A REVIEW OF THE MAJOR MULTIDIMENSIONAL SCALING MODELS FOR THE ANALYSIS OF PREFERENCE/DOMINANCE DATA IN MARKETING

Abstract

Multidimensional scaling (MDS) represents a family of various spatial geometric models for the multidimensional representation of the structure in data as well as the corresponding set of methods for fitting such spatial models. Its major uses in Marketing include positioning, market segmentation, new product design, consumer preference analysis, etc. We present several popular MDS models for the analysis of consumer preference or dominance data. The first spatial model presented is called the vector or scalar products model which represents brands by points and consumers by vectors in a T dimensional derived joint space. We describe both individual and segment level vector MDS models. The second spatial model is called the multidimensional simple unfolding or ideal point model where both brands and consumers are jointly represented by points in a T dimensional derived joint space. We briefly discuss two more complex variants of multidimensional unfolding called the weighted unfolding model and the general unfolding model. Here too, we describe both individual and segment level unfolding MDS models. We contrast the underlying utility assumptions implied by each of these models with illustrative figures of typical joint spaces derived from each approach. An actual commercial application of consideration to buy large Sports Utility Vehicle (SUV) vehicles is provided with the empirical results from each major type of model at the individual level is discussed.