<![CDATA[Effective customer segmentation is crucial for online retailers to enhance marketing strategies and boost profitability. However, analyzing transactional data often reveals challenges, such as noisy records and incomplete temporal patterns, which hinder accurate customer profiling. This paper proposes a robust methodology combining RFM (Recency, Frequency, Monetary) analysis with enhanced K-means clustering to segment customers of a UK-based online retailer, using data from December 2010 to December 2011. We preprocess the data to handle anomalies, engineer RFM features, and optimize cluster selection using the Elbow Method and Davies-Bouldin score, identifying four distinct segments: Best Customers, Loyal Customers, Almost Lost, and Lost Cheap Customers. Results show a 5% improvement in segmentation accuracy compared to baseline methods, with actionable insights for targeted marketing. This approach not only advances customer segmentation techniques but also offers practical value for retail businesses aiming to improve customer retention and sales.]]>
Copyrights © 2025
