Online reviews, social check-ins are tested as credit scoring data, amid skepticism
A string of rave reviews from customers online can boost your business – and may help you qualify for credit.
One of the major credit bureaus is evaluating the use of alternative data such as online customer reviews, social media check-ins and “website sophistication” to assign credit scores to small-business owners with thin or blank credit files. Experian is now pilot testing a program that would draw on such unconventional metrics to assist small, new-to-market firms with no credit footprint in getting loans and credit cards.
The credit industry has in recent years embraced alternative data to gauge the borrowing risk of consumers who lack sufficient credit history to get a traditional FICO score. Those efforts have mostly focused on bill payment information from utility and telecommunication providers and landlords, but a few firms are touting the use of social media data to evaluate prospective borrowers.
However, mixing social media with consumer credit scoring raises concerns over privacy, relevance and fairness. Similarly, some lenders and consumer advocates have reservations about scoring a small business based on what its customers say about it online.
Can one bad review make you a credit risk?According to Experian, about 20 to 40 percent of all inquiries for commercial credit come from businesses that have no credit bureau files. Eric Haller, executive vice president of Experian DataLabs, said many of those companies tend to be brand-new.
“They haven’t developed any trade relationships that are being reported yet, but they could be very viable and active businesses,” Haller said. “It begs the question, how do we serve the extension of credit to these businesses, when there’s definitely a need and a high probability that those obligations will be paid back?”
Experian’s model captures alternative data that can indicate how a new business is performing to predict its ability to pay back loans. For instance, if lots of customers “check in” on social media while dining at a recently opened cafe, it suggests a high volume of foot traffic, and thus, revenue. Similarly, a steady stream of product shipments or positive online reviews from happy customers can be signs of success.
“It’s basically taking another crack at understanding the health of a business – whether it’s growing and thriving, or it’s shrinking and hitting some real challenges,” Haller said. “We’ve determined that we can extract signals that are significant and predictive enough that lenders will be confident in the results.”
Online reviews can make or break a small business, and it doesn’t take many disgruntled Yelpers to create a bad reputation. Haller stressed that a single bad review from an angry or vindictive customer can’t derail a business’s credit chances in Experian’s model.
“As we extract signals from comments, there has to be consistency,” he said. “If there isn’t a strong enough signal that’s being tracked from comments, that’s not weighted in the model.”
Haller also noted that some customer complaints suggest a business is thriving – “I had to wait an hour because the place was packed,” for example.
Skeptics question relevance, predictive value
The other major credit bureaus have not embraced the idea of scoring small businesses based on social media and other unconventional forms of alternative data.
A TransUnion spokesman said in an email the topic is “not applicable” to his company. Bob Crutchfield, vice president and commercial insight consulting leader for USIS Commercial Services at Equifax, said there is little appetite among Equifax’s clients for such a credit risk model.
“Frankly, a lot of our customers are a bit skeptical because you always have to think about the regulatory concerns,” Crutchfield said. “How would somebody dispute a credit score that had social media data in it? You have to think about those types of things.”
There are also concerns over whether social media data and customer reviews are truly predictive of risk and relevant to credit decisions. Ira Rheingold, executive director of the National Association of Consumer Advocates, said it’s preferable to review accounting data – such as revenue, expenses and capital – to assess the borrowing risk of a small business with little or no credit history.
“You’re taking this subjective information which may or may not have anything to do with creditworthiness,” Rheingold said. “I think it’s fairly risky, and that’s not the way decisions about granting credit should be made.”
Haller said Experian’s pilot model is being tested against reams of in-house historical data on other small businesses that have successfully established credit histories. He said the model likely won’t be marketed to lenders unless it has the proven potential to help at least a million businesses establish credit.
Meanwhile, lenders can test Experian’s model out themselves if they are skeptical.
“When we pitch to a lender, we have all the analysis we’ve done and we have the model we’ve already built,” he said. “We can show them how it performs and then they assess whether or not they want to take the risk with us and test it out in real life.”
The use and experimentation of alternative data is not new in the consumer or business spaces.
Equifax’s Crutchfield noted that utility, telecommunication and trade credit relationship data have been used to score businesses for years. Traditional consumer credit data can often be used to fill any remaining information gaps when scoring a business, Crutchfield said.
But Experian is testing uncharted waters. If it’s successful, the use of social media data could improve the business prospects of millions of “credit invisible” small merchants who consistently make their customers happy.
As for any business that routinely draws scathing online reviews, poor feedback could be a sign of weak overall performance and, thus, a risky borrower and potentially reduced lending opportunities.
Alternative data “gives them the opportunity to change a ‘no’ to a ‘yes,’” Haller said. “If there’s a string of negative comments reviewing a business, it may be that they’re going to get the same answer they got from traditional methods. If there are a lot of positive comments that can turn a ‘no’ into a ‘yes.’”