Multidimensional scaling and cluster analysis techniques are commonly employed for the analysis of consumer perceptions of products. However, within the past 10‐15 years, a growing volume of research has shown that the processes underlying similarities judgements of stimuli are incompatible with the fundamental underlying axioms of these techniques. A series of papers in the psychometrics and cognitive psychology literatures by Tversky and his associates have demonstrated the inability of these procedures to handle similarities data from many domains by virtue of the restrictive assumptions they impose on the data. Recently, several procedures have been proposed that overcome the limitations of traditional multidimensional scaling and cluster analysis techniques. The potential benefits are illustrated of applying two of these newer techniques, additive similarity trees (ADDTREE) and extended similarity trees (EXTREE) in the context of marketing research. Consumers′ similarity judgements data are presented from three disparate product domains (newspapers, shops and breakfast cereals). In each case, non‐metric multidimensional scaling and average linkage cluster analysis yield less interpretable solutions than ADDTREE. In the case of the newspapers data, much richer insights are obtained with reference to EXTREE. The paper concludes with a discussion of the implications for market research studies and the development of consumer behaviour theory.