Industry and Services

Customer churn prediction system

The Challenge

Our client, a health insurance company with more than 3 million customers, is facing the need to reduce the churn rate at the end of each policy term.

| The Outcome

Our predictive analytics model anticipated the risk of policyholder churn and helped our client retain them through targeted renewal campaigns.

How we did it

• Implementation of a predictive analytics model which analyzes historical customer, transaction and churn data.
• Detection of customers with a high probability of abandonment in the following 3 months.
• Integration of the model in a web application that allows retraining and obtaining statistical and inferential information from the updated data set.
  • Operating on different data sources: policies, benefits, receipts and withholdings.
  • Model development, training, validation and testing are implemented.
  • Threshold, statistics, inferences and dropout scoring system are set.
  • Built on H2O.ai, R and Shiny architecture.

Increased detection of customers at risk of churn

Reduction of the churn rate

Shall We Talk?

If you need to know more specifics or are interested in having us assist your organization, please use the form below.