Intelligent Customer Segmentation in Digital Commerce Using K-Means Clustering

Abstract

The rapid rise in e-commerce has forced companies to have good knowledge of customer behavior and tailor the marketing strategies accordingly. This study discusses the appropriateness of the K-Means algorithm for customer segmentation from behavioral and demographic data obtained by a systematic Likert-scale survey. Clusters with high interpretability were obtained and validated through silhouette analysis with values up to 0.75, indicating high internal consistency. Key findings show that female interviewees prefer shopping by mobile to a far greater extent than male interviewees, while male interviewees are more responsive to promotional emails and SMS. Younger and middle-aged users are similarly more susceptible to social media advertising, with older segments having more neutral or selective orientations. These results illustrate the complexity of customers' behavior and that demographic and behavioral data should be combined in segmentation studies. By its demonstration of the value of clustering techniques in providing insightful customer profiles, this study contributes to practical and methodological applications to data-driven decision- making in e-commerce. Future research is encouraged to expand the dataset size and incorporate more advanced methods such as predictive modeling and sentiment analysis to further improve segmentation precision.