E-commerce Customer Segmentation Using Machine Learning

Abstract

E-commerce has increased exponentially over the past few years, with opportunities for growth as well as daunting tasks for businesses selling in competitive online spaces. One of the persisting challenges is how to identify and retain valuable customers whose purchasing behavior changes quickly in response to promotions, new product launches, or social media. This study responds by developing a hybrid segmentation model that incorporates Recency-Frequency-Monetary (RFM) measures with the K-Means clustering algorithm. Using transactional-level data, it construct behavioral features and apply clustering to detect patterns not normally captured by static thresholds. Four segments are revealed through analysis: high-value loyal purchasers, mid-value customers with growth potential, disengaged segments at risk for churn, and a small premium spending segment. RFM affords interpretability, and K-Means detects latent structure that yields analytical insight. Overall, the findings provide managers with concrete recommendations for loyalty programs, reactivation campaigns, and premium services, showcasing how machine learning can complement the role of traditional metrics in e-commerce.