Viaduct’s David Hallac on Capturing the Value of Automotive Data

Driving remains one of the quintessential American activities. Over 90 percent of US households have at least one car, and drivers spend an average of around 290 hours a year in their vehicles. And, thanks to the auto industry’s recent embrace of data technology, much of that time on the road now produces an enormous amount of useful information. David Hallac is the founder and CEO of Viaduct, a new company creating a machine learning platform for managing, analyzing, and monetizing that data.

“Ten years ago, car companies were building millions of vehicles, and they were not collecting any data. And over the past handful of years, they’ve flipped the switch and are generating more data than almost any other industry out there. This is this an entirely new data stream,” says Hallac, who estimates that 60 to 70 percent of new vehicles currently send at least some data to the cloud, and that by 2020, the figure will be closer to 95 percent. While the information is already being put to use—by, for example, automakers trying to anticipate component failures (and possible recalls) and insurance companies looking to identify customers’ “signature” driving behaviors—Hallac notes that “a lot of the machine learning methods that have come out of [Silicon Valley] are not meant for the scale or the type of data that’s coming from cars in particular.”

Hallac says that Viaduct’s platform will enable users “to take their data, regardless of the way they captured it or the specific sensors they’re recording or the hardware in the vehicle.” He adds that the company’s ultimate goal is to provide the go-to product for a range of businesses that hinge on accurately tracking, assessing, and predicting the performance of cars and drivers—not just automakers, but parts manufacturers and fleet managers—allowing them to “analyze this data at scale and capture the market in a way that automotive companies ten years ago wouldn’t have even dreamed.”