The Marketing Book 5th Edition

Quantitative methods in marketing 207

Table 9.1 Main multivariate methods and their marketing applications

Method Based on Marketing

applications

Main advantages Main limitations

Factor

analysis

Identification of

relationships among

variables and

establishing the

‘weight’ (factor

loadings) for these

variables

Determine

corporate marketing

images, consumer

behaviour and

attitudes

Data reduction,

identification of the

main constructs

(factors that

underline the data

characteristics)

Applicable only to

interval-scaled data

Latent

analysis

Investigation of both

manifest and latent

factors by

estimating these

latent parameters

Segmentation

research, market

structure analysis

(Dillon and Mulani,

1989)

Could be used for

investigating causal

systems involving

latent variables

Difficulties in

estimating the latent

variables

Cluster

analysis

Developing

similarity or

dissimilarity

measures

(coefficients), or

distance measures,

to establish clusters

association

Primarily for

segmentation

studies and strategy

(Saunders, 1994)

Enables classification

of brands, products,

customers,

distributors, etc.

Different clustering

methods could

generate different

clusters

Multidimensional

scaling

Calculating the

proximity (or,

alternatively, of

dominance) among

attributes/variables

and respondents

Market research,

market share

analysis (Coates et

al., 1994), market

segmentation, brand

positioning, etc.

Presents the entire

structure of

variables, making it

easier to visualize

and interpret

relationship/

similarities among

data

Different software

packages required

for different types

of data input

Conjoint

analysis

Measurement of

psychological

judgements by

measuring the joint

effect of two or

more independent

variables on the

ordering of a

dependent variable

Consumer research

(Vriens, 1994),

advertising

evaluations (Stanton

and Reese, 1983)

Enables calculation

of preferences.

Suitable for product

design and attitude

measurement

Measures first utility

rather than

behaviour

Correspondence

analysis

Graphical technique

for representing

multidimensional

tables. For

procedure, see

Figure 9.3

Selling functions in

bank branches

(Meidan and Lim,

1993), market

segments, track

brand images

Can be used for

analysing binary,

discrete and/or

continuous data.

Facilitates both

within- and

between-set

squared distance

comparison. Fast,

easy to interpret

Limited applications

in marketing

because of lack of

suitable software