The Harmony Of Humans And Machines To Mitigate Risk

Small business financial service providers are beginning to realize the value in linking their business customers to lending services. PayPal launched its Working Capital unit in 2013, for example, while QuickBooks took to partnering with other lenders to fill in its small business financing gaps.

One of the latest to do so is small business payments processing firm JetPay. The company revealed earlier this month that it struck a partnership with small business lending platform Credibly to offer its SME customers access to loans.

According to Credibly Cofounder and Chief Revenue Officer Ryan Rosett, this trend is part of FinServ players’ recognition that they need to meet the holistic needs of small businesses.

“The last thing a credit card processor would want,” Rosett recently told PYMNTS, “is that somebody leaves their platform because they don’t offer a service,” like small business financing.

But it makes sense for other reasons, too. For example, QuickBooks already has access to troves of a small business’ financial data, making it easier for a lender to come in and take a look at that information to underwrite a loan. A partnership with JetPay puts Credibly in a similar position, and Rosett said his firm will certainly use data from JetPay’s card processing to gain insight into an SME’s financial health.

“We’ll look at their credit card processing volumes,” he said, though adding that doing so is only a small part of the picture when assessing the creditworthiness of a small business.

“Often, credit card processing might depend on the business,” he explained. “It may be 90 percent of their business, or 5 percent. We don’t necessarily want to rate their business purely on their credit card processing volume.”

Today, in the age of Big Data, machine learning and the Internet of Things, lenders have the chance to aggregate fields of data that provide deeper insight into a small business’ operations and financial health beyond how much money is coming into a business via credit cards.

“We look at so many variables when making a determination,” Rosett said. A small business’ bank statements, social media and a slew of other sources are aggregated to create Credibly’s “profitability model.”

“The more data each company has to assess and determine the creditworthiness of a small business owner, the quicker the decision will be,” he explained.

[bctt tweet=”‘The more data each company has to assess … the quicker the decision will be.'”]

So, how do lenders weed out the good data from the bad? By balancing data analytics automation with the good, old-fashioned human element. At Credibly, Rosett said, this comes in the form of its data science team, which he described as the “core” of its business.

“There is no doubt that there are a lot of analytics we use, but there are certain circumstances that analytics can’t pick up on,” he said. For example, Rosett pointed to rental history. There may not be another way to know that a small business is a few months behind on rent unless someone actually contacts the landlord. Plus, some information streaming in simply isn’t valuable when underwriting a small business loan — for example, the subjective opinion of a restaurant. But other data, like information about management changes that can be found on Yelp, may be highly useful.

“It’s difficult to fully underwrite a loan based on automation,” Rosett said. “We look at it holistically. Automation certainly guides us and speeds up the process, but there is a human element to what we do in making decisions.”

In this Big Data age, companies are beginning to struggle with taking in too much data and knowing which data to analyze. Rosett said Credibly’s data science team is continually working with different scoring models, weighing statistics in various ways to reach reliable conclusions and predictions.

Experimenting with this balance of data and using data analytics automation in harmony with the human element will be where small business lenders truly find their footing in the coming years, Rosett said.

“Alternative data sets will affect small business lending in the future; it’s constantly evolving,” he said. “I think it’s certainly where lending is going, and making decisions quickly and making them intelligently is ultimately what I think everybody in this business is trying to do.”