Visualizing Consumers' Preference for Products by Ordinal Unfolding using PrefScal


Speaker


Abstract

In various areas of marketing research, data are gathered on the preference of consumers for products. Unfolding is a technique for the analysis of such preference data. The purpose of unfolding is to represent the preferences of consumers for products as distances between points in a low dimensional space, such that a consumer is positioned closest to the product he or she prefers the most. An advantage of unfolding is that it allows direct comparison of products, even though the data contain only product to consumer information. Nonmetric unfolding additionally treats the preference rank orders ordinally, which seems a logical choice for preference data.

The major problem with nonmetric unfolding has been that it often results in perfect, but degenerate solutions in which all distances between consumers and product ideas are equal. However, recently Busing, Groenen, and Heiser (in press) proposed a penalty approach and succeeded to avoid the degeneracy. Their approach is implemented in a computer program called PrefScal which is scheduled to appear in SPSS 14. In this seminar, we show in a non technical way how the degeneracies in unfolding come up and why the PrefScal approach succeeds in avoiding degeneracies in nonmetric unfolding. We apply PrefScal to consumer preferences on new product propositions for ready-made soups. Based on the unfolding analysis recommendations can be made as to what product ideas should be considered for market introduction.