Churn Prediction Model and Segmentation in Insurance Industry

Abstract

Increasing competition in US Insurance Industry pushes companies to raise marketing costs and decrease prices to get new customers. As a result, customer acquisition costs sky rocketed, and the only way for companies to be profitable is to keep stable retention and reduce churn. Without data-driven decisions, companies struggle to reduce churn and eventually lose their momentum in a hardly competitive industry. The main goal of this paper is to analyze customer data, describe churn behavior, and develop action able recommendations to decrease customer churn. We developed a churn prediction model and segmented churnedcustomers. Segmentation combined with model results were used to develop segment-specific recommendations for the business. The business implication of this research is a churn reduction strategy designed specifically for each customer segment.