Deriving joint space positioning maps from consumer preference ratings

Abstract

Joint space multidimensional scaling maps are often utilized for positioning analyses and are estimated on survey samples of consumer preferences, choices, considerations, or intentions so as to provide a concise spatial depiction of the competitive landscape including relevant dimensions or attributes, competing brands, and consumers in the same joint space representation. Care has to be given concerning the underlying scale properties of such survey data so as not to distort the resulting joint space positioning map. We present a new joint space multidimensional scaling procedure for positioning analyses for displaying the structure in such survey data when such common ordered successive category measurement scales such as Likert, Edwards, semantic differential, etc., are employed. We present the technical details of this stochastic ordered preference multidimensional scaling vector model as well as the maximum likelihood estimation-based algorithm devised for parameter estimation. Favorable comparisons are made with several existent multidimensional scaling methods in representing the internal structure for such data in marketing positioning studies. An actual commercial positioning application concerning large sports utility vehicles consideration to buy judgments is presented with predictive validation comparisons with other multidimensional scaling joint space procedures.