scorecardresearch

Could Yelp indicate whether your business deserves a loan?

New York-based JP Morgan Chase is incorporating data compiled online to offer quick small-business loans of up to $250,000.
New York-based JP Morgan Chase is incorporating data compiled online to offer quick small-business loans of up to $250,000.(TIMOTHY A. CLARY/AFP/Getty Images/File)

Online lenders and startups have been mining Yelp and Facebook for several years, checking out how many followers a business has and whether loan applicants are posting photos of extravagant vacations, to discern whether potential borrowers are a good risk.

But now, brick-and-mortar banks are stepping into the world of social media monitoring, checking these sites and other nontraditional data sources, such as utility records and Amazon.com payment history, for financial clues in making loan decisions.

Banks are turning to social media surveillance as they go head-to-head against their more nimble online competitors, who promise easier access to loans and quick decisions.

Advertisement



“Social media is definitely something we have to grow with and something we have to take seriously,” said Joe Panunto, senior vice president of financial services sales at Dun & Bradstreet Inc., which provides credit information about small businesses to banks. The company incorporates information from its monitoring of sites such as Facebook, Instagram, and Twitter into verification reports about a business that banks can use for their loan decisions. “Ultimately, that’s what this is about, finding new customers or finding out more about existing customers to service them better.”

In April, New York-based JP Morgan Chase, the country’s largest bank by assets, rolled out its partnership with OnDeck Capital Inc. to offer quick small-business loans of up to $250,000. Chase is using its own criteria for approving loans, but is incorporating the data OnDeck collects on borrowers, including social media information. OnDeck officials said that information gleaned from social media is one of dozens of pieces of information that go into developing the company’s credit profile.

When Eastern Bank launched a fast, online loan service for small businesses earlier this year, it needed a quick way to help verify companies, understand their potential revenues, and spot red flags. The largest community bank in Massachusetts decided to scour social media and online data, including legal filings, to better understand the small businesses looking for a loan of under $100,000, said Dan O’Malley, head of Eastern Labs, the technology incubator within the bank that designed the express loan.

Advertisement



Businesses with plenty of positive Yelp reviews or a strong social media following could hint at a reliable revenue stream. And that, along with more traditional information from income statements, can be a good indicator whether the borrower will repay the loan, he said.

Eastern’s computer algorithms and searches have proven to be more effective than humans, who in the past would rely on conversations with a borrower and a potential site visit in verifying a loan applicant’s business, O’Malley said.

“As more and more data is available online to use, we need to use it in responsible ways,” he said.

Eastern hasn’t rejected any customers based solely on their online reviews, but the information is helping broaden the bank’s understanding of its customers, he said.

Poor online reviews, along with less-than-ideal back account balances, would likely prompt Eastern to ask a potential borrower for more data, such as tax returns, before signing off on a loan, O’Malley said.

The change in the credit landscape has prompted even Fair Isaac Corp., developer of the FICO score, to tweak its formula.

The firm has introduced a new scoring system to evalute the creditworthiness of some 53 million US consumers who don’t have a banking or credit history. Many are young and haven’t built a payment history or are older and haven’t opened up a new account in years. FICO is working with several large banks to test this new model, which also pulls data from cable television and telephone bills and public property records, said David Shellenberger, a senior director of the firm.

Advertisement



And FICO hasn’t ruled out social media data searches in the future.

“We continue to look and evaluate opportunities from a variety of different data sources,” Shellenberger said.

Alternative small business lenders such as Atlanta-based Kabbage Inc. and OnDeck have been more aggressive about tapping social media postings and nontraditional sources, such as Amazon.com payments, to make decisions about whether to approve a loan.

But traditional banks have been more reluctant to trade established credit checks, such as FICO scores, for Facebook.

Concerns over running afoul of regulatory requirements, especially on consumer loans such as mortgages, credit cards, and auto loans, have kept many banks from using social media searches for loan decisions.

Federal regulators have warned financial institutions that if they are relying on social media data to make credit decisions, they must ensure that they aren’t discriminating against borrowers. And social media companies that allow their data to be scraped for lending decisions could be more strictly regulated as credit-reporting agencies, the Federal Trade Commission has suggested.

Facebook last year limited access to its user data by third-party applications, in response to privacy concerns.

Banks and credit unions will likely start incorporating the new data from social media and other sources into their loan approval models, instead of abandoning their old ways entirely, said Dave Buerger, president of CUneXus Solutions Inc., a California-based technology company that is helping lenders, including Digital Federal Credit Union in Marlborough, introduce faster, online auto loans.

Advertisement



The company’s technology can incorporate social media and online information about a consumer into the loan review process, Buerger said.

But banks are still trying to determine whether these new data models will prove reliable in predicting whether a borrower will pay back a loan, especially during an economic downturn, he said.

“I think there will be an evolution,” he said. “Everybody is testing new things.”


Deirdre Fernandes can be reached at deirdre.fernandes@globe.com. Follow her on Twitter @fernandesglobe.