Back to success stories

Identifying Those Most “Likely to Pay”

The Challenge

A US based Auto Insurer wanted to gain a better understanding of who would have the ability to pay during a subrogation process.

The objective was to better focus their efforts in recovering claim payments.

The Solution

By combining the Powerlytics Data Platform with the Auto Carrier’s data, Powerlytics was able to build a “likely to pay” score to help identify the consumers that had the greatest propensity of payment.

Likelihood to Pay Score Development Process:

  • Appended full ZIP+4 data set to the Auto Carriers’ sample
  • Split file into “training” and “testing” segments
  • Tested multiple “likelihood to pay” models
  • Identified top model parameters and overlaid on full client sample
  • Utilized top variable outcomes to design “likely to pay” score

The Impact

The “likely to pay” score identified a meaningful group with a 75%+ likelihood to pay. As a result, the Insurer could leverage data to improve subrogated claims reserve forecasting and better focus its resources to both improve collections and reduce labor costs.