In the prior blog post, Finding good customers amongst bad credit scores, we talked about using alternative data to identify better performing customers. Specifically, customers with subprime credit scores. In the post, we described the results of using a Scorenomics scorecard with BeyondMyScore data:

A risk rated 1 account is a meaningfully better risk than is a risk rated 3 account. Note the average FICO scores displayed for each risk rating. They are all similar, in the 625 range. The Scorenomics Risk Rating is providing new information beyond the credit bureau score, which helps segment seemingly similar customers.

Those results are for low-scoring customers who had already been approved. How about for applicants who were marginally declined for credit? Could it segment them too?

Marginal decline: just missing the cut

Lenders use underwriting criteria to determine who they will approve for new credit. The output of the criteria is typically a score that indicates likelihood to default. These scores can be sorted from highest to lowest. Within that ordering, there will be a cutoff where the predicted risk of default exceeds the projected profitability for an account.

It looks something like this:

One thing you should notice. Not all declines are the same. There are “firm declines” and “marginal declines”. A firmly declined applicant presents meaningfully negative characteristics, and there is no chance for a change in decision. A marginally declined applicant has some positives, but enough negatives to outweigh them, so no approval. The applicant’s underwriting score fell just short.

But might there be a way to find the diamonds in the rough?

Scorenomics alternative data to find approvals among marginal declines

One of our clients – a personal loan provider – ran a pilot to determine how Scorenomics could help them identify better risks. Marginally declined applicants were offered a “second look” via BeyondMyScore. By completing the BeyondMyScore financial health platform, they would automatically be approved. I don’t suggest this is an optimal way to treat all marginally declined applicants long-term. But for a limited pilot to determine efficacy of identifying better risks, it was a smart play.

The personal loan provider tracked the performance of the customers approved under this Scorenomics second look program. Our client did provide an early snapshot of their payment performance: performance on up to the first six installments for each customer. Most customers were up-to-date, but a decent number had missed a payment already.

We applied that scorecard we had previously developed to these marginally declined BeyondMyScore completers. Would it also work on this new population? The chart below shows how the risk ratings segmented on payment performance.

The model proved itself on this marginally declined population. In the initial set of installments, it was successfully segmenting on payment risk. Those numbers you see at the bottom of each bar – in the 590’s – are the average VantageScores for the applicants. The average VantageScores are relatively similar for all three Scorenomics risk ratings. BeyondMyScore’s alternative data provided new information beyond what is available in an applicants credit report.

Note: our client developed its own scorecard with the BeyondMyScore data. After several months of monitoring the performance of the customers, the client deemed the pilot a success and rolled it out for all marginally declined customers.

Click here to find out how Scorenomics BeyondMyScore^® can help you find profitable growth among your marginally declined customers.

Signup to be notified when we’ve posted a new article to the Scorenomics Blog!