Feature Story

Online lenders harvest big data to extend loans where banks cannot

by Amy Cortese

online lending

Welcome to the brave new world of online lending, where certain factors may someday be more important than your credit score.

Online sellers that ship to California customers make better borrowers.

People who use Apple products are a good credit risk.

Businesses with active Facebook pages are less likely to default on their loans.

And loans seekers using first-generation email accounts raise red flags.

Welcome to the brave new world of online lending, where factors such as your social network, what browser you use or your shipping patterns may someday be more important than your credit score.

Individuals and small business borrowers have been flocking to a new crop of online lenders that promise cash in the bank, fast. In some cases, the borrowers have already been turned down by a bank. But increasingly, they are bypassing banks altogether and going straight online.

On Kabbage.com, an online lender that makes short-term loans to small businesses, the entire application process takes as little as 7 minutes.

The secret?

Big data. Rather than green-shaded bankers, online upstarts like Kabbage, OnDeck and others employ data scientists who crunch hundreds or thousands of data sources to assess whether a person or a business is a good credit risk. By looking at current and realtime data, rather than solely backward-looking credit histories, they are able to make loans that banks, which are heavily regulated and tethered to legacy IT systems, cannot. In other cases, the loan amounts are too small for a bank to be interested.

Alternative lenders approve around 60 percent of small business loans, compared to just 20 percent or so approval rates for big banks, according to Biz2Credit, a site that matches small business borrowers with lenders.

See Also: Data virtualization for better decision making

“We’re serving a market that was not being served,” says Kathryn Petralia, head of operations at Kabbage, which has funded over $1billion in small business loans to date, in amounts as low as $2,000.

Like other online lenders, Kabbage scours a wide range of untraditional sources of data for clues to creditworthiness. Through a partnership with UPS, for example, Kabbage has found that shipping patterns can be a good indicator of credit risk. In fact, a Kabbage underwriting model using only shipping data outperformed one that relied on FICO, the credit score that has been the gold standard among bank lenders for decades.

While factors like social media and shipping patterns are predictive—and can be as or more effective than a FICO score—real-time financial information is the real bread and butter for online lenders. Borrowers typically give online lenders access to their bank account and other financial data, such as QuickBooks. That allows the lenders to validate and monitor a company’s sales, in real-time, and deduct loan payments on a daily or weekly basis.

Kabbage’s interest rates are high—the average APR equivalent is in the low 40 percent range, says Petralia. But it loans to businesses that might not otherwise be able to access any credit at all.

Similarly, other online lenders are using Big Data to lend to people with tarnished credit or “thin files,” meaning they have not built up a credit record. The Consumer Financial Protection Bureau estimates that some 26 million Americans—or one in ten adults—are “credit invisible.” That leaves many people without access to credit—and few traditional lenders willing to take a chance on them.

“We score thousands of data sources,” says Mike Armstrong, marketing chief at Zest Finance. “We use that technology to make loans to hardworking Americans overlooked by traditional lenders.”

Some of the biggest experiments are taking place overseas, where consumer lending and credit is less developed. New York-based Lenddo has built an underwriting algorithm based on its experience making loans using social media indicators to people around the globe that have traditionally lacked access to credit. (In the U.S., many social media sites have policies that prevent lenders from using their data to make consumer credit decisions, so for now it is typically used in small business underwriting or to supplement credit decisions). Lenddo plans to license its algorithm to financial institutions and lenders.

And Jumo, a South African startup, uses cell phone data to lend to thin file customers there.

Even the established consumer credit rating agencies are beginning to take note. TransUnion, one of the big three consumer credit rating agencies, has created a new product called CreditVision that analyzes a broader range of data than its traditional system, such as how often an applicant’s address changes, checking account history, and even data from payday lenders.

For all the promise, however, data models have not yet completely replaced human underwriters who are trained to assess risk, at least not for small business loans. Candace Klein, chief strategy officer at Dealstruck, a San Diego-based online small business lender, says that relying solely on big data for loan approvals can be risky. “Big data modeling should be complementary to loan underwriting, rather than a replacement.”

# # #

The contents or opinions in this feature are independent and may not necessarily represent the views of Cisco. They are offered in an effort to encourage continuing conversations on a broad range of innovative technology subjects. We welcome your comments and engagement.

We welcome the re-use, republication, and distribution of "The Network" content. Please credit us with the following information: Used with the permission of http://thenetwork.cisco.com/.