Churn Prediction and Effective Campaign Management using Predictive Analytics

Project Description

The predictive model solution involves:

  • Churn Management: For the Churn Management, the solution built provides output in the form of churn probability scores for the subscribers. The Subscribers are sorted in descending order of churn scores to divide them into ten deciles (groups). The top 3 deciles capture 88% of churners. The model is refreshed every 7 days.
  • Segmentation Management: In case of Segmentation management, the model output divides the subscribers into 6 clusters with subscribers within the same cluster having similar characteristics. The subscribers identified by the models are targeted through promotion campaigns. Effectiveness and efficacy of campaigns to be determined by having control groups

Potential business benefitted from the above exercise resulted in revenue increase of around USD 1.8 Million.

Client: Large India-Africa Telecom Service Provider