Data-Driven Churn Prediction: Evaluating ML Models for Business Retention Strategies

Abstract

Customer churn, which is the loss of customers to a service, is a major threat to business profits, especially in competitive industries like telecommunications, finance, and e-commerce. Identification of at-risk customers early enough helps organizations to take proactive retention efforts, thus minimizing loss of revenues. The research study will focus on forecasting customer churn through the application of a machine learning (ML) algorithm to provide strategies to enhance customer retention strategies. A comparative analysis was done using five ML models: logistic regression, K-Nearest Neighbors, support vector machine (SVM), decision tree, and random forest (RF). To have a better understanding of the performance of the models, several metrics were applied, namely, accuracy, sensitivity (recall), specificity, precision, and F1 score. In terms of performance, RF showed the highest performance and the best accuracy (80.74%) and F1 score (61.34%), and the specificity and precision were high. SVM also demonstrated stable results, especially in sensitivity, which makes it appropriate for detecting churn-prone customers. The results indicate the advantage of ensemble algorithms such as RF because of their strong nature and generalization on complex data forms. This study addresses existing research gaps by standardizing data preprocessing, evaluation metrics, and model benchmarking, while incorporating multiple performance indicators beyond accuracy to effectively handle class imbalance and real-world decision-making trade-offs. The study further provides insights into model interpretability, deployment feasibility, and computational cost, enabling organizations to select churn prediction models based on both predictive performance and practical applicability.



Received: 19 July 2025 | Revised: 28 October 2025 | Accepted: 6 January 2026



Conflicts of Interest

The authors declare that they have no conflicts of interest to this work.



Data Availability Statement

The data that support the findings of this study are openly available in Kaggle at https://www.kaggle.com/datasets/blastchar/telco-customer-churn/data.



Author Contribution Statement

Ramshankar Nagarajan: Methodology, Software, Validation, Resources, Visualization. Siva Subramanian Raju: Conceptualization, Formal analysis, Investigation, Data curation, Writing – original draft, Writing – review & editing, Supervision, Project administration. Athiraja Atheeswaran: Methodology, Software, Validation, Visualization.