3 Measures of Dealership Risk for Auto Finance Companies | Auto Finance News | Auto Finance News

3 Measures of Dealership Risk for Auto Finance Companies

Why isn’t there a Fico score for measuring dealerships? Auto finance companies use several different tools like D&B, LexisNexis, and AutoCount to measure dealerships. Sophisticated companies go further and build their own scorecards. While figuring out which dealerships are healthy and reliable is a common goal, the industry approaches this problem in a very fractured manner.

Clearly there is demand to rate dealers consistently. If there was a benchmark score that predicted credit risk, or related “out of business” risk, it would be very valuable. General Forensics is working it. In the charts below, they share simple practical signals you can start using today. General Forensics argues the lift you will see from these signals will drive interest in their comprehensive multi-factored risk model and the other dealership and collateral monitoring services they offer.

Consider the following signals for dealership management without getting bogged down in he complexity of machine learning:

Signal No. 1: Lot Size

Dealer lot size is a significant predictor of risk (discussed in this article from May 2017). Large dealerships are historically 23% more likely to survive the next nine months than smaller dealerships. The result aligns with intuition, but quantifying it as shown here allows improving operational policy. The chart conveys the degree that business size impacts risk. Commensurately, how much it should impact your engagement models and anticipated customer lifetime value. The chart does not say to avoid a small dealership! There is a lot of value in doing business with them; many are healthy and growing and present volume opportunity. However, it’s important to capture the differential risk of size similar to how you capture the differential risk of consumers based on Fico scores.

Signal No. 2: Inventory Replacement

Inventory replacement rate describes how efficiently a dealership acquires and markets fresh inventory to replace cars that have already been sold. When the value of this measure is less than 100% the dealership is shrinking, and if the value is 0% the dealership is not acquiring any new inventory at all. This chart describes how replacement rate impacts survival risk. It shows dealerships with less than a 25% replacement rate have an 18% higher risk of not surviving nine months. Did you expect the relation would be steeper? This specific chart doesn’t correct for seasonality fluctuations in acquisition policy. Regardless, if replacement rate remains low for several consecutive months, the risk of survival is increased as the dealership is shrinking towards 0 cars on lot.

Signal No. 3: Turnover

Turnover captures how much inventory a dealer sells in a month as a percentage of total inventory offered for sale. The chart shows how the inventory turnover rate impacts survival risk. Those dealerships that tend to move less than 20% of their inventory in a month have a 14% higher risk of not surviving 9 more months. Recognizing this relationship (and by extension, customer future value), you can implement an operational policy that allows you to better identify dealerships that stand to do well and protect yourself from dealerships that are struggling.

More Complex Signals …

Machine-learning models have the ability to integrate many signals (i.e. features) together and part of their magic is the machine learns to account for all the complex interactions of those features. That’s why it takes qualified data scientists to help manage them! Nonetheless, this chart nicely depicts a relatively additive relation of just two features. The pattern holds up for over 20,000 dealerships analyzed in this study. You can see from the chart that using the two features “lot size” and “turnover” there is already a 30% differential in risk of survival between the left and right most bars. By adding more features and nuances, and using a machine-learning algorithm to balance out all the interactions, you can further optimize risk from dealerships and engage them more effectively.


Try using the signals above in your dealership models, including within your targeting strategy and operation policies. If you need help, General Forensics provides information services that manage dealership risk and collateral risk. They also provide professional services to test, optimize, and operate on any type of data source or system.

Josh Wortman is a data scientist at General Forensics. 

  Like This Post

Sign Up Email List