Predicting E-commerce Churn and Boosting LTV

This working paper presents a machine learning framework for predicting customer churn in e-commerce and quantifying its financial impact on Customer Lifetime Value (LTV). Using publicly available transactional datasets, customer-level behavioral features were engineered using RFM methodology and extended engagement metrics. Multiple classification models (Logistic Regression, Random Forest, XGBoost) and survival analysis methods were evaluated to predict churn probability and customer lifespan. The study translates predictive performance into measurable business impact through LTV simulation and retention intervention modeling. The objective is to connect predictive modeling directly to revenue protection and long-term margin efficiency. Author: Hanzel LacidaVersion: 1.0Date: February 27, 2026