Customer Migration Patterns: Evidence from a North American Retailer

Abstract

In this study, we assess how marketing activities influence the extent to which customers become more (or less) profitable to a company over time. Using data collected from an apparel and household goods multichannel retailer over a three-year period beginning in 2001, we apply a Hidden Markov Model to a longitudinal data set from a cohort of 4,165 customers over a three-year period. We find three segments: inactive, occasional, and loyal. We find 23 unique migration patterns among these segments. Price promotions have a positive effect on increasing migration patterns, but were not significant for stable patterns. Catalogs have a negative impact on stable migration patterns and were not significant for increasing migration patterns. Finally, we find coupons have a negative impact on both stable and increasing migration patterns.