Like so many lenders, OneMain Holdings has a tremendous amount of data about how its customers have behaved over the years.
However, the subprime lender is trying to break from the pack and figure out how to use all that information to its advantage.
In the years immediately following the financial crisis of 2008, the company, then known as Springleaf Holdings, was plotting its future course and part of that meant understanding its customers better. It then had 90 years of experience as a lender but needed to better harness that knowledge.
-
A new class of fintech startups is using loan applicants' social networks to determine creditworthiness as the banking industry debates the merits of alternate underwriting methods.
January 22 -
Rejiggering the credit-scoring system wont do much to expand credit access. Instead, banks and financial services companies should start thinking about new ways to evaluate the situational factors that influence borrowers reliability.
August 10
In 2010 the company began investing in technology upgrades and major hires that would strengthen its data-analysis capability.
"Banks sometimes don't effectively mine the data they have, and that's a key differentiator of what we do," said Dave Hogan, the chief analytics and marketing officer, who joined the company in 2012 after stints at PNC Financial Services, JPMorgan Chase and MBNA.
Analytics have been a hot topic in banking for several years now, but OneMain's use of such tools underscores a major advantage that incumbent institutions have over financial startups: tons of data. Indeed, banks that are able to utilize their data better can serve their borrowers better, said Ann Armstrong, marketplace lending lead at KPMG.
"They have the opportunity to leverage analytics and really offer a different customer experience for lending," she said. "For example, there is a big opportunity to be a lot more specific for interest rates granted to individuals. One person may be less creditworthy at one interest rate, but more [creditworthy] at another rate. The more information a lender has on the borrower, the more prescriptive it can be with product offerings."
Like many of the startup fintech consumer lenders that have arisen in recent years, OneMain mines "alternative" data as part of the mix to determine creditworthiness, along with traditional credit scoring. This nontraditional data can include utility payment history, bounced checks, and even whether a potential borrower is using a Mac or a PC, Hogan said.
But as a company that dates back to 1920, OneMain can more effectively use such data, since it can overlay it onto historical data to glean insight on how credit decisions in the past may have changed if this newer alternative data were used, Hogan said.
"We take a lot of this [alternative data] and test it on real-world scenarios that actually happened," he said.
The company, which has $6.4 billion in net receivables, is doing now to help it inform credit decisions by taking new data, machine learning or other advanced analytic techniques and applying them to historically approved or declined populations, Hogan said.
"We can then measure the ability the new data or techniques have on generating improved performance," he added. "Being able to test against historical populations also dramatically reduces cycle times as we don't have to wait days or months for results to come in when testing new approaches."
OneMain said the tools have helped it maintain customer-acquisition costs amid rising competition from newer startup entrants into the market, improve customer conversion rates and keep loan-loss rates in the 5% to 6% range.
As with so much technology investments, it is hard to say how OneMain's analytics efforts are showing up in the company's performance, analysts say. For instance, it might be hard to peg how useful the changes will be until they go through a full credit cycle.
Also, it is tough to compare its loss rates to peers since OneMain doesn't really have any that release data, analysts say. For instance, banks are in a different business and even a company like Lending Club aims for customers with better credit scores than OneMain's.
Its largest competitor was the
The loss rate for the legacy Springleaf "was much better than legacy OneMain's," said Bob Ramsey, an analyst at FBR Capital Markets. "And Springleaf's numbers look a lot better than they did during the crisis and are a little better than even before that, but how much of that is from the macroeconomic environment and how much is from a greater use of analytics?"
Hogan said the company's focus on analytics isn't a response to the slew of fintech startups that have arisen in the consumer lending space. In fact he says that more competition in lending can only help.
"We are certainly happy at the number of competitors that have entered and brought attention to this industry," he said. "As a whole it's a good thing because it drives innovation and, in the end, more access to credit."
A major difference between OneMain and several fintech startups is its reliance on stores. The company has 1,800 branches in 43 states.
It is, however, making a major push to be more digitally focused. For instance, it launched an online lending unit called iLoan in 2013. Ramsey said given the move, he is pleased the company is incorporating analytics into its credit underwriting.
"Their digital strategy is an important part of how they are finding and acquiring new customers," Ramsey said. "And the more you lend online and not face to face, the more important robust data analytics become."
Robert Barba contributed to this article.
