Identifying & Predicting customers who have a high probability of churn using Machine Learning
The client is one of the leading healthcare providers in North America. The company has been a recipient of Canada’s Best Managed Companies award since 2006.With over 100 locations across the country. It has more than 13,500 staff members and provides care to over 350,000 clients.
The company wanted to understand the characteristics of churned customers and the parameters that led to it. They were looking for a model that would predict the churn probability of their new enquiries. The churn predictor model had to be integrated their CRM tool which would help them to understand the probability of churn when a new customer acquisition happens in a particular branch.
Both in house transactional and external demographic historic data used to develop machine learning algorithm, which helped us to understand the various parameters that led to customer churn and create a churn probability predictor. This was integrated with their CRM tool. When a new customer details are entered in client’s CRM, the predictive ML based algorithm works at the back end and predicts the probability of churn of that customer in a pop up window. For high Churn probability, the front office executive offered different attractive packages, discounts and special offers to retain the customer .
Result & Value Adds
- Increased sales conversion by 30%
- Increased cross selling by 13%
- Reduced churn by 13% during 8 months period.
- Increased customer loyalty by 15% during 8 month period due to the great customer experience