Customer and Product Clustering in Retail Business

Abstract

The customer segmentation plays an important role in modern retail marketing strategies. It is a common practice to classify customers involved in a loyalty programme based on frequency and value of their shopping. In addition to this information, we also utilize the characteristics of bought products, specifically the product category. We use k-means to cluster the customers and determine the key factors in the clustering. We also explore the relationship between customer clustering and product clustering. We cluster products into broader categories based on shopping behavior of the customer segments. This approach significantly reduces the size of the product clustering problem and results in meaningful product supercategories. As an application, we analyze a Czech drugstore company.