Abstract

General/observable approaches to market segmentation are commonplace when more is known about the age, income, and social class of customers than the needs and motivations that reside behind purchasing behavior. Although less accessible, product-specific/unobservable approaches to segmentation based on values, attitudes, and life styles are useful for segmenting customers distinguished more so by how they feel about and respond to brands and vendors than their demographic characteristics. Using data from a nationally representative consumer panel, demographic and psychographic segmentations of credit card customers were developed with agglomerative and k-means clustering, and the effects of segmentation on transactional and relational response differences were examined with multivariate analysis of covariance. Results indicate that novice customers segmented on household size, age of household head, income, and consumer debt differed systematically in terms of overall and co-branded credit card use. Segmenting long-tenured customers on product-specific risk, money savviness, debt, and deal proneness values had a significant but somewhat less powerful effect on relationship satisfaction and cross-buying. However, when household size and income were accounted for in the psychographic segmentation, the effect on relational response differences was magnified threefold. Implications for segmentation theory and practice are considered.