Abstract: Personalized marketing has emerged as a critical marketing strategy as a result of the success of e-commerce and the accessibility of digital marketing data. It is well known that different groups of customers might react differently to the same marketing strategy, owing to their individual preferences. As such, we propose a pairwise subgrouping approach that can be used to identify subgroups and categorize similar marketing effects into groups. Specifically, we model customers’ purchase decisions as binary responses under a generalized linear model framework, while incorporating their longitudinal correlation. We penalize the pairwise distances between heterogeneous effects to formulate subgroups, where a subgroup is associated with a unique marketing effect. We establish the theoretical consistency of the subgroup identification in the sense that the true underlying segmentation structure can be recovered successfully. Here, we also establish the parameter estimation consistency. We conduct numerical studies and apply the proposed approach to IRI marketing data on in-store display marketing effects. The results show that the proposed method outperforms competing methods in terms of identifying subgroups and estimating marketing effects.

Key words and phrases: Alternating direction and method of multipliers, individualized modeling, marketing egmentation, minimax concave penalty, subgroup identification.