Customer Segmentation in Food Retail Sector: An Approach from Customer Behavior and Product Association Rules

Abstract

In competitive markets, customer segmentation improves customer loyalty and business performance, but in practice, these analyses are carried out using simple relationships in dashboard, or Microsoft Excel’ sheets, which do not show customer behavior. Data segmentation in the era of big data has changed this paradigm with some techniques that try to decrease bias. In this research, four segmentation techniques are tested with a large set of data from a retail store. CLARA (Clustering Large Applications Algorithm) and Random Forest algorithms both were the best. Through the RFM (Recency, Frequency, Monetary) approach, eight customer segments were found, where Champions customers spend more money and return frequently to the retail store. In addition, each segment of customer buys following a model, this was demonstrated with the a priori algorithm. Finally, some strategies are given into which products should go together and how to distribute them so that customers can find them, as well as the best-selling products.