Predictive Modeling of Customer Churn in the Home-to-Home Program Industry: A Time-to-Churn Analysis.

Customer churn prediction is a critical area of study in industries where long-term relationships significantly impact profitability, such as the Home-to-Home Program industry. This thesis explores the development of predictive models and survival analysis techniques to identify at-risk customers and understand the timing of their potential churn. By leveraging machine learning models alongside survival analysis methods like the Cox Proportional Hazards model, this research provides a comprehensive framework for understanding and mitigating churn. The results demonstrate the importance of customer tenure and behavioral patterns in predicting churn, offering actionable insights for improving retention strategies. The study highlights the broader applicability of these methods across industries, emphasizing their potential to enhance customer satisfaction, optimize resource allocation, and drive long-term business success.