When Krishna Gopinathan began his career as a data scientist, he helped create the fraud detection software now used by FICO to protect more than 2 billion credit cards worldwide. Now he’s turning his attention to programs that can determine a potential borrower’s ability to pay back a loan using somewhat unconventional pieces of information.
Gopinathan co-founded San Diego-based Global
Analytics in 2003, aiming to provide loans to so-called underbanked consumers with limited access to cheap credit and who often turn to payday lenders. Global Analytics focused its efforts in the U.K., eventually becoming the second-largest online installment lender in the country with revenues of more than $100 million.
But Gopinathan splintered off last year with two Global Analytics advisors to form Applied Data Finance, seeking to apply similar data work to U.S. consumers. ADF, which operates the consumer-facing brand Personify Financial, received a $50 million credit facility in April from Victory Park Capital to launch its loan products.
“This is an opportunity to scale a business much larger, target the underbanked with the benefit of all the latest technologies and do it in the U.S.,” Gopinathan said.
ADF, co-founded by hedge fund manager Dan Zwirn and investment banker Eric Schwartz, offers unsecured loans up to $10,000 payable over up to two years with interest rates between 20 percent and 100 percent APR. The higher end of that scale is about midway between traditional bank loans and payday loans, and Gopinathan said consumers are “fully-aware” they are getting a relatively high-priced loan.
ADF is only available for borrowers in Georgia and Missouri now, though licenses for a half-dozen states and a bank sponsorship that would allow loans of up to 36 percent APR in 48 states are expected by January, Gopinathan said.
Rates and eligibility are currently determined through traditional sources of data, such as credit bureau information and bank account statements. But Gopinathan and his teams in San Diego, New York and Chennai, India, are working on other data collection approaches that could expand their borrower pool. He offered the example of a pizza deliveryman applying for a loan and getting perceived by an underwriting algorithm as a high-risk borrower.
“But what if you find out through LinkedIn that he’s actually a fourth-year MIT student with an internship at Google,” Gopinathan said. “If you just think pizza, he’s one level. But with all this extra information, it makes more sense that this person could be capable of taking a bigger loan, pricing it a bit lower and giving more time to pay it back.”
If implemented, that type of analysis could advance some of the work Gopinathan used at Global Analytics, which said it could pull data in some cases from Facebook, such as an applicant’s number of friends or time spent at a particular address.
The Falcon System
Gopinathan got his start at San Diego’s HNC Software in the early 1990s before it was acquired by FICO. There, he developed the Falcon fraud detection system, which has become a worldwide standard. Michael Thiemann, who hired Gopinathan at HNC, said industry experts at the time believed credit card fraud could not be fixed technically and could only be solved by catching criminals in the act.
“We decided to prevent them from getting any value out of the card,” Thiemann said. “We built a profile of every card and in real time compared the spending pattern to what they’d done in the past and decided how likely it was that the card holder had their card. Krishna’s breakthrough was creating individual models of every card holder.”
Gopinathan eventually tapped Thiemann to be Global Analytics’ CEO in 2009. The reason Global Analytics never made a serious push into the U.S. market, according to Thiemann, was that an early joint venture deal came with a one-year noncompete clause here, so the company turned its attention to the U.K. Success there absorbed most of Global Analytics’ capital, preventing expansion into another county, he said.
Thiemann left Global Analytics last year, spinning off one of the company’s divisions to offer interest-free loans for consumer goods as an employer-provided benefit.
Gopinathan was first fascinated with data science at college in the late 1970s after hearing a lecture on mathematical models that could predict pollution in Lake Ontario. The field now has far more predictive power, bolstered by the flow of information from consumers every day on social media and other online sources, Gopinathan said, clearing the way for companies such as ADF.
“The sheer quantity and variety of data available has skyrocketed,” he said. “We are emitting data packets in all directions. Our life has been captured in data in so many more ways than it used to and the scope with which predictive technology can be used has dramatically increased.”
Applied Data Finance
CEO: Krishna Gopinathan
No. of local employees: 20
Headquarters: San Diego
Company description: Online unsecured lender offering loans of up to $10,000 as an alternative to payday lenders
Key factors for success: Backed by one of the founders of credit card fraud protection software, ADF hopes to predict borrowers’ ability to pay using nontraditional information sources and analysis