Churn prediction model for housing insurance in Portugal

Abstract

This paper presents a predictive churn model for a Portuguese insurance company. The model is designed to provide insights into the churn behaviour of customers with housing policies, and to predict the probability of churn in the period around the renewal date of a policy for a specific policyholder. The model will be used in a proactive retention marketing campaign, and will be employed as a simulator for the pricing team of the insurance company. The goal of the model is to improve retention and to deploy demand-based pricing strategies.