Alternative online lender ZestCash believes the market for short-term emergency loans is getting larger, and it's probably right.
To find out how much larger its market can get, the Los Angeles-based firm, which positions itself as an alternative to payday lenders, has finished work on a new credit decisioning infrastructure that can run multiple underwriting models in parallel, each with a different focus, to better analyze credit risk and expand its customer base. ZestCash has also introduced Hollerith, a new set of models that, in conjunction with its expanded analytics, is part of a goal to extend credit to 25 percent more people and increase repayment by 20 percent.
"Many of the underbanked have real jobs," says Douglas Merrill, founder and CEO of ZestCash. "They aren't drug dealers, they are folks that are working but live paycheck to paycheck. They don't need this credit to fund a lifestyle, they are short and need to make payments."
ZestCash was created in 2009 to offer loans to people that its credit engine finds reliable, but who would probably not qualify for a bank loan. Merrill, the former CIO of Google (GOOG), co-founded ZestCash with Shawn Budde, the former head of subprime cards at Capital One (COF).
ZestCash operates in the fast growing market for online lenders. Other firms include BillFloat and peer-to-peer lending firms such as Kivo, Prosper and Lending. Another new online lender, Kabbage, offers advances for small business among its lending services.
ZestCash recently received a new round of investment and has been adding staff with expertise in analytics and other emerging data techniques to allow it to make decisions on short-term loans, generally a few hundred dollars payable over six months, at about half the cost of payday lenders.
ZestCash's new decisioning method runs myriad models in tandem, rather than in isolation. The goal is to mine richer, more varied data on payments and financial histories to separate borrowers who are poor risks from borrowers who simply need some cash to make it to the next paycheck.
Typically, lenders use single regression models that look at about 15 points of data to make credit decisions. ZestCash runs about 10 unique underwriting models simultaneously that consume thousands of raw data elements collected from borrowers and third parties.
The models then transform this data into tens of thousands of meta variables to assess key customer behaviors such as fraud, short-term and long-term credit risk, or the amount of money a borrower will likely repay.
The models are then "ensembled" by the firm's internal analytics and math engine to arrive at a final underwriting decision that more accurately predicts credit risk.
ZestCash uses search engine analysis-style proprietary algorithms that often work counter to traditional scoring methods to locate borrowers who may be having a tough time but are still paying attention to their finances. For example, someone who has recently declared bankruptcy, generally considered to be a bad bet for a loan, may get approved by ZestCash because the recent declaration suggests that person is still being diligent about his or her finances.
"We're using Google and Netflix techniques in the credit space," Budde says. Netflix's Cinematch uses purchase histories and its inventory to make movie recommendations based on user preferences. Google offers analytics that generate intelligence on how people visit and use websites, and gives marketers tools to tailor campaigns.
