The technologies that make online personal lenders tick

Michael Lock, senior vice president of lending partnerships at Upstart
Upstart can do real-time verification of 73% of its loans, said Michael Lock, senior vice president of lending partnerships at Upstart. The other 27% use digital-forward methods to gather the required information, such as by asking the applicant to photograph the document with their phone rather than email a PDF.

Though customer satisfaction is generally highest among people who apply for loans in person rather than online or by phone, two online lenders are defying this norm. One, Marcus by Goldman Sachs, rose to the top of J.D. Power's recent ranking of personal lenders. Another, Upstart, came in seventh out of 22 (the ranking reflects just those borrowers who recognize Upstart as the lender). American Express, which ranked third, also seems to be getting a satisfaction boost from technology advances.

"That in-person ability to ask questions, probe, understand the need, and make recommendations resonates with customers," said Craig Martin, managing director and global head of wealth and lending intelligence at J.D. Power, explaining the overall preference for human loan officers. 

J.D. Power's 2022 U.S. Consumer Lending Satisfaction Study reported that personal loans were "filling the void left by pandemic-era relief efforts," with customers gravitating toward the competitive rates, easy access and variety of options; a July study from the company found they were most commonly used to pay off debt and to supplement income due to lost wages. Experian found that the number of personal loan accounts has increased by 16% over the past year, while TransUnion reported in August that the number of consumers with credit cards and personal loans reached record highs in the second quarter of 2022. At credit unions, unsecured lending grew 13% in the first six months of 2022, compared to 0% annual growth in the first six months of 2021, according to CUNA Mutual Group.

A look at how the top-performing digital lenders operate can be instructive to bank and nonbank lenders alike. For instance, the lenders with the happiest customers often use artificial intelligence and machine learning to speed up processes. They provide online application forms that prioritize what customers care about most, such as size of loan and monthly payment, rather than asking for customer information right off the bat.

Marcus is an example of the latter.

Martin says a J.D. Power colleague tried to take out a personal loan at his local bank and  was put off when he was told to go online or make an appointment. 

"He went to Marcus and it was easy, fast, and straightforward," said Martin. "Even once he completed the process, they came back with follow-up communication about when his payment was due."

Ilya Gaysinskiy, global head of engineering for Goldman Sachs' consumer business and head of the consumer and wealth management platform, chalks up customer satisfaction to a few elements, one being the design of the application.

"We did a lot of research to figure out what resonates the most," said Gaysinskiy. For example, the bank found that customers want to know how much they will pay per month. Sign-up begins by asking applicants to select the size of the loan they are looking for and the approximate monthly payment they are targeting. Then it gets into personal information such as name and annual income.

"We look at it from a customer-centric perspective, or what they are trying to achieve, versus the information we need to get the loan approved," said Gaysinskiy.

Another is the call center, which is staffed at all hours. The agents aim to answer all customers within 30 seconds. 

"That continues to build trust," said Gaysinskiy. "On one hand, there is a modern digital experience. On the other, there is the peace of mind that a human is a phone call away."

Upstart Holdings emphasizes speed. The company relies heavily on artificial intelligence and machine learning to approve and underwrite loans.

Late-payment rates are rising at nonbanks that lend to people with lower credit scores. "We're probably entering a stretch where you're going to see a separation between those that are relatively good underwriters and those that are not," one analyst said.

August 7

"Upstart executes well on the application approval process, including how fast they get people approved," said Martin. Customers with lower credit may expect a rigorous application review with more required documentation, but Martin finds that Upstart delivers a more seamless experience that exceeds customer expectations.

Michael Lock, senior vice president of lending partnerships at Upstart, says it takes an average of five minutes to complete the application and every applicant gets an instant credit decision. He also says that 73% of loans are instantly approved and fully automated without  applicants having to send in documentation.

"They're a good example where technology plays a key role," said Martin.

Lock credits the machine learning models Upstart has honed for about eight years. Approximately  three-quarters of the decision is made using traditional credit data, but Lock says Upstart digs deep into credit reporting data, "not just the loans I have and the payment history," he said. "Traditional models look at a very small subset of data in credit files." Another quarter of Upstart's credit decision relies on alternative data, including the applicant's education level and profession, to make predictions about employability.

"There has been some controversy over us using the education factor — 'you just want to lend to people who went to Harvard and Duke' — but the model looks at average income levels of an exiting student," or recent grad, said Lock.

The company has also been using artificial intelligence and machine learning over the last four years to verify certain details to lessen the burden on customers to back up their assertions with documentation such as W-2s. Broadly, Upstart draws from third-party databases to verify that, for example, the annual income a nurse in Oklahoma enters on their application is in line with what third-party sources register for the same job and location.

"By connecting to databases, we can do this real-time verification in 73% of instances," said Lock. For the 27% of loans that cannot be verified digitally, Upstart favors digital-forward methods of uploading documentation, such as asking the applicant to take a picture of a document with their phone instead of emailing a PDF.

"Email is the new snail mail," said Lock.

American Express, which makes its personal loans available to consumer cardholders only, lets customers check for pre-approval by logging into their accounts. Most cardholders know if they are approved within minutes, said Kristen McGinnis, vice president of U.S. personal loans at American Express. The company also pre-populates applications.

"American Express has a leg up because they know the customer, know their needs and can position an offer appropriately based on behaviors and habits of that customer," said Martin.

For reprint and licensing requests for this article, click here.
Fintech Consumer lending
MORE FROM AMERICAN BANKER