Predictive Churn Modeling: Proactive Retention

• The Challenge: Identifying "at-risk" customers before they leave, especially in imbalanced datasets where churn is a rare but costly event.

• The Solution: Developed a classification model (Random Forest/Logistic Regression) focusing on Precision and Recall rather than simple accuracy.

• The Impact: Isolated churn precursors (e.g., low usage, lack of tech support sign-ups), allowing marketing teams to intervene and protect revenue.