International Research Journal of Education and Technology

The exponential growth of e-commerce platforms has resulted in the generation of vast volumes of transactional and behavioral customer data. While this data holds significant potential for strategic decision-making, many organizations struggle to extract meaningful insights regarding customer purchasing behavior, value contribution, and engagement patterns. Traditional mass-marketing strategies are increasingly ineffective in competitive digital markets, necessitating data-driven customer segmentation approaches. This study proposes a machine learning-based customer segmentation framework for e-commerce platforms using Recency-Frequency-Monetary (RFM) analysis combined with K-Means clustering.

The proposed system preprocesses transactional data through data cleaning, transformation, normalization, and feature engineering techniques to ensure analytical integrity. RFM metrics are computed to capture temporal and monetary purchasing behaviors, which are then utilized as input features for clustering. The optimal number of clusters is determined using evaluation techniques such as the Elbow Method and Silhouette Score analysis. The resulting customer segments provide actionable insights into high-value customers, loyal customers, potential churn segments, and low- engagement users.