Customer segmentation using an extended rfm model and clustering algorithms in e-commerce

Abstract

Customer segmentation is a critical step in the efficient utilization of customer data and maximization of profitability in the e-commerce sector. Segmentation studies can yield economic benefits for firms and provide a range of insights based on customer data. This study proposes an extended RFM framework to address the shortcomings of the traditional RFM model, using customer transaction data from an e-commerce company for the period 1 January 2024–31 December 2025. The proposed framework integrates additional dimensions—campaign share, basket depth, and the standard deviation of inter-order intervals—alongside the conventional recency, frequency, and monetary values to improve segmentation. Subsequently, several clustering techniques employed for segmentation, including K-means, K-medoids, and fuzzy C-means, were considered. To determine the optimal number of clusters and assess the model fit, the results of the algorithms were evaluated using a quality index computed with multiple indices, such as the Silhouette, Dunn, and Davies–Bouldin indices. The proposed extended RFM model extends the traditional RFM framework by integrating additional behavioral dimensions such as price sensitivity, shopping regularity, and basket depth. This enriched representation of customer behavior allows for more discriminative and actionable segmentation, thereby enhancing target customer identification and enabling more precise product recommendation strategies.