The Challenge

Accept more card applicants and keep risk losses stable

A top 10 small business card issuer wanted to gain small business card market share in 2021. They believed two things held them back: low approval rates and unfavorable terms. Improving their underwriting models would address both concerns. And to do that, they’d need more robust data — beyond what they pulled from credit bureaus 
and the Small Business Financial Exchange.

The card issuer believed they were overlooking many eligible small business card applicants, simply because they didn’t have enough data about the business. New data segmenting healthy and unhealthy businesses would improve their underwriting models so they could increase acceptance rates while keeping risk losses stable.

Ideal state: the card issuer could identify the healthy businesses their competitors might overlook or deem too high risk.

The Solution

Improve underwriting models with data to better predict delinquency events and spend amounts

To get a more holistic view of applicants’ financial health, the card issuer introduced Enigma’s Merchant Transaction Signals data into their models to backtest for signal — specifically card revenues, revenue growth rates, transaction stability, and average transaction size.

Results

The team found:

• The accuracy of their models increased by around 30% for the target population. Card revenue data improved their modeling of spend amount, while transaction and growth rate data improved delinquency prediction.
• The new data enabled them to increase their approval rate by roughly 7 percentage points.
• Within the first twelve months of implementation, they expect to generate around $7.5 million in new revenue with these improvements to their underwriting model.