Perceptual Mapping

Data for this session is available in Data – Perceptual Mapping

Perceptual Mapping

Skander Esseghaier

1

In this session, you will learn:

„

„

„

How to construct a map of product locations in the perceptual space of consumers How to do it using Minitab What attributes you should use when constructing a perceptual map 2

What is Perceptual Mapping „

„

„

„

A technique to understand the position of brands as consumers perceive them The output is a map of product locations in the perceptual space of consumers Though consumers may think about a number of attributes in evaluating products, it may be possible to summarize these attributes because consumer perceptions along these attributes may be correlated We can use factor analysis to find this reduced perceptual space and map the products in this space 3

Perceptual Map… Final Output Perceptual Map for Cars 1.5

VW Golf

Dodge Neon

1 0.5

Economy

Camry

Taurus 0

-1.5

-1

-0.5

0

0.5

1

1.5

2

-0.5 -1

Lexus ES 300 BMW325

-1.5 Fashion

4

Perceptual Maps: An Illustration Using the Car Market „

Cars Considered „ „ „ „ „ „

Ford Taurus Toyota Camry Volkswagen Golf BMW 3-Series Lexus ES300 Dodge Neon

5

Survey… „

Each respondent is asked to rate 6 cars on a number of attributes on a 1-7 scale „ „ „ „ „ „

Affordability Practicality Classiness Sportiness Youth Appeal Fun to Drive 6

Data… Respondent 1 1 1 1 1 1 2 2

20 20 20 20 20 20

Car Taurus

Afford 7

Practical 5

Classy 2

Sporty 5

Youthful 5

Fun 5

Neon Camry Lexus BMW VW Taurus Neon

Taurus Neon Camry Lexus BMW VW 7

Quick and Dirty Sense of the Data: Looking at the Correlation Matrix

Afford Afford Practical Class Sporty Youth App Fun

1 0,013004 -0,61826 -0,34601 -0,0619 -0,17834

Practical

Class

Sporty

1 -0,48435 1 -0,77767 0,798962 1 -0,72219 0,257668 0,636417 -0,73562 0,573691 0,852368

Youth App

Fun

1 0,64312

1

Fair amount of correlations between variables Æ indicates that Factor Analysis may be useful 8

Simple Approach: Plot Attribute by Attribute Practicality Vs Sportiness 1.5

BMW 1

Lexus Neon

0.5 0

-1.5 Sportiness

-1

-0.5

0

0.5

-0.5 -1

1

1.5

Taurus Camry Volkswagen

-1.5 Practicality

Can lead to too many maps

9

First Step: Do Principal Component Analysis (PCA)

„

„

This allows us to select the # of factors PCA uses the correlation matrix of the data and constructs factors „ „

if there are n variables we will have n factors first factor will explain most variance, second next and so on…

10

Minitab Output of PCA: Eigen Analysis

Eigenanalysis of the Correlation Matrix Eigenvalue Proportion Cumulative

3.7336 0.622 0.622

1.3383 0.223 0.845

0.4641 0.077 0.923

0.2558 0.043 0.965

0.1514 0.025 0.991

0.0568 0.009 1.000

84.5% of variance in 6 variables is explained by just 2 factors 11

Minitab Output of PCA: Scree Plot Scree Plot of Afford-Fun 4

Eigenvalue

3

2

1

0 1

2

3

4

5

6

Component Number 12

Second Step: Do Factor Analysis „

„

Perform factor analysis with the factors selected from Step 1 Interpret resulting factors „ „

„

use factor loadings and loading plot to interpret factors if it is not interpretable use rotation options until we get something that can be interpreted

Look at factor equations and factor scores „

score plots will be useful 13

Unrotated Factor Loadings: Variable’s Correlation with the Factors Unrotated Factor Loadings and Communalities Variable Afford Practica Class Sporty Youth Ap Fun

Factor1 0.376 0.849 -0.774 -0.965 -0.740 -0.890

Factor2 -0.841 0.367 0.529 0.050 -0.434 -0.158

Variance % Var

3.7336 0.622

1.3383 0.223

Communality 0.849 0.855 0.879 0.934 0.736 0.818 5.0719 0.845 14

Interpreting Factors: Looking at Loading Plot without Rotation Loading Plot of Afford-Fun Class 0.5

Second Factor

Practica

Sporty 0.0

Fun

Youth Ap -0.5

Afford

-1.0

-0.5

0.0

First Factor

0.5

15

Rotated Factor Loadings: Variable’s Correlation with the Factors Rotated Factor Loadings and Communalities Varimax Rotation Variable Afford Practica Class Sporty Youth Ap Fun

Factor1 0.063 -0.922 0.435 0.829 0.857 0.860

Factor2 0.919 0.075 -0.831 -0.498 0.036 -0.279

Variance % Var

3.2045 0.534

1.8674 0.311

Communality 0.849 0.855 0.879 0.934 0.736 0.818 5.0719 0.845 16

Interpreting Factors: Looking at Loading Plot with Rotation Attribute-Factor Relationship (Loading Plot) 1

Affordability

0.8 0.6 0.4

Practicality

0.2

Youth

0 -1 Factor 2

-0.5

0 -0.2

0.5

Fun

1

-0.4 -0.6 -0.8

Sportiness Classiness

-1 Factor 1

17

Naming Factors

„

Can we name these factors?

„

This highlights the subjectivity involved here

18

How Did Cars Score on Fashion and Economy Factors? Factor Score Coefficients Variable Afford Practica Class Sporty Youth Ap Fun

Factor1 0.206 -0.330 -0.003 0.211 0.327 0.266

Factor2 0.602 -0.135 -0.446 -0.154 0.193 -0.008

Fashion = 0.206 Affordability - 0.330 Practicality - 0.003 Classiness + 0.211 Sportiveness + 0.327 Youthful Appeal + 0.266 Fun Economy = 0.602 Affordability -0.135 Practicality -0.446 Classiness - 0.154 Sportiveness + 0.193 Youthful Appeal - 0.008 Fun

We standardize the variables and then take the average of the 20 consumer’s ratings 19 on the standardized variable to plug in

Alternative Approach To Compute Factor Scores for Each Car

„

Store Minitab’s factor scores for each car based on each consumer’s ratings „

„

we will have 20*6=120 numbers

Average the factor scores across consumers for each car „

we will get the factor score for each car

20

Where are Cars Located in the Perceptual Space? Perceptual Map for Cars 1.5

Volkswagen

Neon

1 0.5

Camry

Taurus 0

-1.5 Economy

-1

-0.5

0

0.5

1

1.5

2

-0.5 -1

Lexus BMW

-1.5 Fashion 21

Applications of Perceptual Maps „

Who are our competitors?

„

On what dimensions do we compete?

„

Where to introduce new products? „ „

you also need to be aware of consumer preferences look for locations with relatively more consumers but limited competition

22

Caveats „

Identify relevant attributes „

„ „

„

don’t miss important attributes (Exploratory Research is important) no point asking about unimportant attributes conjoint analysis may be useful in identifying what attributes are important to consumers

Identify discriminating attributes „ „

don’t use primary attributes (like cleaning power of detergents) there should be real perceptual differences on average for the product 23

Step 1: Choose Number of Factors to Extract Do Principal Component Analysis (PCA) „

In Minitab select Stat>Multivariate>Principal Components…

„

Select the variables you want to factor analyze in Variables box

„

„

Select “Correlation” as the data that will be analyzed; this will mean that the data will be standardized and therefore each variable will have equal effect Ask for Scree Plot (using Graphs button) which graphs the amount of variance explained by each factor 24

Step 2: Perform Factor Analysis Do Factor Analysis „

in Minitab, Stat>Multivariate>Factor Analysis….

„

number of Factors to extract should be from Step 1

„

„ „

try “None” rotation for a start (else try Varimax or others if it doesn’t work) In Graphs: select loading plot (score plot is not useful here) In Storage: in the scores box store the factor scores by selecting 2 variables 25

Step 3: Plot the Perceptual Map „

Take the average of the factor scores for each car

„

Use these average scores to plot the perceptual map

26