When you apply for a loan, most lenders will check your credit score and verify your income. But not Upstart Holdings (UPST -4.81%). The company uses more than 1,000 data points in its artificial intelligence-driven process in order to make smarter lending decisions. In this Fool Live video clip, recorded on April 5, Fool.com contributor Matt Frankel, CFP, and Industry Focus host Jason Moser discuss Upstart's unique approach and why it's such an important competitive advantage.

10 stocks we like better than Upstart Holdings, Inc.
When investing geniuses David and Tom Gardner have a stock tip, it can pay to listen. After all, the newsletter they have run for over a decade, Motley Fool Stock Advisor, has tripled the market.*

David and Tom just revealed what they believe are the ten best stocks for investors to buy right now... and Upstart Holdings, Inc. wasn't one of them! That's right -- they think these 10 stocks are even better buys.

See the 10 stocks

 

*Stock Advisor returns as of February 24, 2021

 

Matthew Frankel: All right, so the most voted up question is from Disk who says, "What are some of the data points that Upstart is looking at the only they have access to, and how are they getting this data?" Obviously, a lot of that is proprietary. They're not going to give away their secret sauce. But the point is, if say a borrower applies that has defaulted on a credit card five years ago, their salary's gone up by X amount a year since then, they have two credit cards that they've paid on fine, things like that. Given that combination, how likely are they to default on a car loan or a personal loan? That's the question Upstart tries to answer. Most lenders just look at the borrower's FICO score and maybe their income to see if they can afford the loan. It's data points like car loan history, or past repossessions, or how long ago was their latest late credit card payment? A bunch of different personal data points. Do they have a bachelors degree? Do they not? That's a big one that a lot of these fintechs are using that is correlated to payment history.

Jason Moser: Yeah.

Frankel: As they build their database, if they make a million loans, they have this data for a million different customers, and can really see the correlation between the certain variables that other lenders just can't. It's not that its data that only they have access to. Obviously, any lender can ask a borrower if they have a bachelors degree, for example. It's just building a big library of this data, and using it in combination with all these other variables to see correlations that other lenders can't see.

Moser: Yeah, I know I've used this example before, but Xoom, the financial remittance company that PayPal (PYPL -0.98%) acquired a while back, that was what they did really well, that they put together this risk assessment model, and it was able to just incorporate a lot of data points that they pulled from transactions, people who are doing business with them, they were able to get information regarding just a million different data points, and then being able to turn that into a risk assessment model to basically determine if someone sends this money, are they good for it? Are we going to have the situation where we've got either fraud or accounts that are unpayable? They were really able to take this and minimize losses. Just goes to show again, it was this idea of taking the data, it's like you get the data that's one thing, but then figuring out what to do it is an entirely different thing altogether. Companies can spend lot of time trying to figure out exactly what to do with that data. The ones that nail it, it can be a real competitive advantage that's difficult to supplant because you can't figure out exactly how to supplant it. You don't know what they did in building that proprietary model to begin with.

Frankel: Yeah, I would agree with that. As a lender, especially when you're targeting the sub-prime market, those little variables are very important, especially if you're not trying to give people 20% interest rates on car loans.

Moser: Yeah.

Frankel: It's really important to do a better job of assessing when people are going to default or not. I remember one statistic from Upstart's presentation that said that most personal loan buyers don't have a credit score that would get a lender's top rates. But most of them have never defaulted on a loan, or made a late payment, or anything like that, so why shouldn't they get the best rates? Upstart's formula tries to really cut through that and level the playing field, and give people the credit that they're due, I guess you would say, better than the traditional model of just looking at your credit score.