Startups that facilitate lending by investors to consumers and small businesses are looking beyond traditional indicators of credit risk, such as an applicant's assets and FICO score, to build more comprehensive credit profiles. The firms are incorporating everything from a person's education to a restaurant's Yelp reviews to rate the creditworthiness of borrowers seeking loans on their platforms. "Automization" an industry buzzword for having the borrower self-generate data, such as by photographing documents with a smartphone, then filtering it electronically lowers costs and increases the speed at which startups can lend.
"We are literally Big Data companies," said Ron Suber, president of Prosper, one of the pioneers in what used to be known as peer-to-peer lending, at the AltFi Summit in New York Tuesday.
Matt Burton, co-founder and chief executive of Orchard Platform, a firm that aggregates marketplace lending platforms much the way Amazon aggregates independent merchants, said he expects this industry to double in total assets to $25 billion next year.
A "chronic underinvestment in technology" is causing banks to miss an opportunity, he said, standing in front of a projected slide of decades-old green code to emphasize his point. The longer banks wait to update their technology, the higher their costs will be to do so.
"Why," he asked, "is now the time they're being disrupted" by marketplace lenders in the small-business and small-balance loan niches? The answer is the upstarts' technology has created efficiencies allowing them to compete with banks, Burton said. Alluding to the green-on-black coding behind him, Burton said banks had to follow peer-to-peer lenders' lead and ditch "the green screen."
Burton, who worked in advertising technology before founding Orchard last year, credited rapidly improving storage technology and pricing for enabling the increased reliance on data.
"It's more expensive to throw away our data than it is to keep it," he said.
Pave, another major player in the field, uses data to focus on lending to millennials who might not have built traditional proof of creditworthiness. Pave uses in-house metrics to determine earning potential and employability through work history, education, and telephone bills to develop profiles of borrowers and apply its own credit scores.
Every company in the peer-to-peer industry seems to have its own internal credit scores based on data points that might not have been accessible, or even existed, until the last few years. The approach makes sense since most peer-to-peer firms see their customer bases as small businesses that might fly under the radar of banks that lack the manpower or the inclination to lend to them.
Alex Kriger, co-founder and chief executive officer of CF4All, a crowdfunding facilitator, said that his company monitors how users behave on its site, and how that correlates to default rates. OnDeck reportedly uses Yelp reviews as a data point for evaluating restaurant borrowers. OpenEnergy and CreditJunction, which specialize in alternative energy and industrial small business borrowers, respectively, tailor their metrics to those specific customers. Burton, in an interview prior to the AltFi conference, used the example of an online bike company's monthly UPS shipping volume as a measure of sales.
Still, not everyone's a data evangelical in the business. D.J. Paul, chief strategy officer for Propellr, a firm that uses crowdfunding to invest in a relatively small but higher priced--number of commercial real estate projects, said he's "leery of drawing conclusions from data sets," and data sets alone.
"You can go very far" using data, Paul said, but, "you can't rely on it 100%." Just as the advanced metrics described by Michael Lewis in "Moneyball" haven't completely supplanted traditional scouting in baseball, underwriters can't be replaced by data in real estate lending.
"Underwriting standards cannot be overlooked, they can't be fudged," Paul said. "At the end of the day you have to have a mind look at it."
Real estate risk depends on subjective factors, like location, not just objective ones, Paul says.
Still, the systemic collection and usage of data also brings transparency to areas that have traditionally been opaque, a factor the marketplace lending industry believes gives it an advantage for packaging securities to financial institutions.
"It's just more granular, a better model," said James Wu of Monja, a San Francisco-based company that specializes in peer-to-peer loan investment.
While securitizations of these loans are scarce, the transparency may help it take off.
"The best way to make investors confident in your lending business is to make all of your data publicly available," said David Snitkoff, another co-founder of Orchard. "That way you show you have nothing to hide, investors can understand the returns, they can understand the risks, and it even forces you to hold yourself to a higher standard because you know everyone can see your data."
For example, Prosper provides an application programming interface for Orchard and University of California-Berkley professor David Swart, an expert on this field, to access its client data for research. At the conference, Snitkoff and yet another Orchard co-founder, chief financial officer Angela Ceresnie, overlaid Prosper's borrower data onto data from the Federal Reserve Bank of St. Louis. The exercise showed that Prosper borrowers have higher debt-to-income ratios than the average American consumer.
"What Prosper and Lending Club did by putting their data out there from the beginning was really transformational," said Ceresnie, who started out as a retail investor in peer-to-peer lending prior to helping start Orchard. "I was able to see the historical performance and make a determination." (She and Snitkoff have previous consumer lending backgrounds from traditional banks.)
In the long term, peer-to-peer lenders see their approach to data as benefiting the banking industry. "This is a good thing for banks," said Ramit Arora, co-founder and president of Biz2Credit, "because it forces them to become more efficient."