Case Study: a leading consumer goods retailer increases approvals and reduces risk with Experian's machine learning acquisition model

A custom machine learning acquisition model for a leading appliance retailer

The client, a leading consumer goods retailer, was using a third-party logistic regression model for credit risk and sought to aggressively grow sales by increasing application approval rates without increasing portfolio risk. They partnered with Experian to develop a custom machine learning acquisition model that could deliver higher predictive power while maintaining transparency and the ability to provide adverse action reasons for declined applicants.

Experian developed a custom gradient boosting model using the XGBoost library. This solution significantly outperformed the incumbent model on key metrics (KS and Gini) by approximately 5%, as validated on out-of-time data. By implementing Experian's model, the client achieved its dual goals: they increased approval rates to drive sales and simultaneously reduced credit risk. The implementation was projected to yield a multimillion-dollar increase in net sales margin annually.