Customer Behavior Profiling in Wholesale Retail Using RFM Analysis and K-Means Clustering

Customer segmentation is a crucial strategy in data-driven marketing and customer relationship management, particularly in wholesale retail environments. This study integrates Recency, Frequency, and Monetary (RFM) analysis with the K-Means clustering algorithm to uncover meaningful customer groups using transactional data from 2022 to 2024. After computing RFM values and applying <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$z$</tex>-score normalization, the optimal number of clusters was determined using the Elbow Method and validated through internal metrics including the Silhouette Score and BSS/TSS ratio. The resulting five-cluster model achieved a Silhouette Score of 0.52 and a BSS/TSS ratio of 0.80, indicating strong intra-cluster cohesion and inter-cluster separation. The identified segments-Platinum, Gold, Silver, Bronze, and Bad-exhibited distinct behavioral patterns that support actionable marketing strategies. Visualizations such as PCA plots and RFM distributions confirmed the robustness of the segmentation. The findings provide a scalable approach to customer profiling in retail and serve as a foundation for personalized marketing, loyalty programs, and reactivation campaigns. This study was conducted using Python and libraries such as pandas, scikit-learn, and matplotlib. A limitation of the study lies in the moderate silhouette score, which indicates room for improvement in future clustering refinements.