Grow the book
without growing the losses.

The risk team assigns a risk score to every business in the portfolio and uses it to sort customers into three buckets: decrease, maintain, or increase credit lines. The score is refreshed each month, so accuracy compounds.

Its models relied on internal data plus bureaus and the Small Business Financial Exchange. Through 2020, volatility in the small business portfolio exposed the gap — without clear visibility into portfolio health, the team was missing opportunities and taking on unnecessary risk.

What it wanted was more timely data carrying signals about a business's financial health: leading indicators of decline and distress, better coverage of its own SMB portfolio, and richer data about growth and revenues.

The team introduced Enigma's merchant transaction data into its models and back-tested against historical data rather than deploying on faith.

Three signals did most of the work: presence of transactions, transaction stability, and three-month growth rates. Those produced the strongest increase in the models' predictive power.

More accurate models meant better customer segmentation — which cuts both ways, reducing risk and uncovering opportunity in the same pass.

Enigma's data increased model accuracy by around 25% for about a third of the portfolio.

Credit line increases made up a large share of the team's new revenue generation, and the improved models identified around 5% of the portfolio as eligible for incremental increases — along with more than 70,000 new high-value businesses.

During the first year of implementation, the team expects those adjustments to generate $30,000,000 in incremental revenue.