 Predicted 2 Total

Segment 1 100% 0% 100%
Segment 2 0% 100% 100%

[ Proportional chance criterion = 52% ]

57
Office Star Data

▪ 40 respondents rated the importance of 6 attributes when

choosing an office supply store: variety of choice, (availability
of) electronics, (availability of) furniture, quality of service,
low prices, and return policy (segmentation data);
▪ Importance was rated on a scale from 0 (not at all important)
to 10 (extremely important);
▪ Data on three descriptor variables are also available: whether
or not the respondent is a professional, the respondent’s
income, and the respondent’s age (discriminant data);
▪ Data on these three descriptor variables are also available for
an additional 300 respondents for whom no segmentation
data were collected (classification data);

58
Prediction model

Parameter estimates Segment 2 Segment 3

(Intercept) 6.19 -4.57
Age -0.106 0.093
Income -0.0000701 -0.0000478
Professional 0.48 3.03

P-values Segment 2 Segment 3

(Intercept) 0.000 0.000
Age 0.000 0.001
Income 0.006 0.089
Professional 0.000 0.000

[ Segment 1 is used as the baseline ]

59
Discriminant variable means by segment

Population Segment 1 Segment 2 Segment 3

Age 40.5 44.2 30.9 49.0

Income 42 500 48 300 32 100 47 500

Professional 0.475 0.333 0.500 0.750

60
Hits-and-misses table (confusion matrix) for Office Star data

Predicted 1 Predicted 2 Predicted 3 Total

Segment 1 13 2 3 18
Segment 2 2 12 0 14
Segment 3 3 1 4 8
Total 18 15 7 40

Predicted 1 Predicted 2 Predicted 3 Total

Segment 1 72% 11% 17% 100%

Segment 2 14% 86% 0% 100%

Segment 3 38% 13% 50% 100%

Overall hit rate = 73%, proportional chance criterion = 37%

61
In-sample model predictions
(first 10 respondents)

Prob(cluster 1) Prob(cluster 2) Prob(cluster 3) Predicted Actual Correct

Respondent 1 28% 11% 60% 3 1 0

Respondent 2 37% 57% 7% 2 2 1
Respondent 3 18% 81% 1% 2 2 1
Respondent 4 22% 42% 36% 2 2 1
Respondent 5 17% 2% 81% 3 3 1
Respondent 6 20% 69% 11% 2 2 1
Respondent 7 36% 50% 15% 2 2 1
Respondent 8 40% 14% 46% 3 1 0
Respondent 9 79% 5% 16% 1 1 1
Respondent 10 35% 64% 0% 2 2 1
Do the discrimination results make sense for the 3-
cluster solution?

Step 1: Deriving market segments

Population Segment 1 Segment 2 Segment 3
Variety of choice 7.53 9.11 6.93 5.00
Electronics 4.58 6.06 2.79 4.38
Furniture 3.45 5.78 1.43 1.75
Quality of service 4.00 2.39 3.50 8.50
Low prices 5.05 3.67 8.29 2.50
Return policy 4.50 3.17 6.29 4.38

Step 2: Describing the market segments

Population Segment 1 Segment 2 Segment 3

Age 40.5 44.2 30.9 49.0

Income 42 500 48 300 32 100 47 500
Professional 0.475 0.333 0.500 0.750

63
Classification results (out-sample) for
300 additional respondents

Prob(cluster 1) Prob(cluster 2) Prob(cluster 3) Predicted

Customer 1 36% 50% 15% 2
Customer 2 88% 5% 7% 1
Customer 3 6% 3% 91% 3
Customer 4 35% 64% 0% 2
Customer 5 25% 1% 74% 3
Customer 6 18% 81% 1% 2
Customer 7 40% 14% 46% 3
Customer 8 42% 48% 10% 2
Customer 9 42% 48% 10% 2
Customer 10 20% 69% 11% 2

Population Segment 1 Segment 2 Segment 3

Size 300 138 108 54
Relative size 100% 46% 36% 18%

64
Recap: Discriminant analysis

• Choose discriminant variables that can be expected to be

predictive of segment membership;

• Run the discriminant analysis and assess the overall quality of

the discrimination based on the confusion matrix (hits-and-
misses table) and the proportional chance criterion;

• Assess the usefulness of individual discriminant variables for

profiling different segments;

• Classify new customers into segments based on their scores

on the discriminant variables;

65
Illustrative confusion matrix

66
Target marketing

• evaluation of the attractiveness of each market

segment and selection of target segments;

• evaluation of market segments based on

– market segment characteristics (attractiveness)
– company objectives and resources (competitive position)

• selection of target segments can result in

– undifferentiated (mass) marketing
– differentiated marketing
– concentrated marketing

67
Portfolio analysis

▪ portfolio models are tools to allocate scarce

resources to different businesses (e.g., product
markets) in a multi-business firm;
▪ steps in portfolio analysis:
□ identify strategic business units (or SBUs);
□ rate each SBU in terms of market attractiveness and
competitive position;
□ decide whether to build, maintain, harvest, or divest a
business;
▪ the goal is to have a balanced portfolio of
businesses which will ensure profitability and
growth in the long run;
68
BCG growth-share matrix

20%

maintain leadership
?
and build future build share
cash cow or divest
market 10%
growth rate

harvest and manage

for maximum divest
profitability
0%
10x 1x .1x
relative market share
69
Steps in constructing a market attractiveness/competitive
position matrix for selecting target markets

▪ List the segments to be evaluated and estimate their size

▪ Identify the key factors determining market attractiveness
(e.g., size, growth, margins, current competition) and
competitive position (e.g., product fit, access, brand
reputation, current penetration)
▪ Assign weights to each factor (e.g., 1 = least important, 5 =
most important)
▪ Rate each segment on the factors (e.g., 1 = worst, 5 = best)
▪ Calculate each segment’s market attractiveness and
competitive position score
▪ Plot each segment in the matrix

70
Office Star data [made up – not in ME]

71
72
Office Star data

73
Recommended strategies based on market
Attractiveness/Competitive Position Matrix

Competitive position

Low Medium High

Opportunities Maximum investment

Market attractiveness

investment Invest to challenge Consolidate

High Build strength the leader
or exit position

Selective
Harvest or Cautious investment
Medium divest investment Build on
strengths

Harvest or Harvest or Protect position

Low divest divest Manage for cash
generation
ISBM segmentation case

• Question 1
Select Run Segmentation from the Enginius dashboard (without Discrimination) on the
data to try to identify the number of distinct segments present in this market.
• Question 2
Identify and profile the clusters that you select. Given the attributes of the ISBM (see
the Perceptual Data for ISBM), which clusters would seem to be core for Ralph’s
retention and customer acquisition campaigns?
• Question 3
Rerun the segmentation with the Discrimination Option. What are the organizational
characteristics that describe each of the segments? Do they make sense? Why?
• Question 4
Using the Segmentation tool in Enginius, re-run segmentation and also include
classification data (Prospects). What firms should Ralph target for his customer
acquisition program?
• Question 6
Comment (very briefly) on the value of the approach and its limitations in providing
both strategic and tactical direction for the ISBM.
75
Positioning

1
Outline

• The concept of product positioning

• Conducting a positioning study

• Perceptual mapping using principal component

analysis

• Incorporating preferences into perceptual maps

2
Learning goals

• Explain the concept of positioning, understand the

fundamental issues in positioning, and be able to write
a positioning statement
• Know the steps in designing a positioning study
• Construct a perceptual map using principal component
analysis and interpret the resulting map
• Incorporate vector or ideal preferences into perceptual
maps and derive marketing insights based on the
distribution of preferences
STP – Segmentation, Targeting, Positioning

All consumers
Product in the market
Marketing Mix

Price Target
Target marketing market
and positioning segment(s)
Communication

Distribution

Marketing strategies
of competitors

4
Positioning

 Position – the way the product is defined by

consumers on important attributes - the place an
offering occupies in consumers’ minds on important
attributes relative to competing products.

 Repositioning – changing the place an offering

occupies in a consumer’s mind relative to
competitive products.
Positioning Examples
Effective Positioning

 Identify relevant set of competitive products serving a target

market
 Identify the set of determinant attributes that define the “product
space” in which positions of current offerings are located.
 Collect information from a sample of customers and potential
customers about perceptions of each product on the determinant
attributes.
 Determine brands current location (position) in the product space.
 Determine customers most preferred combination of determinant
attributes
 Examine the fit between preferences of market segments and
current position of brands.
 Identify positions where new brands might be placed.
 Write positioning statement or value proposition to guide
development and implementation of marketing strategy.
Perceptual Map
Perceptual Map
Central questions in positioning
(based on Rossiter and Percy)

 A brand’s positioning should tell customers

□ what the brand is – what category need it satisfies (brand-
market positioning),
□ who the brand is for – what the intended target audience is
(brand-user positioning), and
□ what the brand offers – what benefits it provides (brand-benefit
positioning)
 The selection of benefits to emphasize should be
based on
□ importance (relevance of the benefit to target customers’
purchase motives in the category),
□ delivery (the brand’s ability to provide the benefit), and
□ uniqueness (differential delivery of the benefit)
What is Positioning?

Category
Need

What the brand is? User

Brand

Benefit(s)
Positioning Statement

 To [the target audience]

 ________ is the (central or differentiated) brand of
[category need]
 that offers [brand benefit(s)]. The positioning for this
brand
□ should emphasize [benefit(s) uniquely delivered],
□ must mention [benefit(s) that are important “entry
tickets”],
□ and will omit or trade off [inferior-delivery
benefits].
Illustrative positioning statement for Volvo automobiles

 To upper-income car buyers

 Volvo is a (differentiated) brand of prestige automobile
 that offers safety, performance, and prestige. The
positioning for this brand
□ should emphasize safety and performance,
□ must mention prestige as an entry ticket to the
category,
□ and will downplay the previous family-car association.
Issues to consider when thinking about positioning

 What are the central dimensions that underlie customers’

perceptions of brands in the product class?
 How do customers view our brand on these dimensions?
 How do customers view our competitors?
 Are we satisfied with the way our customers view our brand
relative to the competition, or are changes required? Are
there opportunities for new product introductions?
 How do perceptions relate to preferences?
 How can we improve our competitive position (market share)
given the distribution of preferences in the market?
Office Star data

 Four different stores are rated on five attributes;

 The stores are Office Star (the target store), Paper &
Co., Office Equipment, and Supermarket;
 The attributes are large choice, low prices, service
quality, product quality, and convenience;
 24 customers rated all stores on all attributes on a
scale from 1 to 6 (e.g., the extent to which Office Star
offers large choice);
 We also know customers’ overall preferences for the
four stores (rated on a 1-5 scale);
Input data
Perceptual data

OfficeStar Paper and Co Office Equipment Supermarket

Large choice 5.2 4.4 3.9 2.3
Low prices 2.1 4.5 2.6 4.1
Service quality 4.2 2.3 3.1 1.8
Product quality 3.7 2.6 3.1 2.9
Convenience 2.7 1.4 4.7 5.1

Note: Values in red if below their column mean minus one standard deviation and in green if above
their column mean plus one standard deviation.
Attribute histograms
Percentage of variance explained by
each principal component (dimension)
Variance explained
Variance and cumulated variance explained, by dimension.

Variance explained Cumulative variance

Dimension 1 66.9% 66.9%
Dimension 2 30.8% 97.7%
Dimension 3 2.3% 100.0%
Dimension 4 0.0% 100.0%
Coordinates for objects (stores) and attributes

Dimension I Dimension II

OfficeStar 0.895 -0.023

Paper and Co -0.358 -0.644
Office Equipment 0.131 0.294
Supermarket -0.668 0.373

Dimension I Dimension II
Large choice 0.620 -0.623
Low prices -0.674 -0.450
Service quality 0.985 0.006
Product quality 0.510 0.305
Convenience -0.159 0.917
Perceptual map for Office Star data
Perceptual map for Office Star data
More Examples on
Perceptual Mapping
Recap: Perceptual mapping

 We have average perceptions for a number of brands on

(many) different attributes/benefits;
 We want to map both the attributes and the brands in a
space of low dimensionality without losing too much of the
original information;
Recap: Deciding on the number of dimensions

Variance Explained
Variance explained as a function of the number of dimensions.
Dimensions / Items 1 2 3 4 5
Total variance explained 0.502 0.265 0.194 0.039 0.000

Cumulative variance explained 0.502 0.767 0.961 1.000 1.000

Statistics (for two dimensions)

Descriptive statistics about the input data.

Proportion Variance
Attribute Mean Variance
Explained

Exciting 6.30 1.043 0.499

Cooling Effect 5.76 1.432 0.942
Chewy 7.08 1.270 0.918
Hard 7.06 0.671 0.911
Long Lasting 5.66 1.376 0.918
Fresh 4.32 0.757 0.909
Flavours 5.96 1.717 0.273
Recap: Interpreting the perceptual map of mints

Positioning Map

Chewy

Nutrine Chlormint

Cooling Effect
Mint-O-Fresh
Exciting
I (50.2%)
Flavours
Fresh
Mentos

Long Lasting

Hard Mahalacto
II (26.5%)
Recap: Interpreting the perceptual map of mints

Positioning Map

Chewy

Nutrine Chlormint

Cooling Effect
Mint-O-Fresh
Exciting
I (50.2%)
Flavours
Fresh
Mentos

Long Lasting

Hard Mahalacto
II (26.5%)
Incorporating preferences into perceptual maps

 Perceptual maps tell us how customers perceive brands, but

they are silent about which brands they prefer;
 In order to understand customers’ choices, we have to
incorporate their preferences;
 Two types of preference models can be distinguished:
□ Vector preferences: an increase in the amount of the
attribute increases preference;
□ Ideal point preferences: there is an ideal amount of the
attribute at which preference is highest;
Vector vs. ideal point preferences

Vector preferences Ideal point preferences

preference preference

Amount of attribute Amount of attribute

Incorporating preferences (cont’d)

 For vector preferences:

□ the brand whose (orthogonal) projection on a consumer’s
preference vector is farthest from the origin (in the
direction of the arrow) is the consumer’s preferred brand;

 For ideal point preferences:

□ the brand that is closest to a consumer’s ideal point is the
consumer’s preferred brand;
A simple way to incorporate preferences

 For vector preferences:

□ calculate the average preference of target customers for each
brand; add these averaged preferences as another “attribute” to
the analysis; alternatively, we could use the market shares of
the brands as a proxy indicator of preference;
□ the preference vector in the resulting map shows the direction
of increasing preferences in the market;
 For ideal-point preferences:
□ introduce a hypothetical ideal brand and have respondents rate
this ideal brand on all the attributes;
□ the location of the ideal brand in the map indicates the most
preferred combination of attributes;
Office Star data

 Four different stores are rated on five attributes;

 The stores are Office Star (the target store), Paper & Co.,
Office Equipment, and Supermarket;
 The attributes are large choice, low prices, service
quality, product quality, and convenience;
 24 customers rated all stores on all attributes on a scale
from 1 to 6 (e.g., the extent to which Office Star offers
large choice);
 We also know customers’ overall preferences for the
four stores (rated on a 1-5 scale);
Input data
Average store preferences

Average preference
OfficeStar 3.46
Supermarket 3.42
Office Equipment 3.21
Paper and Co 2.71
Perceptual-preference map
Customer preferences (excerpt)

Dimension I Dimension II Dimension III

Lori 0.099 -0.089 -0.991
Mary -0.176 -0.040 -0.984
Radjeep -0.523 0.391 -0.757
Antoine -0.073 0.522 -0.850
Yoshi -0.134 0.415 -0.900
Hubert 0.707 -0.074 0.704
Michael 0.720 0.613 -0.324
Elisabeth -0.050 0.105 0.993
Mike 0.607 0.362 -0.707
Hal -0.208 -0.361 -0.909
Positioning by segment
Segment-specific individual preferences
Segment-specific average preferences
Average store preference by segment

Average
Segment A Segment B Segment C
preference

OfficeStar 3.46 4.60 2.50 3.00

Supermarket 3.42 2.40 3.90 4.75
Office Equipment 3.21 3.20 3.40 2.75
Paper and Co 2.71 1.90 4.00 1.50
Average store preference by segment
Market share predictions based on preferences
Recap: Ideal vs. vector preferences

Brand B

Preference vector

Ideal brand

Brand A
Brand A
Recap: Preference mapping with an ideal brand

Positioning Map

Hard
Mahalacto

Ideal Mint

FreshLong Lasting
Exciting Flavours
I (41.2%)

Nutrine Mentos
Mint-O-Fresh
Cooling Effect

Chlormint
Chewy
II (27.1%)
Recap: Vector preferences for 3 respondents

Positioning Map

Chewy

Nutrine Chlormint

Cooling Effect
Mint-O-Fresh
Exciting
I (50.2%)
Flavours
Fresh
Mentos

Long Lasting
R3

Market Share Hard Mahalacto

II (26.5%)
Recap: Vector preferences for 3 respondents

Positioning Map

Chewy

Nutrine Chlormint

Cooling Effect
Mint-O-Fresh
Exciting
I (50.2%)
Flavours
Fresh
Mentos

Long Lasting
R3

Market Share Hard Mahalacto

II (26.5%)
Positioning the Infiniti G20

1. Describe the two (or, if applicable, three) dimensions underlying the perceptual
maps that you generated. Based on these maps, how do people in this market
perceive the Infiniti G20 compared with its competitors?
2. Infiniti promoted the G20 as a Japanese car (basic version $17,500) with a
German feel, basically a car that was like the BMW 318i ($20,000), but lower
priced. Is this a credible claim, given the perceptions and preferences of the
respondents?
3. Which attributes are most important in influencing preference for these cars in
the three segments (S1, S2 and S3) shown on these maps? To which segment(s)
would you market the Infiniti G20? How would you reposition the Infiniti G20 to
best suit the chosen segment(s)? Briefly describe the marketing program you
would use to target the chosen segment(s).
4. What ongoing research program would you recommend to Infiniti to improve its
evaluation of its segmentation of the market and positioning of its G20?
5. Summarize the advantages and limitations of the software provided for this
application.
[Focus on the first three questions]
Segmentation and
Targeting
1. STP (Segmentation, Targeting, and
Positioning)
2. Needs-based segmentation
3. Cluster Analysis
4. Discriminant Analysis

• Market segmentation is the subdividing of a market into distinct

subsets of customers.

Segments

• Members are different between segments but similar within.

STP - (Segmentation, Targeting, Positioning) is a Decision Process

• To identify and select groups of potential customers...

− Organizations, Buying Centers, Individuals
• Whose needs within-groups are similar and whose needs between-groups are
different (S)
• Who can be reached profitably (T)
• With a focused marketing program (P)

Segmentation Targeting Positioning

Create competitive
Identify segments Select segments advantage

Marketing resources are focused to better meet customers needs and deliver
more value to them

Customers develop preference for brands that better meet their needs and
deliver more value

Customers become brand/supplier loyal, repeat purchase, communicate

favorable experiences

Brand/supplier loyalty leads to increased market share and creates a barrier

to competition

Fewer marketing resources needed over time to maintain share due to brand
or supplier loyalty

Profitability (value to the firm) increases

Short term segmentation applications:

• Salesforce allocation/call planning
• Channel assignment
• Communication program
• Pricing
• Today’s competitors and my current relative advantage to the
customer

Longer term:
• Emerging needs
• New and evolving market segments to serve
• Planning for segment development/growth
• Not in kind competition/threats (satisfying customer needs in
different ways)
• Lead user identification and management
• Market driving (vs. customer focused)

1. Define segmentation problem Business

Understanding
Data Understanding

2. Identify data needs

3. Conduct market research Data Preparation

Deployment
4. Build segmentation database

5. Define market segments Data Modeling

6. Describe market segments

Evaluation
7. Implement results

• Internal assessment and planning

• Objective(s) of segmentation? (critical step!)
• Resources?
• Constraints?
• …

• Database review
• Primary data available?
• Secondary data available?
• …

What kind of data is most suited to answer your business problem?

• Socio-demographic? Firmographic?

• Past behavior? Purchase data? Scanner data?

• Deep needs? Benefits sought? Perceptions?

• Do you have the data you need, readily available?

NO YES

Market research Database marketing

• Collect data, do segmentation • Use the available data, usually

analysis on a small (random) sample behavioral
of customers, but leverage the
insights across the entire
market/customer base

Understand how my brand is

perceived by different segments I Identify shopping habits among my
in the market. existing customers (e.g., big
spenders, promotion-driven
customers, etc.)

• Qualitative study
− Interviews, sources, materials
− “Deep needs” identification
− Decision-making process assessment
− Which criteria to use, what questions to include?

• Quantitative study
− Sample design (which respondents/informants?)
− Questionnaire development
− Data collection

Assemble different data sources, with two types of data:

1- Segmentation variables (bases)

• Characteristics that tell us why segments differ
(e.g., needs, wants, benefits, solutions to problems, usage situation, usage
rate, decision processes, past behavior)

2- Discriminant variables (descriptors)

• Characteristics that tell us how to find, reach, identify segments
(age, income, education, profession, lifestyles, media habits, use occasions;
industry, size, location, organizational structure)

• Two core approaches to segmentation

− Need-based segmentation
− Choice-based segmentation (next week)

• How many segments?

• How are they defined?

− Segmentation variables

• How can they be described, reached?

− Discriminant variables

© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-commercial educational
purposes only is authorized if this notice appears on every copy. 14
A Four-Phase Process for Conducting a
Successful Segmentation Project
Phase I Phase II Phase III Phase IV
Planning and Design Qualitative Assessment Quantitative Measurement Analysis and Implementation

Internal Assessment Qualitative Quantitative

& Planning Research Survey
Segmentation
• Objective(s) of • Interview Materials • Sample Design Analysis
segmentation Development ·
· · • Cluster Analysis
• Resources
· · • Portfolio Analysis Implementation
• Constraints Through
• Qualitative Data • Positioning Analysis Database Toolsc
Collection
Database · • Questionnaire • Call Center
Review · Development
· • Web
• Primary data already • “Deep needs”
· • Sales call patterns
available Identification
·
· Classification Tool • Promotion
• Secondary data · Development
• ….
• … • Decision-Making
• Data Collection • Discriminant function
Process Assessment
·
· • Binary (CART) tree
Prototype ·
·
Implementation · • …
Exercises

• What ifs?
• Relevant groups
involved?
Basic Idea: Do segmentation analysis on a small (random) sample of customers, but leverage the insights across the
• ….. entire customer base.

• Bases—characteristics that tell us why segments differ (e.g., needs,

preferences, decision processes).

• Descriptors—characteristics that help us find and reach segments.

(Business markets) (Consumer markets)

Industry Age/Income
Size Education
Location Profession
Organizational Life styles
structure Media habits

Consumer Markets B2B Markets

Segmentation Needs, wants, benefits, solutions to problems, Needs, wants, benefits, solutions to problems,
Bases usage situation, usage rate. usage situation, usage rate, size*, industry*.

Descriptors Age, income, marital status, family type & size, Industry, size, location, current supplier(s),
Demographics gender, social class, etc. technology utilization, etc.

Psychographics Lifestyle, values, & personality characteristics. Personality characteristics of decision makers.

Behavior Use occasions, usage level, complementary & Use occasions, usage level, complementary &
substitute products used, brand loyalty, etc. substitute products used, brand loyalty, order
size, applications, etc.

Decision Making Individual or group (family) choice, low or high Formalization of purchasing procedures, size &
involvement purchase, attitudes and knowledge characteristics of decision making group, use of
about product class, price sensitivity, etc. outside consultants, purchasing criteria,
(de)centralizing buying, price sensitivity,
switching costs, etc.

Media Patterns Level of use, types of media used, times of use, Level of use, types of media used, time of use,
etc. patronage at trade shows, receptivity of sales
people, etc.

• Customers are grouped together based on their similarities of

− Needs
− Wants
− Lifestyles
− Behavior…

Customers in segment “X” highly value prestige, peace of mind, and

are not price-sensitive (bases). This segment is mostly composed of
men with high income who have been loyal customers for more than 3
years (descriptors).

A simple survey to understand customers’ needs

• Context: B2B, IT industry, assembly parts, direct marketing (catalog)

• Question 1: “On a scale from 1 to 7, how many items/references would you

like to find in our catalog?”
− 1 = very few (key references only, but easy to find)
− 7 = many (very large choice with many references)

• Question 2: “How much technical assistance and support would you expect
from us in your decision-making process?”
− 1 = none (I know exactly what I want and don’t need help)
− 7 = a lot (I will require a lot of support and expertise from you)

Each respondent can be represented by a line in a table

• There are 8 respondents
• The first respondent answered “1” to the first question, and “7” to
the second question

Q1 Q2
Customer 1 1 7
Customer 2 6 6
Customer 3 5 2
Customer 4 2 7
Customer 5 2 6
Customer 6 7 2
Customer 7 7 7
Customer 8 6 1

Each respondent can also be plotted on a chart

7
Q1 Q2
Customer 1 1 7 6

Q2: technical assistance

Customer 2 6 6 5
Customer 3 5 2
4
Customer 4 2 7
3
Customer 5 2 6
Customer 6 7 2 2

Customer 7 7 7 1
Customer 8 6 1
0
0 1 2 3 4 5 6 7 8
Q1: size of catalog

How many segments of customers do you see?

7
Q2: technical assistance

0
0 1 2 3 4 5 6 7 8
Q1: size of catalog

Intuitively, customers can be segmented into three segments with

homogeneous needs
8

7
Segment 1 Segment 2
6 • Large catalog
• Small catalog
Q2: technical assistance

• Lot of support
• Lot of support
5

2 Segment 3
• Large catalog
1 • Not much support

0
0 1 2 3 4 5 6 7 8

Q1: size of catalog

How to reach that result more formally?

1. At first, consider each customer as a separate segment

2. Find the two segments/customers that, if grouped together, would

lead to the lowest loss of information

3. Continue merging segments/customers until such merging would

lead to an unacceptable loss of information

We start with 8 segments. Which should be grouped first?

7
Q2: technical assistance

0
0 1 2 3 4 5 6 7 8
Q1: size of catalog

Which should be grouped next?

Q2: technical assistance 6

4
These two first
These two customers are the most 3
similar. Treating them as if they had
exactly the same needs will lead to the 2
smallest loss of information for the firm.1

0
0 1 2 3 4 5 6 7 8
Q1: size of catalog

Which should be grouped next?

7
Q2: technical assistance

0
0 1 2 3 4 5 6 7 8

Q1: size of catalog

Which should be grouped next?

7
Q2: technical assistance

0
0 1 2 3 4 5 6 7 8

Q1: size of catalog

Which should be grouped next?

7
Q2: technical assistance

0
0 1 2 3 4 5 6 7 8

Q1: size of catalog

Which should be grouped next?

7
Q2: technical assistance

0
0 1 2 3 4 5 6 7 8

Q1: size of catalog

Which should be grouped next?

7
Q2: technical assistance

3
Large loss of information!
2
This new segment groups customers with very
different needs in terms of size of catalog.
1
Customers’ needs in that segment are not
homogeneous anymore. We should not go that
0
far. 0 1 2 3 4 5 6 7 8

Q1: size of catalog

You can segment your customers into… One segment

(not very useful, though)

7
Q2: technical assistance

0
0 1 2 3 4 5 6 7 8

Q1: size of catalog

Real-life applications will be much more complex:

• Many more questions (dimensions)

• Many more respondents (points)

8
7
6
5
4
3
2
1
0
0 1 2 3 4 5 6 7 8

• We need a more systematic approach to treat large datasets

− A more formal definition of “similarity”
− A more formal definition of “loss of information”
© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-commercial educational
purposes only is authorized if this notice appears on every copy. 33
Formal definition for “similarity”

We define “similarity” as the Euclidean distance between

the responses of two respondents

d ab  xa  xb    ya  yb  ...  za  zb 

2 2 2

• dab is the distance/similarity between respondents a and b

− The greater the distance, the smaller the similarity

• xa and xb are their answers to question “x”

− The formula can deal with as many questions as required

• Questions need to use the same scale – or be standardized

The dendogram shows how much

distance separates the next two closest
segments

• The value on the y-axis

indicates how much
distance we need to travel
to join two segments

• i.e., how much

information is lost
when two segments
are grouped together

The dendogram shows how much

distance separates the next two
closest segments

• If there is a sudden
jump, stop

(here  3 segments)

Once segments are formed, describe them by their means

Segmentation Variables
Means of each segmentation variable for each segment

Segmentation variable Overall Cluster 1 Cluster 2 Cluster 3

/ Cluster
Q1 4.75 1.7 6.0 6.5
Q2 4.5 6.7 1.7 6.5

• Much easier to deal with than with each customer individually

Name the segments

But beware. Segment names will stick. Name them carefully
8

7
Clueless? Demanding?
6 • Large catalog
• Small catalog
Q2: technical assistance

• Lot of support
• Lot of support
5

2 Self-service?
• Large catalog
1 • Not much support

0
0 1 2 3 4 5 6 7 8

Q1: size of catalog

• Often, surveys include additional information about respondents, not

related to their needs, but informative, called discriminant variables
(descriptors)

• They do not describe what they want or why they buy

They describe who they are and how they can be reached:
− B2C : Age, Gender, Profession, Income, Number of children, Magazines they
read, TV shows they watch, etc.
− B2B : Industry, Revenues, Sales, Number of employees, etc.

Segments should not be formed on these variables!

Describe the segments by their discriminant variable means

Discriminant Variables
Means of each discriminant variable for each segment.

Discriminant variable / Cluster Overall Cluster 1 Cluster 2 Cluster 3

Number of Employees 2.93 4.20 3.10 1.50

% Export 0.46 0.48 0.78 0.12
Computer industry 0.26 0.60 0.14 0.05
Telecom industry 0.34 0.31 0.60 0.11
Other industry 0.40 0.09 0.26 0.84

• Customers in segment 2 export a lot, and can be found mostly in the Telecom
industry

Informs us about who they are, but not what they want

How can we assign all customers/prospects to the right segment?

• If they replied to the survey, that’s easy…

• But most have not

Can I predict to which segment a customer will belong

based on his/her descriptors?

Remember:

• Segmentation variables (bases) define the segments.

This information is only available from the survey

• Discriminant variables (descriptors) describe the segments.

This information is in the survey, but also often readily available from
secondary sources

How to assess whether discriminant variables predict

cluster memberships well?
• Confusion matrix
• Hit rate

Confusion Matrix
Comparison of cluster membership predictions based on discriminant data, and actual cluster
memberships. High values in the diagonal of the confusion matrix (in bold) indicates that discriminant data
is good at predicting cluster membership.

Actual / Predicted cluster Cluster 2 Cluster 2 Cluster 3

Cluster 1 79.6% 15.1% 05.4%

Cluster 2 17.6% 72.5% 09.8%
Cluster 3 06.2% 18.8% 75.0%

Hit Rate (percent of total cases correctly classified) 76.88%

The most recurring question is often:

Which segment to go after?

(with what product/offering, at what price, communicating
primarily on what benefits, etc.)

Undifferentiated SEGMENT A

Differentiated SEGMENT B

Focused SEGMENT C

• Overall appeal of that segment

− Size
− Growth rate
− Margins, revenues
− Volumes…

vs.

• Relative advantage of the firm to serve this segment

− Access, channels
− Brand, reputation
− Product, service fit
− Production capabilities…

• Overall appeal of that segment

− Size “External/objective” strength
− Growth rate An attractive segment is equally
− Margins, revenues appealing to all firms competing in
− Volumes… that market

vs.

• Relative advantage of the firm

to serve this segment
− Access, channels “Internal/subjective” strength
− Brand, reputation This dimension is specific to each
− Product, service fit firm’s competitive situation
− Production capabilities…

Both dimensions (internal and external strengths) have characteristics

that make them difficult to assess

• They are complex

− They have many underlying, sometimes conflicting dimensions

• Different stakeholders within the firm will weight these underlying

dimensions differently
− Ex., market share vs. targetability vs. solvability

• They need to be assessed subjectively

− Hard to get hard numbers on “product fit” or “reputation”

• The GE McKinsey Matrix has been widely used to measure the relative
appeals of business units or divisions

• Assessed by:
− Market attractiveness (“external” appeal)
− Business unit strength (“internal” appeal)

• The GE/McKinsey matrix is a multi-dimentional version of the much

more simple BCG Matrix (Boston Consulting Group)
The BCG Matrix
www.12manage.com
Business Growth Rate (Market Growth)

High
Select a Divest
few the others

Liquidate

Low

High Low
Relative Position (Market Share)

The GE/McKinsey matrix in 6 steps:

1. List segments, and estimate size of that segment

2. List dimensions that drive segment attractiveness

• Size, volume, growth, margins, current competition…

3. List dimensions that underlie competitive advantage of the firm

• Product fit, access, brand reputation, current penetration…

4. Rate each segment on each dimension (1=worst, 5=best)

• Forces group discussion, builds consensus

5. Weigh each dimension (1=least important, 5=most important)

• Again, forces group discussion, builds consensus

6. Compute simple summed multiplications ( ratingsweights), and plot

Horizontal Axis (ratings, weights)

On a scale from 1 to 5, rate Segments on each factor, and weight the importance of each factor.

Competitive
Advantage Segment 1 Segment 2 Segment 3 Weights
Product fit 5 3 4 5
Direct access to
3 1 4 3
customers
Brand reputation 4 1 3 2

Vertical Axis (ratings, weights)

On a scale from 1 to 5, rate Segments on each factor, and weight the importance of each factor.

Segment
Attractiveness Segment 1 Segment 2 Segment 3 Weights
Size 3 5 1 2
Margins 3 5 1 5
Current
3 1 5 4
competition

Market Size
On a scale from 1 to 20, please enter market size for each item.

Segment 1 Segment 2 Segment 3

Market Size 6 10 2

© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-commercial educational
purposes only is authorized if this notice appears on every copy. 54
Illustration
Horizontal Axis (ratings, weights)
On a scale from 1 to 5, rate Segments on each factor, and weight the importance of each factor.
Competitive
Advantage Segment 1 Segment 2 Segment 3 Weights
Product fit 5 3 4 5
Direct access to
3 1 4 3 Specific to the firm doing
customers this exercise
Brand reputation 4 1 3 2

Vertical Axis (ratings, weights)

On a scale from 1 to 5, rate Segments on each factor, and weight the importance of each factor.
Segment
Attractiveness Segment 1 Segment 2 Segment 3 Weights
Size 3 5 1 2
But they might
Margins 3 5 1 5 disagree on their
Current relative importance
3 1 5 4
competition
Every firm competing in this market should agree on
these ratings

Market Size
On a scale from 1 to 20, please enter market size for each item.
Segment 1 Segment 2 Segment 3
Market Size 6 10 2

© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-commercial educational
purposes only is authorized if this notice appears on every copy. 55
Illustration
The management team of this
Horizontal Axis (ratings, weights)
company believes that the product
On a scale from 1 to 5, rate Segments on each factor, and weight the importance of each factor.
fit between their offerings and
Competitive
segment 1’s needs is very good.
Advantage Segment 1 Segment 2 Segment 3 Weights They also believe that, in this
Product fit 5 3 4 5 market, product fit is the most
Direct access to important driver of competitive
3 1 4 3 advantage.
customers
Brand reputation 4 1 3 2

Vertical Axis (ratings, weights)

Segments have widely differing
On a scale from 1 to 5, rate Segments on each factor, and weight the importance of each factor.
sizes, and in general, the bigger
Segment the segment, the more attractive it
Attractiveness Segment 1 Segment 2 Segment 3 Weights
is. For this company, in this
Size 3 5 1 2 market, however, size seems to be
Margins 3 5 1 5 much less important than
Current competitors’ strength.
3 1 5 4
competition

Market Size
On a scale from 1 to 20, please enter market size for each item. Watch out !
Segment 1 Segment 2 Segment 3 A higher number means “better”, not “more”. Here,
competition is very fierce on segment 2, hence the low
Market Size 6 10 2
number.

Segment 2
Segment Attractiveness

Segment 1

Segment 3

Competitive Advantage

Segment 2
Segment Attractiveness

Segment 1

Segment 3

Competitive Advantage

© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-commercial educational
purposes only is authorized if this notice appears on every copy. 58
Example Criteria for Determining Which
Segments to Serve
Criterion Examples of Considerations
I. Size and Growth
1. Size • Market potential, current market penetration
2. Growth • Past growth forecasts of technology change

II. Structural Characteristics

3. Competition • Barriers to entry, barriers to exit, position of competitors,
ability to retaliate
4. Segment saturation • Gaps in the market
5. Protectability • Patentability of products, barriers to entry
6. Environmental risk • Economic, political, and technological change

II. Product-Market Fit

7. Fit • Coherence with company’s strengths and image
8. Relationships with segments • Synergy, cost interactions, image transfers, cannibalization
• Entry costs, margin levels, return on investment
9. Profitability

© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-commercial educational
purposes only is authorized if this notice appears on every copy. 59
Selecting Segments Using GE/McKinsey
Portfolio Matrix
Competitive Strength
Low Average High
1.0 3.0 5.0
5.0

Attractive
Market Attractiveness
Neutral

3.0
Unattractive

1.0

Market Segment 1 Market Segment 4

Market Segment 2 Market Segment 5

Market Segment 3 Market Segment 6

= $10MM Sales

© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-commercial educational
purposes only is authorized if this notice appears on every copy. 60
Example Criteria for Segment Selection
Relative Market Attractiveness Analysis Relative Competitive Strength Analysis

1 = completely disagree; 2 = somewhat disagree; Weight Score Weighted 1 = completely disagree; 2 = somewhat Weight Score Weighted
3 = neutral; 4 = somewhat agree; 5 = completely (1 - 5) Score disagree; 3 = neutral; 4 = somewhat agree; 5 = (1 - 5) Score
agree completely agree

1. Large Served Available Market Size 0.00 1. Strongest Global Presence 0.00

2. Strong Market Segment Growth Potential 0.00 2. Largest Market Share 0.00

3. Profitable Market Segment Customers 0.00 3. Best-in-Class Product Quality 0.00

4. Weak Supply Base Negotiating Position 0.00 4. Powerful Brand Image 0.00

5. Weak Customer Negotiating Position 0.00 5. Premier Distribution Network 0.00

6. Absence of Strong Competition 0.00 6. Highly Effective MARCOM 0.00

7. Hard for Customers to Use Alternate Products 0.00 7. Right-Sized Production Capacity 0.00

8. Barriers to New Competitors Entering Market 0.00 8. Competitive Unit Cost 0.00

0.00 9. Superior Technical Depth and Breadth 0.00

0.00 10. Strongest Customer Relationships 0.00

Totals 0.00 0.00 Totals 0.00 0.00

© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-commercial educational
purposes only is authorized if this notice appears on every copy. 61
Segment Strategies Based on GE/McKinsey
Portfolio Analysis
Capitalize on Challenge for Top Build Preemptive
Attractiveness Position Position
•Selectively position to improve base •Invest for growth •Invest for growth
business •Build on major strengths •Diversify globally
•Create well-defined and rigidly •Protect areas of vulnerability •Consolidate position
enforced cut-off criteria
•Focus on long-term cash flow
•Monitor closely
(adaptation of Porter Five Forces Model)

Harvest Preserve Position Expand Selectively

Market Segment Attractiveness

•Reduce fixed costs •Make selective investments to •Seek most desirable areas of segment
•Minimize capital expenditures maintain position •Trade share for earnings as
•Reduce product line •Seek opportunistic growth appropriate
•Seek opportunistic sale •Apply strengths in more attractive •Monitor closely for further decline
areas

Exit Strict Cash Flow Protect/Refocus

Management
•Exit market or prune product line •Selectively invest for earnings
•Minimum commitment of resources
•Determine timing so as to present •Defend strengths
maximum value •Protect position in attractive areas
•Refocus to attractive segments
•No further commitment of resources •Monitor for harvest or divestment
timing

Competitive Strength (Available Levers)

© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-commercial educational
purposes only is authorized if this notice appears on every copy. 62
Segment Economics for Driving Segment
Selection
Average acquisition
cost Average Churn Average NPV* Key satisfaction
Segment Euro Percent Euros drivers

A 125 9 220
After-sale support

B 90 8 185 Product quality

C 85 11 57
Convenience

D 70 14 -36 Price

*NPV calculation assumes 5-year timeline, discount rate of 80% and average annual
contributions of €95, €75, €40 and €10, respectively
Source: McKinsey & Co. Report, October 2001

Choice-based segmentation is the process of categorizing

customers into groups (a.k.a. segments, clusters) based on their
response likelihood (to buy/not to buy, to chose brand A/B/C).

…Similarities (of needs, wants, past behavior) only matter to

the extent that they are predictive of customer choices.

• In need-based segmentation, we
carefully divide data between
segmentation and discriminant
Need-based segmentation
data, and build segments based on
the former only…

Segmentation
variables • …Because we don’t want to build
(bases) segments on the base of irrelevant
data
− E.g., demographics, so what?
Discriminant
variables
(descriptors)

• In a typical choice-based segmentation, the model decides what’s

relevant, and everything becomes segmentation data (to predict
likelihood of response)

Need-based segmentation Choice-based segmentation

Response

Segmentation
variables
(bases) Segmentation
variables
(predictors)
Discriminant
variables
(descriptors)

• CART (for classification and

regression tree) is one of the
most popular choice-based
segmentation tools

1. Select the predictors (segmentation variables)

• Past behavior
• Demographics
• …

2. Select the response, to be predicted/explained

• Buy or not, Donate or not (0/1)
• Brand choice (A, B, C)
• Purchase amount ($X)

3. Group all the respondents/customers into one big pool

• The “parent node”

4. For each available predictor, one by one, split the population (the parent node)
into subgroups (the child nodes), and check to what extent the child nodes are
• More homogeneous (within)
• More distinct (between)

5. Keep the split that works best

• Where “best” is usually measured by a statistical index, such as entropy, Gini index,
RMSE, etc.

6. Repeat for each child node

7. Stop when some criteria are met

• No further improvement,
• Not enough data to keep going

LET’S TRY TO PREDICT

THESE RESPONSES

• Predictors:
− Students’ profiles (analytical mind, etc.)
− Students’ specialization (finance, etc.)
− Current satisfaction (course, instructor)

• Response:
− Likelihood to take an advanced elective course (1..5)

Enjoy_Instructor <= 4 Finance <= 4 Marketing <= 2

19 enr. 31 enr. 22 enr.
Response : Response : Response :
moyenne 2.74 moyenne 3.29 moyenne 2.5
minimum 1 minimum 1 minimum 1
maximum 5 maximum 5 maximum 5
37 enr. 37 enr. 37 enr.
Response : Response : Response :
moyenne 3.11 moyenne 3.11 moyenne 3.11
minimum 1 minimum 1 minimum 1
maximum 5 Enjoy_Instructor > 4 maximum 5 Finance > 4 maximum 5 Marketing > 2
18 enr. 6 enr. 15 enr.
Response : Response : Response :
moyenne 3.5 moyenne 2.17 moyenne 4
minimum 2 minimum 1 minimum 3
maximum 5 maximum 3 maximum 5

• Satisfaction with • Intend to specialize • Intend to specialize

current instructor? in finance? in marketing?
• Poor predictor • Slightly better • Excellent predictor

5 segments
Enjoy_Instructor <= 4
8 enr.
Response :
moyenne 1.5
minimum 1
Strategic consulting <= 3 maximum 3
14 enr.
Response :
moyenne 2
Marketing <= 2 minimum 1 Enjoy_Instructor > 4
22 enr. maximum 4 6 enr.
Response : Response :
moyenne 2.5 moyenne 2.67
minimum 1 minimum 2
maximum 5 Strategic consulting > 3 maximum 4
8 enr.
Response :
moyenne 3.38
minimum 2
37 enr. maximum 5
Response :
moyenne 3.11
minimum 1
maximum 5
Marketing_Analytics <= 4
12 enr.
Response :
moyenne 3.83
minimum 3
Marketing > 2 maximum 5
15 enr.
Response :
moyenne 4
minimum 3
maximum 5 Marketing_Analytics > 4
3 enr.
Response :
moyenne 4.67
minimum 4
maximum 5

In CART, the predicted response can be…

• A binary decision
− Buy / Do not buy
− Donate / Do not donate
− Register / Do not register

• An integer or a real number

− Amount
− Likelihood of…

• A choice out of a set

− Brand A / B / C
− Choice A / B / C

Enjoy_MM <= 3
10 enr.
Course_Name :
Strategic Marketing Consulting : 7 70%
Response <= 4 Marketing Engineering : 3 30%
32 enr.
Course_Name :
Strategic Marketing Consulting : 13 40.6% Enjoy_MM > 3
37 enr.
Marketing Engineering : 6 18.8% 22 enr.
Course_Name :
Other : 5 15.6% Course_Name :
Strategic Marketing Consulting : 18 48.6%
Applied Marketing Analytics : 8 25% Strategic Marketing Consulting : 6 27.3%
Marketing Engineering : 6 16.2%
Other : 5 13.5% Marketing Engineering : 3 13.6%
Applied Marketing Analytics : 8 21.6% Other : 5 22.7%
Response > 4 Applied Marketing Analytics : 8 36.4%
5 enr.
Course_Name :
Strategic Marketing Consulting : 5 100%

• Replace “preferred course name” by “preferred brand”, and you’ll

have a perfect marketing application of CART
© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-commercial educational
purposes only is authorized if this notice appears on every copy. 75
Limitations of CART

• Each split reduces the size of the population left in the node
− If you want to go deep, you need a lot of data!
− Risk of overfitting the data
(i.e., finding by chance a relationship that does not exist)

• The method only select a subset of predictors

− Those that are not selected are the least important
− But they still have predictive value
− Yet, they are overlooked

• Within each node, response is still heterogeneous

A choice model is a mathematical model that predicts the

likelihood of an observed choice/response based on related
characteristics data (or predictors)

• Observed choice
− Buy / not buy (e.g., direct marketers)
− Brand bought (e.g., packaged goods)

• Predictors
− Demographics
− Attitudes, perceptions
− Market conditions (price, promotion, etc.)
− Past behavior, pattern of previous choices

• Link between the two

− The model predicts customers’ probabilities of purchase…
− …And in the process, reveals importance weights of predictors
(some might have little weight, hence being bad predictors)

• Predictor “Interest in marketing” (scale 1-5)

• Response “Take this elective course” (0 or 1)

100%

90%

80% 90%
86%
70% 71%

60%

50%

40%
33%
30%

20%
17%
10%

00%
1 2 3 4 5

(*) For this illustration, we’ve assumed that an answer of 1 to 3 is equal to “no”, and an answer of 4 or 5 is equal to “yes” .
The 86% figure here means that 86% of the respondents who answered “4” to the marketing question, answered either
“4” or “5” to the target question.

1
response 
1 e   weightscharacteristics 


Observed Inferred Observed

(0/1) to provide (including an intercept)
best fit

1
Course     2.9  1.19 Marketing
1 e

100%
90% • This term is positive and
80%
significant
70%
60%
50% • The higher the student’s
40% interest in marketing…
30%
20%
10%
• …the more likely he/she
00%
is going to take the
1 2 3 4 5 elective course

Observed Predicted

© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-commercial educational
purposes only is authorized if this notice appears on every copy. 82
Observations / Choice Strategic_Co Entrepreneu Analytical_ Marketing_A Enjoy_Cours Enjoy_Instru
Choice (0/1) Marketing Finance Graduation
data nsulting r Mind nalytics e ctor
Enter id 1 here 1 5 1 2 4 5 5 5 5 5

Data
Enter id 2 here 1 5 1 5 2 4 5 1 5 5
Enter id 3 here 0 1 4 1 5 4 4 5 4 4
Enter id 4 here 1 2 2 4 2 4 2 3 4 4
Enter id 5 here 0 1 5 2 4 5 3 4 4 5
Enter id 6 here 0 1 2 4 5 3 4 3 4 4
Enter id 7 here 1 3 2 5 3 4 3 5 4 4
Enter id 8 here 0 1 1 1 3 3 1 3 1 4
Enter id 9 here 0 1 4 4 2 4 3 4 4 5
Enter id 10 here 0 2 4 1 4 4 4 3 4 5
Enter id 11 here 0 1 5 3 2 4 3 1 2 2
Enter id 12 here 0 1 5 5 4 4 4 2 4 4
Enter id 13 here 0 1 4 3 3 3 2 4 4 5
Enter id 14 here 0 1 4 1 3 4 2 5 4 4
Enter id 15 here 0 1 3 2 1 3 4 3 3 5
Enter id 16 here 0 3 5 4 2 4 4 1 3 4
Enter id 17 here 0 1 1 3 5 4 4 5 4 4
Enter id 18 here 1 5 1 3 4 2 3 4 4 4
Enter id 19 here 0 2 5 4 2 5 4 2 5 5
Enter id 20 here 1 4 1 1 1 2 3 3 4 4
Enter id 21 here 1 3 4 4 2 2 3 4 4 5

• 306 respondents
Enter id 22 here 0 3 1 3 2 4 3 3 5 5
Enter id 23 here 0 2 1 1 2 4 2 4 2 4
Enter id 24 here 0 1 1 1 4 1 2 3 4 5
Enter id 25 here 1 2 4 5 5 5 3 5 4 5
Enter id 26 here 0 2 2 5 1 2 3 3 3 4
Enter id 27 here 0 4 1 4 3 4 4 3 4 5
Enter id 28 here 1 5 1 3 1 4 4 3 4 4
Enter id 29 here 0 1 5 1 5 5 3 4 2 3

• 1 choice Enter id 30 here

Enter id 31 here
Enter id 32 here
0
1
1
2
3
4
2
1
1
2
3
3
1
1
4
3
4
3
2
4
1
5
4
5
3
3
5
4
3
5
Enter id 33 here 1 1 1 1 5 5 4 4 2 5
Enter id 34 here 1 1 3 5 5 4 3 3 5 5
Enter id 35 here 1 5 1 4 2 5 3 5 4 5
Enter id 36 here 1 4 1 1 5 5 5 1 5 5

• 9 predictors
Enter id 37 here 1 3 2 2 5 5 2 5 4 4
Enter id 38 here 1 4 1 3 3 2 4 4 5 5
Enter id 39 here 1 4 1 4 4 4 2 3 4 4
Enter id 40 here 0 5 3 4 4 4 5 3 4 4
Enter id 41 here 1 4 2 3 1 4 5 3 3 4
Enter id 42 here 1 4 2 3 1 4 5 3 3 4
Enter id 43 here 0 1 4 4 1 4 3 4 4 4
Enter id 44 here 1 5 2 4 1 4 4 2 5 4
Enter id 45 here 0 1 4 4 3 3 2 4 3 3
Enter id 46 here 1 4 1 5 4 4 2 4 5 5
Enter id 47 here 1 3 3 5 4 5 3 3 3 5
Enter id 48 here 1 4 1 3 3 4 5 1 4 4
Enter id 49 here 1 1 5 3 5 4 3 5 4 4
Enter id 50 here 1 2 4 4 2 5 5 1 3 4
Enter id 51 here 0 2 1 3 3 4 3 1 4 4

© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-commercial educational
purposes only is authorized if this notice appears on every copy. 83
Results
Coefficient Estimates
Coefficient estimates of the Choice model. Coefficients in bold are statistically significant.
Variables / Coefficient estimates Coefficient estimates Standard deviation t-statistic
Marketing 1.19 0.20 6.08
Finance -0.30 0.13 -2.25
Strategic_Consulting 0.41 0.14 2.88
Entrepreneur 0.24 0.15 1.64
Analytical_Mind 0.51 0.22 2.35
Marketing_Analytics -0.06 0.21 -0.29
Graduation 0.40 0.16 2.49
Enjoy_Course -0.13 0.24 -0.53
Enjoy_Instructor 0.12 0.33 0.37
Const-1 -7.28 1.74 -4.17

• The most important drivers of choice are:

− Interest in marketing
− Analytical mind
− Strategic consulting
− Close to graduation
− Interest in finance (-)
© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-commercial educational
purposes only is authorized if this notice appears on every copy. 84
Provides a good fit?
Confusion Matrix on Estimation Sample
Comparison of observed choices and predicted choices (based on MNL analysis).
High values in the diagonal of the confusion matrix (in bold), compared to the non-diagonal values, indicate high convergence between
observations and predictions.
Analysis has been performed on the estimation dataset, and measures the goodness-of-fit of the model.

Observed / Predicted Choice Response Dummy

Response 114 18
Dummy 42 132

• 80% of respondents are well classified

− 114 correct “yes” + 132 correct “no”, over 306 observations
− Very good !

• There are more “false positive” then “false negative”

− 42 predicted as “yes”, but indeed “no”
− 18 predicted as “no”, but indeed “yes”
− Are we missing something?

• Not only used to predict a yes/no choice

• But also to predict a one-out-of-many choice

− Brand, product, option…
− Same logic, slightly more complicated formulation

− σ 𝑤𝑒𝑖𝑔ℎ𝑡𝑠×𝑐ℎ𝑎𝑟𝑎𝑐𝑡𝑒𝑟𝑖𝑠𝑡𝑖𝑐𝑠
𝑒 Brand A
𝑝Brand A =
− σ 𝑤𝑒𝑖𝑔ℎ𝑡𝑠×𝑐ℎ𝑎𝑟𝑎𝑐𝑡𝑒𝑟𝑖𝑠𝑡𝑖𝑐𝑠
σ𝑒 All Brands
All brands

Targeting

1. Calibrate model coefficients on a sample of customers

• Different customers (test)
• Same customers, different period in time

2. Apply choice model to a larger list of customers

• Predict likelihood of choice
• Rank order, from least likely to most likely

3. Target customers based on predictions

© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-commercial educational
purposes only is authorized if this notice appears on every copy. 87
Who to target?
Respondents / Choice Response
probabilities probability
Customer 15 0.993
Customer 26 0.976
Customer 20 0.964
Customer 19 0.947
Customer 12 0.944
Customer 8 0.943
Customer 7 0.934
Customer 24
Customer 27
0.917
0.901
• A list of 30 customers, with their respective
Customer 17
Customer 5
0.893
0.837
likelihood of response
Customer 18 0.816
Customer 16 0.804
Customer 21 0.801
Customer 22 0.801
Customer 28
Customer 11
0.778
0.698
• Which ones to target?
Customer 2 0.675
Customer 1 0.442
Customer 13 0.374
Customer 3 0.362
Customer 14
Customer 10
0.335
0.283
• It depends on the purpose of the action
Customer 6 0.259
Customer 30 0.229
Customer 29 0.214
Customer 9 0.132
Customer 25 0.128
Customer 23 0.123
Customer 4 0.035

© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-commercial educational
purposes only is authorized if this notice appears on every copy. 88
Goal #1: elicit a purchase
Respondents / Choice Response
probabilities probability
Customer 15 0.993
Customer 26 0.976

If the goal is to:

Customer 20 0.964
Customer 19 0.947
Customer 12 0.944
Customer 8
Customer 7
0.943
0.934 • Send a catalogue to ask for an order
Customer 24 0.917
Customer 27
Customer 17
0.901
0.893
• Place a call to elicit a sale
Customer 5
Customer 18
0.837
0.816 • Send a direct mail to ask for donation
Customer 16 0.804
Customer 21
Customer 22
0.801
0.801
• …
Customer 28 0.778
Customer 11 0.698
Customer 2 0.675
Customer 1
Customer 13
0.442
0.374 Then the customers who are most likely to answer
Customer 3
Customer 14
0.362
0.335
are the primary targets
Customer 10 0.283
Customer 6 0.259
Customer 30 0.229
Customer 29
Customer 9
0.214
0.132
MAXIMIZE PROFITABILITY
Customer 25 0.128
Customer 23 0.123
Customer 4 0.035

© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-commercial educational
purposes only is authorized if this notice appears on every copy. 89
Goal #2: influence behavior
Respondents / Choice Response
probabilities probability
Customer 15 0.993
Customer 26 0.976
Customer 20 0.964
Customer 19
Customer 12
0.947
0.944 If the goal is to:
Customer 8 0.943
Customer 7
Customer 24
0.934
0.917
• Send a coupon to ease trial/purchase
• Change, modify perceptions
Customer 27 0.901
Customer 17 0.893
Customer 5 0.837

• …
Customer 18 0.816
Customer 16 0.804
Customer 21 0.801
Customer 22 0.801
Customer 28 0.778
Customer 11 0.698
Customer 2
Customer 1
0.675
0.442
Then the customers who are potential
Customer 13
Customer 3
0.374
0.362
switchers are the primary targets
Customer 14 0.335
Customer 10 0.283
Customer 6 0.259

MAXIMIZE MARKETING IMPACT

Customer 30 0.229
Customer 29 0.214
Customer 9 0.132
Customer 25 0.128
Customer 23 0.123
Customer 4 0.035

© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-commercial educational
purposes only is authorized if this notice appears on every copy. 90
Why?
Respondents / Choice Response
probabilities probability
Customer 15 0.993
Customer 26
Customer 20
0.976
0.964
• These customers will be loyal anyway
Customer 19
Customer 12
0.947
0.944 • No need to send coupon
Customer 8 0.943
Customer 7
Customer 24
0.934
0.917
• No need to “convince” them
• They are already convinced
Customer 27 0.901
Customer 17 0.893
Customer 5 0.837
Customer 18 0.816
Customer 16 0.804
Customer 21 0.801
Customer 22 0.801
Customer 28 0.778
Customer 11 0.698
Customer 2 0.675
Customer 1 0.442
Customer 13 0.374
Customer 3 0.362
Customer 14 0.335
Customer 10 0.283
Customer 6 0.259
Customer 30
Customer 29
0.229
0.214
• These customers will never buy from you
Customer 9
Customer 25
0.132
0.128 • Don’t waste valuable time and resources
Customer 23 0.123
Customer 4 0.035

• The first derivative of the Logit function gives us the marginal impact
of a change in a variable (e.g., marketing actions)
• When probability of choices is near 50%, impact of marketing
actions is maximized

1 3

0.8 2.5
2
0.6
1.5
0.4
1
0.2 0.5
0 0
-5 -4 -3 -2 -1 0 1 2 3 4 5 -5 -4 -3 -2 -1 0 1 2 3 4 5
LOGIT FUNCTION LOGIT’S FIRST DERIVATIVE

A score is a characterization of a customer’s most likely

behavior or potential.

In the most simple case, a score is simply the result of a choice

model (e.g., likelihood of purchase).

In more complex cases, a score can be the result of a

combination of choice and predictive models

Predicted
Likelihood donation
of donation Score
Fundraising context: amount
Donor 1 34% € 153 € 52.7
Donor 2 48% € 61 € 29.0
• Likelihood of donation Donor 3 38% € 36 € 13.8

(choice model) Donor 4

Donor 5
7%
71%
€ 185
€ 19
€ 13.7
€ 13.5
Donor 6 33% € 34 € 11.2
Donor 7 26% € 42 € 11.0
• Predicted donation amount Donor 8 5% € 58 € 2.8
in case of donation Donor 9 5% € 52 € 2.6
(e.g., regression analysis) Donor 10 3% € 83 € 2.2
Donor 11 2% € 67 € 1.6
Donor 12 2% € 68 € 1.6
• Score = combination of both Donor 13 2% € 60 € 1.4
(expected donation amount) Donor 14 1% € 124 € 1.4
Donor 15 5% € 24 € 1.3
Donor 16 3% € 45 € 1.2
Donor 17 1% € 130 € 1.0
Two donors can have similar scores Donor 18 1% € 73 € 0.6
with different underlying behavior (*) Donor 19 1% € 20 € 0.3
Donor 20 2% € 14 € 0.2

A score class is a grouping of customers whose scores fall within a

given range.

Predicted
Likelihood donation
of donation Score
amount
Donor 1 34% € 153 € 52.7
A
Donor 2 48% € 61 € 29.0
Donor 3 38% € 36 € 13.8
Donor 4 7% € 185 € 13.7
Donor 5 71% € 19 € 13.5 B
Donor 6 33% € 34 € 11.2
Donor 7 26% € 42 € 11.0
Donor 8 5% € 58 € 2.8
Donor 9 5% € 52 € 2.6 • Donors are grouped into 4
C
Donor 10
Donor 11
3%
2%
€ 83
€ 67
€ 2.2
€ 1.6 classes: A, B, C, D
Donor 12 2% € 68 € 1.6
Donor 13 2% € 60 € 1.4
Donor 14 1% € 124 € 1.4
Donor 15 5% € 24 € 1.3
Donor 16 3% € 45 € 1.2
Donor 17 1% € 130 € 1.0 D
Donor 18 1% € 73 € 0.6
Donor 19 1% € 20 € 0.3
Donor 20 2% € 14 € 0.2

© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-commercial educational
purposes only is authorized if this notice appears on every copy. 96
Score classes vs. choice-based segmentation
Enjoy_Instructor <= 4
8 enr.
Response :
moyenne 1.5

Similarities:
minimum 1
Strategic consulting <= 3 maximum 3
14 enr.
Response :
moyenne 2
Marketing <= 2 minimum 1 Enjoy_Instructor > 4
22 enr. maximum 4 6 enr.
Response : Response :
moyenne 2.5 moyenne 2.67

• They group customers based on likelihood of choices

minimum 1 minimum 2
maximum 5 Strategic consulting > 3 maximum 4
8 enr.
Response :
moyenne 3.38
minimum 2
37 enr. maximum 5
Response :
moyenne 3.11

• They are simple to use (e.g., a few segments/classes)

minimum 1
maximum 5
Marketing_Analytics <= 4
12 enr.
Response :
moyenne 3.83
minimum 3
Marketing > 2 maximum 5
15 enr.
Response :
moyenne 4
minimum 3
maximum 5 Marketing_Analytics > 4
3 enr.
Response :
moyenne 4.67
minimum 4
maximum 5

Differences:
• Score classes are more complex…
Predicted
Likelihood
donation Score
of donation
amount
Donor 1 34% € 153 € 52.7
A
− They need regular updates of response models
Donor 2 48% € 61 € 29.0
Donor 3 38% € 36 € 13.8
Donor 4 7% € 185 € 13.7
B
− They are more complete, more accurate,
Donor 5 71% € 19 € 13.5
Donor 6 33% € 34 € 11.2
Donor 7 26% € 42 € 11.0

they use all data available Donor 8

Donor 9
Donor 10
5%
5%
3%
€ 58
€ 52
€ 83
€ 2.8
€ 2.6
€ 2.2 C
− They are harder to “get” intuitively
Donor 11 2% € 67 € 1.6
Donor 12 2% € 68 € 1.6
Donor 13 2% € 60 € 1.4

(who’s in score class “A”, and why?) Donor 14

Donor 15
Donor 16
1%
5%
3%
€ 124
€ 24
€ 45
€ 1.4
€ 1.3
€ 1.2
Donor 17 1% € 130 € 1.0 D
Donor 18 1% € 73 € 0.6
Donor 19 1% € 20 € 0.3
Donor 20 2% € 14 € 0.2

CLASSIC TRADE-OFF --- ACCURACY/PERFORMANCE vs. SIMPLICITY/USABILITY

Need-based Choice-based

• Customers are grouped together • Customers are grouped together

based on their similarities in based on their (predicted)
profiles
− Needs similarities in future behavior
− Wants – Likelihood of donation, of purchase, of
− Lifestyles choice
− Past behavior
80% of customers in segment “X”
Customers in segment “X” highly value are expected to select our premium
prestige and peace of mind, are not highly offering.
price-sensitive, and many of them have
been loyal customers for more than 3 years.

THE MARKET = ONE SEGMENT

Mass marketing

FEW SEGMENTS (3~6)

High-level view of the market, used to define company-wide strategy

MORE SEGMENTS (10~30)

More-detailed view of the market, to target specific
groups, customize campaigns, optimize operations

ONE CUSTOMER = ONE SEGMENT

One-to-one marketing

THE MARKET = ONE SEGMENT

Average response of the market

MORE OR LESS COMPLEX

CHOICE_BASED SEGMENTATION
(3~30 segments)
SCORING
CLASSES

CHOICE-BASED MODELS
Individual predictions and scores

• A large charity sends a direct mail solicitation to its donors for its
Christmas campaign

• A few key figures:

− Mails sent 301 500
− Donations 17 200
− Return rate 5.7%
− Total donations 992 000 €
− Average donation amount 57.7 €
− Mailing costs 182 500 €
− Net margins 809 500 €
− Return on investment + 443%
− Fundraising ratio 18.4%
(e.g., needs 18 cents to collect 1 €)

They used several choice models (scoring)

• Responses:
− Likelihood of donation
− Donation amount

• Predictors:
− Recency
− Frequency
− Amount
− Activity over the years
− Demographics
− …

• They built a score…

Score = Likelihood of donation  Donation amount

• And ranked all their donors by decreasing order

What would have been the financial results, had we only

solicited the top [X]% of our donors?

(e.g., the 10% of our donors who received the highest scores,
the top-50%, etc.)

Collecte brute Couts Collecte nette Ratio de collecte

1,000,000.00 € -30%
900,000.00 €
800,000.00 € -25%

700,000.00 €
-20%
600,000.00 €
500,000.00 € -15%
400,000.00 €
300,000.00 € -10%

200,000.00 €
-5%
100,000.00 €
0.00 € -0%
0 1 Donateurs sollicities
2 (de O a 301,500)
3 4 5

HIGHEST-SCORE DONORS LOWEST-SCORE DONORS

Collecte brute Couts Collecte nette Ratio de collecte

1,000,000.00 € -30%

900,000.00 €
-25%
800,000.00 €

700,000.00 €
-20%
600,000.00 € 20% of the donors
= 76% of donations
500,000.00 € -15%
= 86% of net margins
400,000.00 €
-10%
300,000.00 €

200,000.00 €
-5%
100,000.00 €

0.00 € -0%
0 1 Donateurs sollicities
2 (de O a 301,500)
3 4 5

Collecte brute Couts Collecte nette Ratio de collecte

1,000,000.00 € -30%

900,000.00 €
-25%
800,000.00 €

700,000.00 €
-20%
600,000.00 € Mails sent 301,500
Donations 17,200
500,000.00 € -15%
Return rate 5.7%

400,000.00 € Total donations € 992,000

Average donation amount € 57.7 -10%
300,000.00 € Mailing costs € 182,500
Net margins € 809,500
200,000.00 €
Return on investment 443% -5%
100,000.00 € Fundraising ratio 18.4%

0.00 € -0%
0 1 Donateurs sollicities
2 (de O a 301,500)
3 4 5

Collecte brute Couts Collecte nette Ratio de collecte

1,000,000.00 € -30%

900,000.00 €
-25%
800,000.00 €

700,000.00 €
-20%
600,000.00 € Mails sent 234,450 301,500
Donations 16,070 17,200
500,000.00 € -15%
Return rate 6.9% 5.7%
400,000.00 € Total donations € 968,400 € 992,000
Average donation amount € 60.3 € 57.7 -10%
300,000.00 €
Mailing costs € 148,700 € 182,500

200,000.00 € Net margins € 819,700 € 809,500

-5%
Return on investment 551% 443%
100,000.00 € Fundraising ratio 15.4% 18.4%

0.00 € -0%
0 1 Donateurs sollicities
2 (de O a 301,500)
3 4 5

Collecte brute Couts Collecte nette Ratio de collecte

1,000,000.00 € -30%

900,000.00 €
-25%
800,000.00 €

700,000.00 €
-20%
600,000.00 € Mails sent 154,200 234,450 301,500
Donations 14,400 16,070 17,200 -15%
500,000.00 €
Return rate 9.3% 6.9% 5.7%
400,000.00 € Total donations € 912,500 € 968,400 € 992,000
Average donation amount € 63.4 € 60.3 € 57.7 -10%
300,000.00 €
Mailing costs € 102,959 € 148,700 € 182,500
200,000.00 € Net margins € 809,541 € 819,700 € 809,500
-5%
Return on investment 786% 551% 443%
100,000.00 € 11.3% 15.4% 18.4%
Fundraising ratio
0.00 € -0%
0 1 Donateurs sollicities
2 (de O a 301,500)
3 4 5

Depending on the managerial objectives, this charity could:

• Collect more with less

− + 10,000 €
− Improve fundraising ratio from 18.4% to 15.4%

• Dramatically improve financial performance

− Same net margins
− Improve fundraising ratio from 18.4% to 11.3%
− Improve ROI almost twofold
− Save 80 000 € in costs

ILLUSTRATION 2

Voters do not wear uniforms!

Most elections are won by a few percentage points only.

How to identify potential swing voters, and target them with

the “right” message?

• 140 questions about deep needs, beliefs and values

− What do you envision for the future?
(fear, hopes, safety, etc.)

− How do you define success?

(family, financial security, community respect, business success)
− Skills and assets you believe you will need to achieve success?

− “Taking care of the country’s children should be our #1 priority”

− “Country should do whatever it takes to protect the environment”

− Etc.

• Which party they have voted for in the last elections

• Found 5 segments
− “Extending opportunities to others” (37%)
− “Working within the community”
− “Achieving independence”
− “Focusing on family”
− “Defending righteousness” (16%)

• Further segmented into 10 “tribes”

• Extending opportunities to others”

− Keyword for Democrats. No need to “convince” them

• “Defending righteousness”
− Pure-blood Republicans. Don’t bother

• Potential swing voters:

− “Working within the community”
− “Achieving independence”
− “Focusing on family”

• Bought discriminant data (descriptors) from various commercial database

companies

• ChoicePoint
− Tax records, court rulings, birth and death records
− Used for checking resumes, loan, credit card, etc.

• Acxiom
− Shopping and lifestyle data on 200,000,000 Americans
− Value of their house, magazines subscribed, books bought, etc.
− Buy all data available, then sold it to those who want to target us

• You are more likely to be a hard core Democrat if…

− You have a cat
− You eat sushis
− You have an Apple computer
− …

• You are more likely to be a hard core Republican if…

− You have a dog
− You drive a Pontiac
− You own a gun
− You go to church
− …

• Go to Acxiom
• Buy addresses of 20,000 people in that region who are likely to belong to the “Working within the
community” segment
• Send a specific direct mail talking about the Democrat promise to help local groups working with
the community

Or…

• Check what TV programs the “Focusing on family” segment members watch, and when, in that
region (e.g., games, TV shows)
• Buy TV spots during these programs
• Advertise plans for children’s education, family welfare…

Not Action Plans …

Little Measurable Value from many

segmentation studies ……
They are one-time Projects that
drain resources

• Instinctively, firms think about target market segments that are:

− Easily defined
− Clear-cut
− And reachable . . .

Seg A
• In practice, market segments are . . .
− Hard to define
− Fuzzy, and
− Overlapping
Overlap

Seg B Seg C

• And, customer needs evolve over time.

Ad in London Newspapers, 1913*

Men wanted for hazardous journey. Small wages,

bitter cold, long months of complete darkness,
constant danger, safe return doubtful. Honor and
recognition in case of success.
— Ernest Shackleton,

Did it work? Absolutely yes!

* It is not entirely clear whether Ernest Shackleton actually placed such an ad. At worst, this makes for an
interesting apocryphal story.

• Identifies segments of customers with differentiated needs.

− How many different segments?
− How do their needs differ?

• Enables segments to be separately targeted/reached (this can be problematic

even if segments have distinctly different needs).

• Finds one or more attractive segments (i.e. a profitable and separate marketing
program can be designed for selected segments).

• Facilitates the implementation of the segmentation scheme as an ongoing

process, not a discrete project.

Barrier Solution
1. The blind man syndrome 1. Prototype sessions
2. Bad history 2. Diagnosis and prescription
3. Salesforce resistance 3. Salesforce involvement
4. Integrating database marketing 4. Value added
5. Project positioning 5. Position as a process or a program

Barrier
1. Wrong segmentation unit
Solution
2. Multiple purchase influences 1. Use the buying location
2. Key informant or integration
3. Segmentation results conflict with
strategy 3. Pre-specify/Constrain the analysis
4. Difficulty in reaching segment 4. Design in a targeting process
members

Barrier Solution
1. Lack of acceptance amongst key use 1. Involvement and education
communities
2. No implementation process 2. Blueprint development
3. No performance metrics 3. Develop performance metrics
4. No continuous 4. Integrate an improvement / updating
improvement/updating process process into the segmentation
5. Reorganizations, mergers and 5. Document success and make it a high
acquisitions ROI investment

6. The ten commandments syndrome 6. A balanced, flexible view

• Choice-Based Segmentation (Chapter 2). Compare with Needs-

Based Segmentation.
• Latent class methods
− Segment based on unobservable characteristics (e.g., price sensitivity).
− Two approaches: (1) Latent cluster analysis, (2) Latent regression analysis.
• Bayesian classification
− Use (updatable) prior knowledge to improve classification accuracy.

• Consumer packaged goods

• Pharmaceuticals
• Banking and insurance
• Services

In summary,

• Use needs variables to segment markets.

• Select segments taking into account both the attractiveness of

segments and the strengths of the firm.

• Use descriptor variables to develop a marketing plan to reach and

serve chosen segments.

• Develop mechanisms to implement the segmentation strategy on a

routine basis (one way to do this is through information technology).

1. Positioning concept
Material designed to accompany the book Principles of Marketing Engineering
and Analytics by Lilien, Rangaswamy and De Bruyn, and the marketing analytics
software Enginius available at http://www.enginius.biz
2. Perceptual mapping
3. Preference mapping

publication is not authorized. Copying and distribution for non-commercial
educational purposes only is authorized if this notice appears on every copy.
© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-
commercial educational purposes only is authorized if this notice appears on every copy. 1
Key Concepts

• Differentiation (What you do to an offering): Creating

tangible or intangible differences on one or more
attributes between a focal offering and its main
competitors.
• Positioning (What you try to do to the minds of
customers): A set of strategies a firm develops to
differentiate its offering in the minds of its target
customers. Successful positioning will result in the
offering occupying a distinct, important, and sustainable
position in the minds of the target customers.

• Unique (Only supplier offering true, four color packaging)

• Superior (Relative to competition, we deliver 20% more
on time delivery)
• Equal but cheaper (commodity) strategy (We will meet
or beat any competitor’s price)

For [target segment], the [product/concept] is [most important

claim] because [single most important support].
Iomega
For [PC Users], the [IOMEGA Zip drive] is the [best portable
storage device] because [it is most cost-effective system].
JC Penney
For [Modern Spenders and Starting-outs in mid-income levels
who shop for apparel, accessories, and home furnishings] we
offer [private-label, supplier exclusive, and national brands] that
[deliver greater value than that of our competitors] because of
[our unique combination of quality, selection, fashion, service,
price, and shopping experience]. (From jcpenney.com)

Pantene
• For [females 18-49 who possess dry damaged hair and
believe they cannot achieve truly healthy/shiny hair] Pantene
is a [hair care system (shampoo/conditioner/ styling aids)] that
offers [“hair so healthy it shines”] because it [“penetrates from
root to tip” through its patented Pro-Vitamin B5 formula].
Microsoft .NET
• For [companies whose employees and partners need timely
information], Microsoft.NET is a [new protocol and software
system] that enables [unprecedented levels of software
integration through XML Web services], because [unlike Java,
.NET is infused into the Microsoft platform, providing the
ability to quickly and reliably build, host, deploy, and utilize
connected applications].

• Mapping: Techniques that enable managers to develop

differentiation and positioning strategies by helping them
to visualize the competitive structure of their markets as
perceived by their customers. The maps are derived
from data of customer perceptions of existing products
(and new concepts) along various attributes, perceptions
of similarities between brands, preferences for the
products, or measures of behavioral response of
customers toward the products.

Consumers face increasing number of choices

• Over 9,000 mutual funds

• Over 500 different make-models of cars
• Over 30,000 products in a typical grocery store
• Over 100,000 prescription drugs
If you can figure out how to stand out in the crowd, you
will have a better handle on your future profitability.

Perceptual Data
Average score each brand achieves on each attribute from your sample of respondents.

Christian Calvin Princesse

Attributes / Brands Aubade Darjeeling Dim Esprit Etam Playtex Wonderbra
Dior Klein Tam-Tam
Esthetism 5.9 5.4 4.3 4.3 3.7 3.8 3.8 3.7 4.9 3.5
Comfort 4.9 4.5 5.1 4.4 5.5 4.6 4.6 4.7 4.8 3.7
Price 5.8 6.9 5.8 4.7 3.1 4.2 2.7 3.6 4.5 4.7
Originality 5.5 5.3 3.7 4.0 3.2 3.2 3.4 2.6 4.7 3.1
Sexy 6.1 5.3 3.4 4.7 2.9 3.2 3.8 2.4 4.6 4.3
Redefine body 4.1 3.3 2.6 3.1 3.1 2.7 2.8 3.4 3.3 5.3
Special 3.8 3.4 2.4 2.8 2.4 2.5 2.4 2.3 3.2 3.5
Adequation 4.1 2.3 3.9 3.7 4.4 4.0 4.8 2.4 4.6 2.4
Seduction 5.7 4.2 3.7 4.0 3.2 3.4 3.7 2.4 4.7 3.7
Attachement to brand 3.9 2.5 2.8 3.0 3.4 2.7 3.2 1.9 3.8 1.9
Purchase frequency 2.3 1.1 1.8 2.2 3.4 1.9 3.6 1.6 3.0 1.3

• Perceptual study about underwear brands

• Average perception scores
• Sample: 19-25 year-old women
Crédits: Clotilde Parisis, Florence Sallé, Stéphanie Weill-Hébert © ESSEC Business School

• Place “competitors” on a map

– Products
– Companies
– Offerings

• Position them along relevant

dimensions
(relevant to the customers)

• Short distance =
fierce competition

• Easy to read, analyze

• Graphical
• 2D map
• Human beings are good with
graphical representations
• Much better than with raw
numbers

• Easy to communicate, to
convince, to act
• Ah-ah! factor

Challengers Leaders

SAP
Information Builders
Cognos

Ability to execute
Microsoft
Business Objects
• So powerful, it’s magic… Oracle
SAS Institute

Hyperion Solutions
Actuate MicroStrategy

Source: Gartner, Business Intelligence Siebel Systems

Niche Players Visionaries

Completeness of vision

As of December 2005
Source: Gartner

• What if 2 dimensions are not enough?

Good advertising Innovation brand

High quality
Well-known brand LAVAZZA blend
SPLENDID Good taste
Wide range
Sales promotion
Familiar brand Elegant
Creamy
packaging
KIMBO Value for price
Strong flavour
For the whole family

Low price
Neapolitan roasting Gift
SEGAFREDO
SAO
Convenient
packaging Arabic blend
MAURO Premium price
Like the coffee bar
Robusta blend
Bitter taste ILLY
Specialized
Local brand brand

Source: Kraft Jacobs Suchard “Marketing food brands in Italy: a case study approach,” Carlo Alberto Pratesi,
British Food Journal (2002), 104(6), pp.413:464

• What if 2 dimensions are not enough?

• What data to collect? Is it hard to do?

• How to link these perceptions to preferences and market

shares?

• When I say SAFETY you think…

“We’ve been saving people in car accidents

for more than 80 years” (Volvo website)

Volvo
• GM first to perform barrier crash test (1934)
• SAAB first to produce cars with a safety cage (1949)
• FORD first to build a fleet of cars with airbags (1971)
• For the first rollover risk of new cars and SUVs (2004),
only the MAZDA RX-8 got a 5-star rating
• Yet…

VOLVO = SAFETY

Drilling company

• A mining and drilling

company
• They came up with a
breakthrough technology
• Faster
• Cheaper

• Yet, it did not sell, until…

Drilling company

FAST
5

This is how the company 4

saw its market…
3
• They were faster and cheaper
then competition 2

• They were “dominating”

1
• Yet, sales were not impressive…
0
0 1 2 3
INEXPENSIVE

Drilling company

This is how their customers

saw the market…
• Their positioning
was not credible
FAST
? INEXPENSIVE

• Their product
became very successful.
once they increased price

Débitel
• Débitel = French, low-cost “reseller” or “aggregator”
– Cell phone company, with no network
– Buys air time from SFR, #1 network in quality
– Like Immix Wireless in the USA

• Perceptions of “network quality”

– SFR = 8.0 / 10
– Débitel = 5.9 / 10
– Same network

• Débitel purchased by SFR, now Simpleo (2008)

Cat litter

• Cat litter
• White vs. blue grains
• Just plain clay
– Painted in blue
– No chemical properties
– No perfume

• Focus
• Clarity
• Brevity
• Coherence

• Wishful marketing = a death wish

– Wishful marketing = Communicate what you’d like to be, rather
than what you are

• Shaping market perceptions = an expensive death wish

• If customers’ attention span is short…

• If you need to convey a highly focused message…
• If their perceptions matter more than your reality…

It is critical to map customers’ mind

• How you are perceived
• How your competitors are perceived
• On which dimensions
• Which dimensions matter

Human brain is not very good at visualizing numbers

• Need a method to represent graphically complex

patterns

Correspondence Analysis
(a.k.a. Perceptual Mapping, Multi-Dimensional Scaling,
Factor Analysis, Principal Component Analysis)

Correspondence analysis allows for the reduction of an

important number of variables (product attributes,
perceptions) into sub-sets of key dimensions (linear
combination of the original variables) by analyzing their
correlations

Perceptual Data
Average score each brand achieves on each attribute from your sample of respondents.

Christian Calvin Princesse

• Perceptual study about underwear brands

• Average perception scores
• Sample: 19-25 year-old women
Crédits: Clotilde Parisis, Florence Sallé, Stéphanie Weill-Hébert © ESSEC Business Schhol

Brands that are close on the

map are close competitors Wonderbra

Customers have similar

perceptions about these

Dimension II
brands
These brands compete on the Aubade

Etam
same dimensions. But which Dim

ones?

(e.g., Dim and Etam are close competitors)

Dimension I

Attribute lines show

along which key Wonderbra
dimensions brands are Redefine body
positioned

Dimension II
Aubade

Etam Seduction
Dim

Comfort

(e.g., Dim and Etam are perceived as very Dimension I

comfortable)

Highly correlated Somewhat correlated

Uncorrelated Negatively correlated

Inexpensive Expensive

How to compare
brands along one Wonderbra

dimension? Redefine body

Dimension II
• Expand dimensions
• Get orthogonal Aubade
Seduction
projections Dim
Etam

Comfort

(e.g., Dim and Etam most comfortable, Aubade Dimension I

in the middle, Wonderbra least comfortable)

Product
quality A
C

Service
quality

Product
quality

A
C

Service
quality

Interpreting the map

Wonderbra

Redefine body
• Highly correlated dimensions Playtex Christian DiorPrice
go toward identical directions Special

(e.g., price, sexy, seduction, special)

Dimension II
Esprit Darjeeling Sexy
Calvin Klein

• Negatively correlated Aubade Esthetism

Originality

dimensions go in opposite
Seduction
Etam Princesse Tam-Tam
Dim

directions (e.g., redefine body, comfort)

Comfort

Purchase frequency
• The further away from the Adequation Attachement to brand

origin, the more significant/ Dimension I

meaningful

• Some dimensions are correlated with customers’

preferences, choices, and ultimately market shares.
Some are not.

• Example:
– Wonderbra is well positioned on the « Redefine Body »
dimension, but we have no guarantee that this characteristic is
desired by the target population

• We need to link perceptions to preferences

• Preference data (one row per respondent)

Preference Data
Preference score data obtained for each brand from each respondent.

Respondents / Christian Princesse

Aubade Calvin Klein Darjeeling Dim Esprit Etam Playtex Wonderbra
Brands Dior Tam-Tam
Respondent 1 5 5 2 5 4 4 4 3 5 5
Respondent 2 4 5 3 4 5 3 3 1 6 4
Respondent 3 6 5 5 4 4 2 3 4 4 5
Respondent 4 5 6 4 4 3 3 3 3 4 4
Respondent 5 6 4 2 5 4 3 5 3 6 2
Respondent 6 5 6 4 4 3 3 3 4 5 5
Respondent 7 5 4 4 5 3 2 4 2 4 3
Respondent 8 6 5 5 4 3 5 4 3 5 2
Respondent 9 7 4 3 4 2 3 3 2 4 2
Respondent 10 4 3 4 4 4 5 5 3 4 2
Respondent 11 5 3 3 3 4 3 4 3 4 2
Respondent 12 5 3 4 5 2 4 4 2 3 3
Respondent 13 4 4 6 5 4 4 3 1 5 3
Respondent 14 5 3 3 5 3 4 5 3 4 2
Respondent 15 5 2 4 5 4 6 4 4 5 2
Respondent 16 4 4 6 3 4 3 4 1 5 1
Respondent 17 3 3 4 2 5 4 4 2 3 1
Respondent 18 5 2 4 4 4 4 4 1 4 4
Respondent 19 5 4 1 4 3 2 2 3 2 7
Respondent 20 5 2 4 4 2 2 4 4 2 7

Wonderbra

Each fuchsia line

represents the Redefine body

Preferences/choices of a Playtex Christian Dior

Price
Special

customer
Esprit Darjeeling Sexy
Dimension II Calvin Klein

• Some dimensions Aubade Esthetism

Originality

barely matter Dim

Etam Princesse Tam
-Tam
Seduction

Comfort

• Customers’ preferences
are heterogeneous Purchase frequency

Adequation Attachement to brand

Dimension I

Stage 1 Design the positioning study

• Select relevant Brands
• Select relevant Dimensions

Stage 2 Collect data

• Develop the survey
• Collect data

Stage 3 Analyze data

• How many dimensions?
• Interpret brands and competition
• Interpret key dimensions and axes
• Identify brands/dimensions not well explained

Design the positioning study

• Select relevant Brands

– Select products which are known by the segment and which can be
evaluated
– Don’t forget to include the leading brands of the product category

• Select relevant Dimensions

– Select descriptive attributes (relevant ones)
– Don’t forget to include most representative attributes of the brands
included (e.g., if you include “Volvo”, don’t forget “Safety”)
– The selected attributes need to be related to known perceptions on
which the products are usually evaluated
– Choose the evaluation scale for the attributes (e.g., 1-7)

Collect data
• Develop the survey
– Use the software to design a template
– Perceptual data
– Preference data

• Collect data
– One matrix per respondent
– Perceptual data: averaged over the sample
– Preference data: one line per respondent
Respondent 1
Attributes / Brands Brand A Brand B Brand C Brand D
Attribute 1
Attribute 2
Attribute 3
Preference Score

Analyze data
• How many dimensions?
• If 2-dimension map not enough, go to 3-dimension map
• Check “variance explained” in the output

Dimensions / Items 1 2 3 4 5 6 7 8
Total variance explained 47% 21% 14% 10% 6% 2% 1% 0%
Cumulative variance explained 47% 67% 81% 91% 97% 99% 100% 100%

• Interpret brands and competition

• Which brands are close to one another?

• Interpret key dimensions and axes

• Which dimensions are correlated? Negatively correlated? Perpendicular?
• Can you give a meaning to the x- and y-axis?

• Identify brands/dimensions not well explained/captured

• Bad positioning?
• Unknown from respondents?
• Flawed or missing data?

• Check how customer perceptions of client products

compare to perceptions of competitors.
• Identify product strengths and weaknesses.
• Select competitors to compete against.
• Determine how much change is needed on key product
attributes to move products to more favorable positions.
• Visually determine impact of communications programs
on market perceptions.

• Leverage strategically advantaged resources:

– Product characteristics
 features, performance, durability, conformance, reliability, style,
etc.
– Service attributes
 delivery, installation, consultation, customer training, repair, etc.
– Personnel
 competency, credibility, courtesy, responsiveness, etc.
– Brand image
 symbols, emotion, personality, etc.

• just clever slogans or gloss

• what you push on your customers -- it is what customers
will realistically grant you
• a one-time activity
• an appeal to everyone
• a way to make your product superior -- it is about finding
a superior position in the minds of your customers.

Positioning is a way to “live your brand.”

© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-
commercial educational purposes only is authorized if this notice appears on every copy. 53
Takeaway Hierarchy of Important
Constructs
1. Awareness (for our analysis, we assumed everyone in the target
segment is aware of the competing brands).
2. Beliefs and Perceptions (these are the locations of brands on
various attributes used in our analyses).
3. Consideration: The set of brands a consumer would consider
buying.
4. Preferences (these represent intrinsic values that a consumer
attaches to different brands, usually based on their beliefs and
perceptions of those brands).
5. Choices/Market Shares: Choices represent pre-dispositions of
consumers to act favorably toward purchasing a particular
product. However, consumers will not necessarily purchase
products they prefer the most. This is because they face several
constraints (e.g., budget constraints) that prevent them from
always purchasing their most preferred products.

• Provides a static model - ignores dynamics of customer

perceptions.
• Interpretation is sometimes difficult.
• Does not incorporate cost or likelihood of being able to
achieve a desired positioning.
• Does not incorporate a “probability model” to indicate
goodness of a map.
• Generally, need about 6 to 8 products in a category to
make the technique useful.

• Life style (self-concept) positioning

• Attribute positioning
• Benefit positioning
• Competitive positioning
• Time-based (e.g., usage occasion) positioning

• Our product is unique (e.g., Perdue chicken).

• Our product is different (e.g., Listerine).
• Our product is similar (e.g., Meister Brau).

Four questions that could trigger positioning analyses

• Who do we need to target?
• What do they think now?
• What do we want them to think, now and from now on? Why?
• How do we do that?

© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-
commercial educational purposes only is authorized if this notice appears on every copy. 58
Hard to See Patterns in Customer
Data Without Analytic Tools
Ratings of nine brands of notebook computers on several
attributes
B1 B2 B3 B3 B4 B5 B6 B7 B8 New
Attractive 5.1 3.6 3.5 5.4 3.9 4.8 5.2 4.0 5.2 4.0
Light 6.0 3.5 5.0 3.9 3.3 5.3 5.0 2.5 5.5 2.5
Unreliable 3.4 4.1 4.5 2.1 4.5 2.7 4.5 3.7 2.5 3.8
Plain 1.5 4.1 2.9 2.3 4.5 2.7 3.5 4.3 2.2 5.2
Battery life 3.3 4.9 4.3 4.1 3.9 3.0 3.5 6.2 3.5 4.0
Screen 3.5 5.3 3.4 6.4 5.4 5.2 3.3 6.0 3.3 4.8
Keyboard 2.6 3.5 2.5 3.4 3.8 3.3 2.8 5.0 4.3 4.7
Roomy 5.5 4.3 5.4 3.1 3.4 3.3 4.7 3.5 4.3 4.2
Easy service 4.5 4.9 3.3 5.0 4.4 4.5 3.3 4.7 3.8 4.5
Expandability 5.5 4.3 5.4 3.1 3.4 3.3 4.7 3.5 4.3 4.2
Setup 5.6 3.5 5.6 5.4 2.5 4.2 5.2 3.3 5.8 2.5
Common 4.1 3.5 3.3 2.9 4.0 4.3 2.2 4.2 3.3 4.2
Value 3.5 4.8 4.4 3.6 3.6 2.7 3.2 4.7 3.5 4.0
Preference 7.4 3.4 4.8 6.6 4.4 7.4 7.1 3.8 6.9 3.3

Et An organic
chemist reads
H H
i) F ii) G diagrams such as
the one here
H2 SO4/H2 O NaOH/H2 O

An Electrical
Engineer reads
diagrams such as
the one here

© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-
commercial educational purposes only is authorized if this notice appears on every copy. 60
Conventional Mapping Using
Snake Charts
1) Company provides adequate insurance coverage for my car.
Does not
Describes it describe it 2) Company will not cancel policy because of age, accident
completely at all experience, or health problems.
| | | | | | 3) Friendly and considerate.
0 1 2 3 4 5 4) Settles claims fairly.
5) Inefficient, hard to deal with.
6) Provides good advice about types and amounts of coverage to buy.
7) Too big to care about individual customers.
8) Explains things clearly.
9) Premium rates are lower than most companies.
10) Has personnel available for questions all over the country.
11) Will raise premiums because of age.
12) Takes a long time to settle a claim.
13) Very professional/modern.
14) Specialists in serving my local area.
15) Quick, reliable service, easily accessible.
16) A “good citizen” in community.
17) Has complete line of insurance products available.
18) Is widely known “name company”.
19) Is very aggressive, rapidly growing company.
20) Provides advice on how to avoid accidents.

1. Select target segment

2. Determine relevant competitive offerings
3. Determine potential differentiator-dimensions
4. Select sample of customers in target segment and get
ratings of competitors on the selected dimensions
5. “View” the results (Perceptual Maps)
6. Relate to preference, choice, or market share
(Preference Maps)
7. Develop positioning statement and associated
strategies

• The arrow indicates the direction in which that attribute is increasing

(The attribute is decreasing in the direction opposite to the arrow).
Thus, airlines positioned farther to the South direction are perceived
as being more punctual and those positioned in the North direction
are less punctual.
• The length of the line from the origin to the arrow indicates the
variance of that attribute explained by the 2D map. The longer this
line, the greater is the relevance of that attribute in helping you to
interpret the map.
• Attributes that are both relatively relevant and close to the horizontal
(vertical) axis help determine the meaning of the axis.
• To position an airline on each attribute, draw an imaginary
perpendicular line from the location of that airline onto that attribute.

Mapping Methods in Marketing

Joint Space Maps

Perceptual Maps Preference Maps (includes both perception &
preference)

Similarity-based Ideal-point model Simple “Joint space

methods (unfolding model) maps” using modified
perceptual mapping
Attribute-based Vector model methods
methods
Vector model
using PREFMAP-3

Ideal-point model
using quasi-metric
Included in Marketing Engineering. approach

Ideal-Point Preference Vector Preference Model

Model

Increasing
Preference

Preference Preference

Decreasing
Preference

Ideal Point

Attribute Attribute
(e.g., sweetness) (e.g., service speed)

Objective—Introduce customer preferences into perceptual

maps:
• A simple ideal point method: Introduce an “ideal” brand as an additional
stimulus evaluated by customers.
1) Combine attribute ratings of ideal brand with the other brands.
2) Compare how similar the ideal brand is to the other brands included in the study.
• A simple vector method: Introduce “preferences” as an additional
variable within the attribute ratings data
1) Analyze the preference variable simultaneously with evaluations of the cars on
each attribute.
 Identify which attributes influence consumer preferences the most
 Identify which brands are most preferred in the aggregate.

2) Analyze preferences separately for each customer.

 Identify target segments that prefer the brand of interest the most.

• Incorporate Individual-level Preference Data

(PREFMAP-3 and Quasi-metric Methods)
– First, develop a perceptual map of competing alternatives.

– Given the fixed positions of the alternatives on the perceptual map, add
the preferences of each customer on the perceptual map (this
approach is called external analysis).

Ideal-Point Map Vector Map

Ideal
Point (I)
Preference
dIA Vector
dIB
A A
dAC
B dAB
C B

(a) (b)
A is preferred twice as much as B. A is preferred to B and B is preferred to C.
(dIB = 2dIA) With reference to A, C is preferred half as
much as B (dAC = 2dAB)

© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-
commercial educational purposes only is authorized if this notice appears on every copy. 68
Preference Data for PREFMAP-3
and Quasi-metric Approaches
• Matrix of inputs (higher numbers indicate greater preference):

C1 C2 C3 C4 ...............
American 2 3 9 3
United 7 7 2 5
US Airways 3 8 3 5
•
•
•
An element of this matrix is the stated preference of each consumer (C1, C2, . . .) for each brand. A
higher number indicates higher preference for that brand.

• Use the Perceptual and Preference data option in Marketing Engineering to

obtain a “joint space” map showing relationships between brands and
preferences between brands for each consumer.
• Interpret the map. The relative locations of the brands are directly provided by
the map. The attributes are represented as vectors. Individuals (i.e., their
preferences) are represented as vectors for the PREFMAP-3 model and as
ideal points for the quasi-metric model

There are several ways to translate Preferences to

Choices, and subsequently, to Market Shares. We
consider two possibilities:
• First Choice or Maximum utility rule
– Consumer chooses the product s/he prefers the most
• Share of preference rule
– Consumer chooses each product in his/her consideration set
in proportion to the relative preference for that product as
compared to the other products.

© Copyright DecisionPro, Inc. 2018. Commercial distribution and any publication is not authorized. Copying and distribution for non-
commercial educational purposes only is authorized if this notice appears on every copy. 70
Evaluating Perceptual and Preference
Maps
• Technical adequacy
– What percentage of the total information (variance) in the raw data is captured in the
map?
– What percentage of the information of each attribute (variance) is captured in the
map?
• Managerial interpretation (example questions)
– What underlying dimensions seem to characterize how customers view the products?
– What is the competitive set associated with the target product or new concept?
– How well is a target product positioned with respect to the existing products?
– Which attributes are related to each other?
– Which attributes influence customer preferences positively? negatively?
– What improvements will enhance the value of a product or new concept?
– Which customer segments have positive perceptions and high preference for the
product?

Apple Segmentation PDF
100% (1)
Apple Segmentation PDF
33 pages
Segmentation, Targeting, Positioning
92% (12)
Segmentation, Targeting, Positioning
11 pages
Post Graduate Program in Business Analytics: M&CRM: STP, Customer Intelligence
No ratings yet
Post Graduate Program in Business Analytics: M&CRM: STP, Customer Intelligence
31 pages
Lec5 Customer Analysis STP
No ratings yet
Lec5 Customer Analysis STP
5 pages
BT4211 Data-Driven Marketing: Product: Segmentation, Targeting
No ratings yet
BT4211 Data-Driven Marketing: Product: Segmentation, Targeting
31 pages
Segmentation and Targeting
No ratings yet
Segmentation and Targeting
20 pages
Market Segmentation, Targeting, and Positioning
100% (1)
Market Segmentation, Targeting, and Positioning
31 pages
Segmentation & Targeting
No ratings yet
Segmentation & Targeting
77 pages
Cluster Analysis
No ratings yet
Cluster Analysis
5 pages
Segmentation and Targeting: Basics Market Definition Segmentation Research and Methods Behavior-Based Segmentation
No ratings yet
Segmentation and Targeting: Basics Market Definition Segmentation Research and Methods Behavior-Based Segmentation
77 pages
Market Segmentation
No ratings yet
Market Segmentation
26 pages
BY Dr. Kumar Saurav Segmentation Targeting and Positioning About The STP Process?
No ratings yet
BY Dr. Kumar Saurav Segmentation Targeting and Positioning About The STP Process?
11 pages
STP: Segmentation, Targeting and Positioning: Presented By: Ravi Ramoliya
No ratings yet
STP: Segmentation, Targeting and Positioning: Presented By: Ravi Ramoliya
35 pages
Milpark PMAR01-6 Topic 3 - PPT - BBA 2018 - Updated 2019
No ratings yet
Milpark PMAR01-6 Topic 3 - PPT - BBA 2018 - Updated 2019
70 pages
Chapter 4 STP
No ratings yet
Chapter 4 STP
44 pages
ADVANCED MARKETING - Teoria
No ratings yet
ADVANCED MARKETING - Teoria
54 pages
Module 3
No ratings yet
Module 3
27 pages
Market Segmentation
No ratings yet
Market Segmentation
11 pages
by Amit Kumar Shuklla (1) .PPT 2
No ratings yet
by Amit Kumar Shuklla (1) .PPT 2
19 pages
Customer Segmentation
No ratings yet
Customer Segmentation
7 pages
04 Market Segmentation &amp Positioning RAC
No ratings yet
04 Market Segmentation &amp Positioning RAC
28 pages
Chapter 6 Marketing Management
No ratings yet
Chapter 6 Marketing Management
10 pages
Market Segmentation Targeting Positioning
100% (3)
Market Segmentation Targeting Positioning
14 pages
Basics of Marketing
No ratings yet
Basics of Marketing
21 pages
Chapter 4 EDIT
No ratings yet
Chapter 4 EDIT
12 pages
Segmentation Targeting Positioning (STP)
No ratings yet
Segmentation Targeting Positioning (STP)
20 pages
1
No ratings yet
1
60 pages
CH 4 (3)
No ratings yet
CH 4 (3)
27 pages
market segmentation.docx
No ratings yet
market segmentation.docx
7 pages
Market Segmentation: Uses and Bases of Segmentation
No ratings yet
Market Segmentation: Uses and Bases of Segmentation
21 pages
The STP Process
No ratings yet
The STP Process
25 pages
Reflection and Discussion Week 2-Dina
No ratings yet
Reflection and Discussion Week 2-Dina
6 pages
Market Segmentation, targeting, and positioning
No ratings yet
Market Segmentation, targeting, and positioning
39 pages
Nawaz
No ratings yet
Nawaz
69 pages
Lecture 5
No ratings yet
Lecture 5
45 pages
Principles of Mrktg. - Q3 Module 7
No ratings yet
Principles of Mrktg. - Q3 Module 7
11 pages
Market Segmentation Targetting and Posit
No ratings yet
Market Segmentation Targetting and Posit
51 pages
Consumer
No ratings yet
Consumer
40 pages
Seg Me Nation
No ratings yet
Seg Me Nation
27 pages
Market Segmentation
No ratings yet
Market Segmentation
43 pages
Strategic Marketing: Course Credit:02
No ratings yet
Strategic Marketing: Course Credit:02
55 pages
Module 2 Segmentation and Targeting
No ratings yet
Module 2 Segmentation and Targeting
43 pages
STP Process
No ratings yet
STP Process
8 pages
Unit 2
No ratings yet
Unit 2
10 pages
Unit 4
No ratings yet
Unit 4
62 pages
doneSESSION 14
No ratings yet
doneSESSION 14
19 pages
Chapter 3 Strategic Market Segmentation
No ratings yet
Chapter 3 Strategic Market Segmentation
33 pages
PM Chapter 3 PART 3
No ratings yet
PM Chapter 3 PART 3
18 pages
BUS Marketing 3
No ratings yet
BUS Marketing 3
24 pages
Market Segmentation
100% (1)
Market Segmentation
40 pages
Identifying Market Segments
No ratings yet
Identifying Market Segments
14 pages
Market Segmentation
No ratings yet
Market Segmentation
21 pages
Strat 4
No ratings yet
Strat 4
36 pages
Marketing STP
No ratings yet
Marketing STP
23 pages
Lesson 15 STP
100% (1)
Lesson 15 STP
24 pages
Chapter 3
No ratings yet
Chapter 3
28 pages
Marketing Matters: A Market Analysis Methodology Leading to a Marketing Simulation Capability
From Everand
Marketing Matters: A Market Analysis Methodology Leading to a Marketing Simulation Capability
Kendall Carey
No ratings yet
Marketing Mastery: 1,000 Ideas for Success
From Everand
Marketing Mastery: 1,000 Ideas for Success
Abed Shouman
No ratings yet
Product Journey
From Everand
Product Journey
The Wordsmith of Wonder
No ratings yet
Global Expansion Route-to-Market Options for Scale-Ups: Guidelines for addressing Route-to-Market choices during international expansion
From Everand
Global Expansion Route-to-Market Options for Scale-Ups: Guidelines for addressing Route-to-Market choices during international expansion
Jan Buis
No ratings yet
Ward Clustering Algorithm
100% (1)
Ward Clustering Algorithm
4 pages
Unit 2 FDS
No ratings yet
Unit 2 FDS
13 pages
Pattern Classification and Scene Analysis PDF
100% (1)
Pattern Classification and Scene Analysis PDF
8 pages
Clustering Monograph DSBA
No ratings yet
Clustering Monograph DSBA
36 pages
ML - 8
No ratings yet
ML - 8
70 pages
2022 BM MRS Cluster Analysis
No ratings yet
2022 BM MRS Cluster Analysis
23 pages
13_Unsupervised_Learning
No ratings yet
13_Unsupervised_Learning
9 pages
Techniques of Cluster Analysis: A Seminar On
No ratings yet
Techniques of Cluster Analysis: A Seminar On
25 pages
Hands On Machine Learning with R 1st Edition Brad Boehmke (Author) pdf download
No ratings yet
Hands On Machine Learning with R 1st Edition Brad Boehmke (Author) pdf download
53 pages
Fundamentals of Data Science Unit 3
No ratings yet
Fundamentals of Data Science Unit 3
15 pages
Machine Learning
No ratings yet
Machine Learning
115 pages
Defining Homogenous Climate Zones of Bangladesh Using Cluster Analysis
No ratings yet
Defining Homogenous Climate Zones of Bangladesh Using Cluster Analysis
11 pages
MCQ Amt 2
No ratings yet
MCQ Amt 2
9 pages
Lec 06 Clustering
No ratings yet
Lec 06 Clustering
44 pages
Path-Based Spectral Clustering: Guarantees, Robustness To Outliers, and Fast Algorithms
No ratings yet
Path-Based Spectral Clustering: Guarantees, Robustness To Outliers, and Fast Algorithms
66 pages
ML NOTES MOD 4
No ratings yet
ML NOTES MOD 4
27 pages
Full Download Introduction to Artificial Intelligence for Security Professionals 1st Edition The Cylance Data Science Team PDF DOCX
100% (2)
Full Download Introduction to Artificial Intelligence for Security Professionals 1st Edition The Cylance Data Science Team PDF DOCX
35 pages
SAP HANA Predictive Analysis Library PAL en
No ratings yet
SAP HANA Predictive Analysis Library PAL en
578 pages
Fuzzy Clustering
No ratings yet
Fuzzy Clustering
6 pages
Clustering
No ratings yet
Clustering
8 pages
MLT Quantum
No ratings yet
MLT Quantum
163 pages
Machine Learning Techniques Quantum
No ratings yet
Machine Learning Techniques Quantum
161 pages
ML Unsupervised Notes
No ratings yet
ML Unsupervised Notes
26 pages
CZ4032 Data Analytics & Mining Notes
No ratings yet
CZ4032 Data Analytics & Mining Notes
16 pages
How To Pass Sem 5 - Comps
No ratings yet
How To Pass Sem 5 - Comps
11 pages
Clustering Techniques
No ratings yet
Clustering Techniques
38 pages
Marielle Caccam Jewel Refran
No ratings yet
Marielle Caccam Jewel Refran
100 pages
ch07 Clustering
No ratings yet
ch07 Clustering
56 pages
Instant Access to Practical Guide to Cluster Analysis in R Unsupervised Machine Learning Alboukadel Kassambara ebook Full Chapters
100% (2)
Instant Access to Practical Guide to Cluster Analysis in R Unsupervised Machine Learning Alboukadel Kassambara ebook Full Chapters
52 pages
DSA Presentation Group 6
No ratings yet
DSA Presentation Group 6
34 pages