LRFM-based customer segmentation using K-means clustering: A data-driven study in the telecom industry

In competitive digital service environments, an accurate understanding of customer behavior and value has become a fundamental requirement for effective marketing decision-making and customer relationship management. The telecommunications industry, due to its multi-channel nature, high transaction volumes, and diverse customer interaction patterns, faces significant challenges in achieving meaningful customer segmentation. The objective of this study is to propose a data-driven framework for customer segmentation in the telecommunications industry using extended LRFM indicators and unsupervised clustering methods. To this end, real customer data, including transactional, behavioral, and loyalty-related information, were first collected and preprocessed. Subsequently, a set of key features was extracted, and the four indicators of Length, Recency, Frequency, and Monetary were calculated for each customer. These indicators were then used as inputs to the k-means algorithm for customer clustering. To determine the optimal number of clusters, the Elbow method and the Silhouette index were employed, and customers were ultimately classified into several distinct clusters. Analysis of cluster profiles revealed significant differences among customers in terms of economic value, level of engagement, relationship stability, and likelihood of churn. The results enabled the identification of segments such as golden customers, valuable loyal customers, low-activity customers, and customers at risk of churn. The findings demonstrate that integrating LRFM indicators with clustering can serve as an effective tool for designing targeted marketing strategies, optimizing resource allocation, and improving managerial decision-making in the telecommunications industry. In addition to its interpretability, the proposed framework is extensible to customer behavior prediction models.