A leading online consumer lender was suffering from an unacceptably high rate of default. It followed industry best practices, using the leading indicator (FICO), and yet, the default rate refused to budge. In addition, a significant percentage of the borrowers did not accept the clients' APR offer for the loan and ended up borrowing from other lenders.
Was this a good thing (Did we refuse to give loans to defaulters?) or a bad (Did we turn away good borrowers with our offer of too-high-an APR?).
Hudson Data was brought in to investigate the problem and to develop actionable insights. We built a comprehensive AI system utilizing both internal data as well as external data with thousands of variables.
This enabled the firm to make online lending decisions in less than 10 seconds, lowered the default rate by more than 50%, and enabled optimal risk-based pricing by precisely placing borrowers in the correct APR bucket.
Our work enabled the client to borrow over a Billion dollars from the market based on the improved borrower-repayment ratio.