Churn Prediction

Home - Solution - Churn Prediction

Churn Prediction

Our clients are only using our product short-term. We have a credit-line and want to increase the utilization period.
Client’s CEO

Client Need

Our client – a credit card issuer operating in the sub-prime loans domain in Eastern Europe reached plateau for their acquisition phase. Even though the portfolio is now rich as number of active customers these use the product only to meet their current need and leave shortly. A strategy to control the attrition rate is needed.

Our Approach

Acquiring new customers is more expensive than retaining the current ones. We have developed a Churn behavioural prediction model that will significantly increase the knowledge of which credit card holders are about to leave the company.

The Solution

A full behavioural churn profile was achieved with the scoring model. Along with other metrics a CRM solution was developed to target and retain customers with tailored proactive offers.

The Model

  • Great discriminative performance;

  • Enriched and reliable customer’s profile.

The Campaign

  • Based on the scorecard;

  • Enhanced with other metrics like time since first debt and current customer lifetime value.


  • Able to differentiate two significant groups that are 22 and 8 times more likely to churn than the average;

  • Multi-segment targeted campaign;

  • The knowledge was integrated into other CRM campaigns as well;

  • The campaign outperformed the costs for development by 3 times just in the first 6 months;

  • 30% better retention rate while AB testing live.

Snapshots of our solution in action

Request Your Demo

Unique Ideas Grow Business