The tried-and-true Fico score worked well in years past, Keith Shields, senior vice president of analytic services at Loan Science, and chief analytics officer at Magnify Analytic Solutions, told the audience at the 14th annual Auto Finance Summit held in Las Vegas this week. But increasingly, traditional lending models are being undercut by custom solutions that allow for more approvals and lower default rates, and one major signal that is often ignored is a borrower’s social media channels.
“A custom score does result in a reducation of defaults,” Shields said. “Everyone can approve a 700, you see the number and ok, that’s easy. But what’s hard is to find a 650 that performs like a 700.” Lenders using only traditional data inputs could miss those 650s that perform like 700s, Shields said. “There are good 550s and bad 550s,” he said, and better analytics can help show which is which. He described a recent examination of Twitter activity that was matched up to default rates.
“Borrowers who used the phrase ‘test drive’ were significantly less likely to default,” he said. Phrases such as test drive indicate engagement, an important factor for borrowers’ likelihood to continue making payments. “Engagement lowers risk,” Shields said.