Predicted churn using limited internal historical data

Churn predictor

A large insurance client faced an increasing cancellation rate of insurance policies and was looking for ways to improve their sales effectiveness and retain their customers. The aim was to predict which individual customers are likely to cancel policies 90 days in advance and for one or more of their policies

The challenge

How PA helped

PA developed and calibrated a model that predicts each month customers likely to cancel their policies for one or more product lines. The model has a 83% success rate of predicting cancellations

Benefit for the clients

As a result each month the salesforce are fed with information on exactly where they need to focus their efforts to retain customers