Customer:
The customer is a telecom company.
Expectation:
Churn is a common problem faced for telecom or any subscriber-based industry. It is very challenging to win lost customers. So, the customer was looking to build a churn prediction system that can add help retain and grow the customer base.
Solution :
As part of this engagement, we developed the solution by utilizing AI/ML techniques to predict churn probability for each customer.
We implemented XGBoost models for binary segmentation of subscriber base. Probability scores for the subscriber base were predicted, representing chances to churn in the next three months.
We used subscriber call records, data usage history, call center interactions, subscriber metadata while developing the solution.
Results :
The solution resulted in better retention programs based on these models and resulted in a 5% reduction in overall churn.