Online Lending

Ford, MCA Funders Take Pages from Tech-Based Underwriting

Alternative lending fever has spilled over into the auto sector, evidenced by the financing arm of automaker Ford’s decision to move beyond FICO and deeper into machine learning for credit decisions. Ford is moving toward alternative lending strategies in an attempt to capture a wider swath of borrowers, including those with “limited credit histories,” and bolster auto sales.

Ford’s decision comes on the heels of a study with fintech play ZestFinance, the results of which favor a machine-learning based approach to credit decisions.

Ford’s decision comes on the heels of a study between Ford Credit and fintech play ZestFinance, the results of which favor a machine-learning based approach to credit decisions.

“There is absolutely no change in Ford Credit’s risk appetite. Ford Credit is maintaining the consistent and prudent standards it has applied for years. This enhanced ability to look at data will help us more appropriately place applicants along the full spectrum of the risk scale. The result will be some that some people may appear on that scale who did not before, and some applications that are approved today might not be approved in the future. The risk appetite remains the same,” Ford Credit spokesperson Margaret Mellott told deBanked.

Until now, there has been no aspect of machine learning in Ford Credit’s underwriting process.

“The study showed improved predictive power, which holds promise for more approvals … and even stronger business performance, including lower credit losses,” according to Joy Falotico, Ford Credit chairman and CEO, in a press release.

Ford is targeting consumers with a lack of credit history, especially the millennial generation.

Tech-Driven Underwriting

While Ford embraces tech-driven underwriting, this style is already knit into the fabric of the MCA and online lending communities.

To name a few, Upstart takes a machine learning approach. FundKite developed algorithmic-based underwriting. UpLyft’s underwriting process has an automated component to it.

Alex Shvarts, CTO and director of business development at FundKite, a balance-sheet based funder, said the company has been writing algorithms since the early days. Now the tech- and algorithm-driven funder wants to expand into small business lending in Q1 2018.

“We’re building our technology to the point that by Q1 next year, we will get into automated loan products. Our technology will be able to underwrite loan products within seconds. We have a lot of data we put together, which allows us to price deals and make offers relatively quickly,” he said.

By a lot of data, Shvarts is referring to hundreds of data points that are used to measure merchant performance. FundKite, which has a default rate of far less than 10 percent, takes the data, reworks and combines it, leading to a fast result.

“Besides the data points we look at the merchant from a collections point of view. If this person or business runs into trouble, could they go out of business or would they be okay?” he said.

That’s where the human element to the underwriting process comes in.

Human Element

While FundKite relies on algorithm-driven underwriting, the funder is not running an online app yet. There is still a need for human participation surrounding data input, information that is then verified by machines.

“The human element is entering the information correctly, and the machine spits out predetermined pricing based on the business data points and industry,” said Shvarts, adding that FundKite views that information in the context of micro-trends in the industry as well as the overall market environment.

“We know that during certain seasons some merchants perform worse than others. The numbers say the merchant should get this, but we dig a little deeper and say no, this merchant can’t handle this much of an advance and repayment along those lines. The final touches are done by humans. Our technology is advanced so that we are able to get to that point a lot faster and more accurately,” Shvarts said.

Second Opinion

Michael Massa, CEO and founder of Uplyft Capital, points to a hybrid approach in the company’s credit underwriting, referring to the automated scoring portion of Uplyft’s underwriting model as a second opinion. “We believe there must be a hybrid of human and automated technology,” said Massa.

Uplyft relies on a proprietary scoring model. The model includes an automated function that attaches a unique rating to the small business based on certain features in the prospective borrower’s profile, such as a home-based versus business location and the number of years the company has been in business, to name a couple.

“It’s only as second opinion for our underwriters, really,” he said, adding that cash flow and affordability are major drivers of the credit decision. “In most cases we price at max affordability for the client while protecting them from overleveraging their accounts, allowing us to provide real help and establish merchant loyalty.”

Second opinion or not the automated function is part of what makes Uplyft a fintech play, setting the funder apart from the banks. “They’re like the payphone and we’re the iPhone. They’re yellow cab and we’re Uber,” said Massa, adding better yet, “we’re Lyft.”

Uplyft is in the process of developing a trio of portals designed for merchants, sales partners and investors to be released shortly. “We are API-ing that now into our CRM,” said Massa.

Merchants can access the portal to apply for funding while sales partners use it to submit files and view a status. Investors can track their participation via the portal. The new portals will be available on the website and through a mobile app that Uplyft is in the final stages of developing.

Uplyft also recently inked an exclusive partnership with an undisclosed software company allowing merchants to link their bank account to the application, capturing six months of actual PDF bank statements in the process.

“It can help us with the initial credit decision and when we’re conducting final verifications. We get the actual bank statement. It’s a legitimate bank statement, not a rendition,” said Massa.

Fintech & Auto Finance

As for the auto industry, don’t be surprised to hear about further collaboration between the automakers and the fintech market. “Financial technology is key … as fintech can contribute to an even more seamless and better personalized vehicle financing experience for the consumer,” according to the Ford press release.