This project investigates how to identify different customer types in real-time and track their behavior change based on analysis of point-of-sale and other data.
With the pervasive availability of smart phones it becomes for the first time economically feasible to operationalize personalized marketing measures also for physical grocery retailing. A particularly interesting and high value target group in this domain is the one of variety seeking, since this group is most likely to respond positively to new offers and recommendations. However, present methods in identifying variety seekers rely on questionnaires and ignores that variety seeking may differ between product categories. In this project, we present a model for measuring variety seeking behavior on a high level of granularity, based on a consumer’s purchases in individual product categories. Our study has three main contributions. Firstly, we contribute to the customer segmentation research stream by providing a novel way for identifying customers’ overall extent of variety seeking as well as their specific variety seeking at a category level. Second, for the most important retail categories we characterize the extent of variety seeking and provide a data-driven approach that is easy to operationalize by practitioners – especially for deploying large-scale personalized marketing measures in social or mobile commerce in physical stores. Finally, we provide a method to reconcile the highly granular category-level results with existing per person typologies.
Customer segmentation, typologies, shopping types, big data, analytics, real-time