Predicting and Modeling Customer Churn in Telecom Industry: An SEM–PLS-Based Analytical Framework

SEM, PLS-Predict, 5G, Customer Churn, Telecom Industry, Predictive Analytics

India’s mobile telecommunications market, the world’s second largest, faces rapid growth and fierce competition, making customer churn reduction a strategic priority. This study develops and validates an integrated framework to identify and predict churn determinants within India’s emerging 5G ecosystem. Data from 1,600 mobile subscribers in Hyderabad were analyzed using a two-stage process involving Exploratory and Confirmatory Factor Analyses, followed by Structural Equation Modeling (SEM). Logistic regression and PLS-Predict were applied to assess predictive accuracy. Results indicate that network quality, service experience, and product value significantly influence churn, while advertisement, brand, and social influence are insignificant. Predictive testing confirmed high accuracy (AUC = 0.84) and strong relevance (Q²_predict > 0). The validated SEM–PLS model offers actionable insights for proactive churn management, emphasizing network reliability, responsive service, and value-added offerings. This study extends existing literature by modeling churn behavior through a structural–predictive approach specific to India’s 5G context.