Credit scores can be predicted by your online behavior, study claims

Information like whether a consumer capitalizes their email address in online forms, when they purchase products on e-commerce sites and what device is used to browse the web can inform how likely they are to default on a loan, according to a new study.

While many unknowns still exist, the early findings could have significant ramifications, including providing those individuals with no credit history an opportunity to begin building a score. It could also lead some customers to change how they act online.

Consumers voluntarily hand over mass amounts of information when they visit websites and much of that data can be used to predict future financial behavior, says study author Manju Puri, a finance professor at Duke University’s Fuqua School of Business.

“It’s as informative as your credit bureau score,” Puri told FOX Business. “The digital footprint and how useful it is probably will move dynamically as people begin to understand it.”

To conduct the study, Puri examined over 270,000 purchases from a German e-commerce site -- akin to Wayfair – from October 2015 to December 2016.

The company partners with two private credit bureaus and analyzes the digital footprint of its users to determine whether to allow consumers to pay for products within 14 days after they are shipped, providing Puri’s team a benchmark for comparison.

The results indicate that your digital footprint can be an accurate predictor of credit scores.

Those with more impulsive purchasing behavior, like customers who shop in the early hours of the morning or arrive at the website via paid advertisements, are more likely to default on loans, according to the study.

Individuals who have iPhones are less likely to default than those who own Android devices, a finding that supports prior studies indicating Apple users spend more and have higher incomes than those who use Google products.

Consumers shopping from a mobile phone are three times more likely to default than those ordering from a desktop computer. And those customers who have their names in email addresses are 30 percent less likely to fail to pay off the money they borrow, the findings show.

Ticker Security Last Change Change %
AMZN AMAZON.COM INC. 183.66 -0.17 -0.09%
WMT WALMART INC. 67.02 +0.32 +0.48%
TGT TARGET CORP. 141.16 -3.44 -2.38%
W WAYFAIR INC. 52.64 -1.92 -3.52%

Put together, both a credit score and a consumer’s digital behavior can give a more holistic view of behavior.

“The two are complements, not substitutes,” Puri said. “If we use the two in isolation, the default predictions are roughly similar. But when we put the two together, we actually do better.” 

And digital behavior can sometimes be even more insightful than the data a credit bureau has. Online information is more forward-looking, according to Puri, and is often a better predictor of future credit scores. It’s also quicker to compile and largely requires no activity on the part of the borrower.


The findings could help those with thin credit histories obtain financing, but could also earn the attention of regulators who may be concerned how the data is employed.

“There’s a lot of information. We need to be careful how we use it going forward,” Puri said.