This document discusses using predictive modeling to increase customer retention for an insurance company. It compares using survival analysis versus true lift modeling. Survival analysis predicts when customers will leave and identifies those at highest risk of non-renewal. True lift modeling identifies which customers are most likely to respond to a retention effort so the company can focus its efforts. The insurance company tested both approaches and found true lift modeling was better for targeting customers they could directly influence through agent meetings. The key takeaway is to select the right model based on the specific business question being asked.