How Square turns unconventional credit data into loans for tiny businesses

Register now

Millions of microbusinesses in America find it hard to get a loan from a bank, forcing their owners to use their personal credit cards or turn to friends and family for help.

But using personal credit amplifies the level of risk signaled in their personal credit scores, which is why Square has developed its own method of evaluating which businesses it can lend to.

Altogether, Square has extended $5.5 billion of credit to 275,000 very small businesses since it founded Square Capital. Its loans range from $500 to $100,000; the average is $6,500.

“We’ve been able, using very creative means, to unlock access to capital for them so we could get them into the financial system and hopefully graduate them into banks and bigger institutions that can provide a broad array of services to these small businesses,” Jacqueline Reses, the company's lead on Square Capital, said this week at American Banker’s Small Business Banking Conference in Los Angeles.

Borrowers include owners of small coffee shops, food trucks, and one-chair hair salons.

“At Square, we have seen hundreds of thousands of examples of small businesses all across America who need a couple hundred dollars or a couple thousand dollars to tide them over in a pinch,” Reses said. “Nobody can give folks like this a loan; it’s a really hard business.”

A coffee shop owner, for example, might need $500 to buy a new cappuccino machine. An electrician might need to fix a truck immediately, a day after paying employees.

“If you need less than $100,000, it’s really hard for banks to create a return on investment on a loan like that because of all the multiple touchpoints that are required,” Reses said.

Women and underrepresented minorities are affected more than others, she said. While 50% of all businesses are approved for loans, women-owned businesses are approved only 30% of the time.

“Another study found that when underrepresented minorities, either black or Hispanic business owners, walk into a bank, they are given far less information and they’re asked far more questions about their personal finances,” she said.

Further constraining credit for microbusinesses is the fact that U.S. banks have closed 2,000 bank branches in low-income neighborhoods and 1,000 rural branches since 2008, Reses said.

Algorithmic lending

Most of Square’s loans are made with no human intervention but based on lots of data, she said.

“We’re actually looking at their business every day, analyzing their data every second to be able to underwrite and facilitate loans,” Reses said, referring to the payment information Square collects from its merchant customers. “Typical lending paradigms reflect them as poor credits. And that creates roadblocks for entrepreneurs getting access to credit.”

Square’s machine learning models take in 400 signals from Square sellers every day. The models consider how consistent the merchant’s revenue is; what time of day its transactions take place; how many customers it has each day; whether those customers come in daily, weekly, or never again; whether sales are increasing or decreasing; and even whether the business’s name is in is email address.

“There are different types of digital signals that we've determined are predictive of a seller's credit,” Reses said.

Overall, 80% of Square merchants are eligible for a loan through Square Capital. And 80% of the loans are made outside major metropolitan areas.

“These are communities that have lost local bank branches,” she said. “We don't need a branch. All we need is an internet connection. Within a few clicks and almost certainly within a day, we're able to deposit money into a small businesses.”

The loans are repaid as a percentage of the merchant’s daily sales.

“We’ve realized the kinds of folks who struggle to get loans are actually not risky at all,” Reses said. “They’re often incredible bets. These tiny businesses are great, consistent credits.”

Most of these loans are paid back within a year and the loss rates are less than 4%, Reses said.

About 50% of the loans are made to women-owned businesses; overall, 18% of small business loans are made to women. Square also provides 30% more loans to minorities than traditional lenders do, according to Reses.

Businesses that have received loans have grown 10% faster than those that haven’t.

Banks stepping up online lending

Traditional banks see the opportunity in lending online to small businesses and are stepping up their use of similar technology.

The same day that Reses spoke, HSBC announced it is white-labeling Biz2Credit’s online lending platform for small businesses, and Fifth Third Bank said it has deployed Fundation’s small-business loan origination platform.

Mary Makfinsky, senior vice president at HSBC, said that even though her bank has a well-known brand, it has only 230 branches in the U.S.

“We’re not that big that we could have done it ourselves,” she said. “We needed a partnership, and that’s going to really help us make it easy for our customers.”

For reprint and licensing requests for this article, click here.
Alternative acquirers Peer-to-peer lending Marketplace lending Online banking Online payments Machine learning Data management Data mining Artificial intelligence Square HSBC Fifth Third