Behavioral Scoring in Collections
December 6, 2021 | By Rick Haskell
Ever feel like collection scores are the red-headed step child of scoring models? For some reason, everyone wants to talk about origination models (called application scorecards), but collection models (behavioral scorecards) don’t often get the credit they deserve. In fact, this whole idea of precisely identifying high and low risk accounts that are already delinquent can be bewildering. I mean, once an account goes past due, collection efforts will begin regardless, right? Well if your institution subscribes to the “blanket every last account with the maximum number of contact attempts allowed by the regulators” ideology, you are wasting valuable time and resources.
What if it was possible to get a similar, or even better collections result with significantly fewer collections staff? Modern tools like AI and ML help you work smart, not hard—and working smart means more profits to your bottom line. Risk management in collections is largely about budget optimization. ML tools like behavioral scoring models are what makes optimization possible.
One of the best uses of behavioral scores is to group accounts into categories based on their likelihood of default. At Lendisoft, each delinquent account is assigned a letter grade of A through E, where A is the highest risk group.
From there, you can create any number of calling schedules using our Champion/Challenger Campaign Builder.
Here’s one example:
First, notice how even account numbers are assigned to the Champion segment, and odds to the Challenger. Next, notice how the Champion Schedule slots all accounts regardless of their Risk Grade to appear in the work queues when they become 6 or more days past due. Also notice how the Challenger Schedule slots accounts using a staggered approach based on their Risk Grade (A’s have the most aggressive schedule, appearing in work queues at only 3 days past due).
After a full month of collection efforts, results are summarized automatically to determine the winning strategy using a net profitability approach. Essentially, Lendisoft determines how much staff is required to support each segment given your portfolio size. Next, we measure the performance result of the 30-day campaign based on real results, and estimate each segment’s forecasted cumulative net loss. In the end we identify which segment was the most cost effective and that’s the winning strategy.
Behavioral scores are also used in deficiency collections, but in reverse. While an “A” Risk Grade in active collections indicates a high likelihood of default (a negative outcome), in deficiency collections an “A” indicates a high likelihood of collectability (a positive outcome)—and having this grade can be a great tool to have when making important decisions in your daily operations.
Behavioral Scores also unlock other possibilities for optimization. For example, Risk Grades can be used to assign your toughest accounts to your top collectors (or vice versa). You can experiment with virtually endless possibilities using our champion/challenger framework, and doing so will make you more efficient over time.
Remember, the most optimal and efficient strategies come through experimentation. A/B testing requires the right tools and the right framework, and Lendisoft brings all of this to your front door. If you are looking to modernize your operations with cutting-edge risk management tools, look no further than Lendisoft. Schedule a demo today!
Rick Haskell is the Founder & Chief Operations Officer at Lendisoft, a SaaS technologies company looking to disrupt the lending industry with its unique blend of enterprise software, risk management tools, and risk consulting services.