Churn Prediction for a global wireless communications service

Virgin Mobile aims to predict churn for their prepaid customers based on their activity patterns and high-level profile information. For prepaid customers, business was more interested in predicting a brief period of inactivity rather than a user not returning to their service or products. Instead of a traditional churn prediction, we were looking at predicting a temporary soft churn where the customer is likely to return to the services and products but is inactive for a long enough time to impact business.

Read More