As Ford Motor Credit Co. refines the first phase of its machine-learning model for credit decisioning, the captive is already working on version 2.0.
“In the fourth quarter of 2018, we launched machine learning into production,” Director of Credit Analytics Yi Lu said during a panel discussion at the American Financial Services Association’s Vehicle Finance Conference last month. “We’ve seen a decent lift in the model’s forecastability and decided as a company that we are going to launch it and test it.”
It will likely take until the third or fourth quarter to see results from the pilot, which was launched for only “a few” segments, Lu said, adding that “we are confident we will have the benefit we initially think.”
Also read: Ford Credit Earnings Fall Despite Higher Lease Penetration, Stronger Auction Values
While version 1.0 uses machine-learning technology to develop a static model for use in credit decisioning, version 2.0. will be a “live, self-adapting learning model,” Lu said. During the panel discussion, Lu presented a business case for machine learning, a concerted marketing effort between Ford Motor Co. and Ford Credit.
“Consider a consumer that currently drives a 2015 C-Max, and we know that the lease will be up soon,” he said. “We use the machine-learning model to figure out whether we should sell this person a similar vehicle or maybe we should upgrade him to a different vehicle. Then we feed the information back to the motor company and Ford Credit to send a personalized offer [to the consumer].”
The goal is that the customized offer will result in a higher “conversion rate,” he added.
For more content like this, check out our upcoming event Auto Finance Accelerate, May 13-16 at the Omni San Diego. Visit www.AutoFinanceAccelerate.com to register.
