Machine learning-based customer segmentation: advances and applications in retail

While machine learning (ML) is widely used for customer segmentation in e-commerce, its application in offline retail remains challenging due to limited data. This study investigates the efficacy of ML-based segmentation using two years of purchase data from a major department store, comparing traditional methods (Monetary, RFM) with three ML clustering algorithms. Results demonstrate that ML-based methods, particularly Hierarchical Clustering, significantly outperform traditional approaches in creating distinct customer segments. Our findings provide strong evidence that ML techniques can offer significant advantages for customer relationship management, even in data-scarce offline environments.