Retail Patronage as a Network: An Alternative Approach to Customer Segmentation

摘要

We investigate how multiple-store patronage behavior gives rise to distinct retail brand portfolios and how different customer groups using these portfolios can be characterized. We combine methods from two fields: blockmodel analysis, a method from social network analysis, which constitutes clusters of actors based on similarity of their relationship patterns, and multivariate statistical analysis on different sociodemographic and purchasing behavior variables. The first enables us to identify homogeneous groups of customers being characterized by similar patronage patterns. With the second, we examine if customers with similar patronage patterns also share common characteristics, thereby allowing for customer segmentation.