How does the concept of behavioral scoring apply to SMB lending?
This is the last blog of our 3 Part Series – “Ask the Data Science Guru”. The two blogs prior featured musings from the JUDI.AI credit science team, focused on the following questions:
Today, our credit science team explains the concept of a behavioral score in the context of SMB lending. Here’s what they had to say on the topic:
A behavior-based score in small business lending puts a grade on business conduct and habits. This score is often used to supplement credit scores obtained from traditional credit bureaus and is meant to factor in alternative data points that can help determine creditworthiness and paint a more robust, current picture of SMB financial health. Behavior scores can also help identify changing risk in time to take corrective actions if needed.
Some major challenges when it comes to small business lending is the low SMB credit score hit rates and the reliance on financial statements that tend to be stale and inaccurate. Reliance on traditional credit scores alone means missed opportunity to attract new borrowers and help existing small business customers. To lend more without taking on massive risk requires you to account for SMB behavioral fluctuations due to swings in revenue, costs, margins, demand drivers, competition, and seasonality. The next AI chapter is all about adding the intake of other relevant and vertical-specific data, creating adaptive credit risk assessment models, and teasing out a behavior score that really matters.
Behavior scoring algorithms have been around for a long time. What is important to think about are the variables that go into your model, and how often it is updated. With real-time cash flow data derived from banking transactions, you are sitting on a data insights gold mine.
Let’s consider variables outside of Beacon Scores, BNI and Business Scores that reflect small business behavior. Judi.ai has validated dozens of powerful risk variables based on cash flow history, finding indicators is a variety of behaviors:
- The debt service coverage ratio (DSCR)
- Increasing trend of NSF transactions
- Changes in cash flow drivers (credits and debits)
In JUDI.AI’s proprietary behavior scoring system we have weighted the importance of certain cash flow metrics based on 25+ million data points ingested and over $1Billion SMB loan applications reviewed. Performance insights are based on a minimum of 3 months of activity prior to decisioning, and a minimum of 6 months after lending. As our SMB lending consortium expands, new data sources are added and changes in the lending environment are experienced, our AI and machine learning models will adapt, and the library of cash flow variables will expand or contract.
If you are interested in exploring a new generation of small business scoring, JUDI.AI’s credit science team would love to start a discussion. Alternatively, you can request a demo of JUDI.AI’s SMB lending platform below.