Customer segmentation in retailing based on retail brand patronage patterns

In this study, we propose an alternative method for customer segmentation based on households' retail brand portfolios. We suggest blockmodeling, a method that is rooted in social network analysis, to identify homogeneous groups of customers being characterized by similar patronage patterns. In a second step, we combine this method with psychometric analyses to examine group differences based on sociodemographic and purchasing behavior variables. To illustrate our approach, we use a sample of 500 households out of a panel data set of 36,502 German households in the fast-moving consumer goods sector. The findings of our analyses yield support for our assumption that our methodology, taking into account the relationship between households and the different retail brands they patronize, provides a broader understanding of the antecedents of households' multiple store patronage strategies, leading eventually to a new form of customer segmentation.