CredoLab Lands $7M Funding, Bringing the “Gini” to US, Elsewhere
Chief Product Officer Michele Tucci said the platform uses 50,000 data points of mobile phone activity to predict a prospective borrower’s debt capabilities. CredoLab serves the 1.7 billion “credit-invisible” customers across the globe that may have some credit history, but not enough for a score, let alone a prime score.
“We do this in real-time: in less than a second a lender anywhere in the world, receives a credit score from Credolab,” Tucci said. “We don’t know the identity of the user; it’s only known to the bank or the lender, not to Credolab.”
CredoLab anonymously collects thousands of mobile data points, uses that data to create behavioral models, and then derives a credit score. The data can be anything- from the type of apps a user downloads, to the number of calendar events created- even the amount of texts the user sends. Is the user a gambler, a gamer, does the user use a work email during the week, and how many calendar events they schedule- all go into the predictive model.
“Some of these micro behavioral patterns could be the type of files being downloaded. Is it mostly music, or is it PDFs- or the percentage of photos taken in the week prior to the loan application that are selfies,” Tucci said. “So these are all indications that we collect and find a correlation we compare and analyze about 1.3 million micro behavioral patterns.”
Tucci said the CredoLab platform offers unmatched speed and predictability for customers’ future credit habits. He said Credolab helps lenders save money because they can better predict how their borrowers will act. Borrowers benefit by the program: Tucci argued that if lenders can better expect how they will be repaid, they tend to lend more.
The team built the platform for the world’s risk managers, whom Tucci knows constantly worry about the health of their transactions.
“Our CEO and founder Peter Bartek has more than 20 years of experience managing risk,” Tucci said. “So he feels the pain of the CROs out there, and our solution is built to address the very specific needs of chief risk officers.”
To explain the CredoLab platform’s accuracy, Tucci used a data metric called the Gini coefficient, a number between 0 and 1 that identifies to which category a request belongs. In this case, the GINI is used to classify borrowers as creditworthy or unworthy based on their mobile data.
“Zero is like flipping a coin; you have a 50/50 chance of getting the decision right. Basically no predictability,” Tucci said. “A GINI of one is like my wife; she’s always right. You know exactly what outcome to expect every single time.”
CredoLab’s platform has a predictive power of 0.6. Tucci cited World Bank economist David Mckenzie, who found for each decimal increase in GINI, there is a 1% cost savings from a risk point of view.Last modified: October 2, 2020
Kevin Travers is a Reporter at deBanked. Email me story tips at email@example.com and connect with me on LinkedIn