Start Your Semantic Engines: TrueCar Looks To Foster Transition Of Vehicle Data From Flat To Structured And Enhanced
Back when he was VP and CTO at Hearst Interactive Media, Mike Dunn advocated the use of semantic technologies for media organizations to rocket-boost their control over content, both for internal operations and for presenting a better face to users out there on the web. (See our story with his insights on that here). Now, Dunn has recently made the move to Truecar, an eight-year-young start-up focused on improving the car-buying process. As CTO, his mission is to modernize its data stack.
How do the two worlds of media and automotive connect? “There’s definitely a connection if you think about content as data,” Dunn told The Semantic Web Blog during a few free moments at the recent Semantic Technology and Business Conference. And, TrueCar gets “the importance of data, even though you don’t always have to throw the semantic web [phrase] in there. But things like sentiment-enhancing and context – those are useful words that don’t confuse people.”
Today, says Dunn, much of the data around vehicles, sales processes, and how cars are customized or configured tends to be fairly flat – that is, either unstructured and/or proprietary, but doors open up when it gains meaning — becomes structured, enhanced and openly known and leveraged from an industry perspective. “That transition, which we believe we’ll be able foster, will allow the creation of additional enhancing services to consumers and the industry at large,” he says.
As TrueCar rolls out new services, he continues, “we want to have a rapid way to deal with the market. For example, services around finance, or accessories, we want them to be API- driven out of a robust, high performance data environment.”
Dunn’s only been at TrueCar for a little more than a month, but says things like the use of HTML 5 microdata are on the agenda to help move the needle in that high-performance data environment. As he looks ahead to opportunities that can build on smarter data, he says he sees possibilities for a ton of semantically-enabled applications for driving new user services.
For example, intelligence about a car model’s wear-and-tear state three to five years down the road doesn’t normally enter into the buying equation. But shouldn’t it? “You might like it today but in 3 years, because of the manufacturer’s choice in materials, a brand [other than the one you may be emotionally connected to] would be better,” he notes. But to understand a vehicle’s residual value at the end of a lease – to come to understand the car as an asset to algorithmically keep track of its worth – some really solid data is necessary, and to build a service to provide that accounting, that data must be structured and easily integrated to the systems that will use it.
“Data has to be managed and enhanced, maybe with some NLP to convert it from systemic naming conventions to something more useful in descriptions to consumers,” he says. “You’ve got to get to the point where you could have it become highly structured so that more rapid integration can occur.”
TrueCar has as its aim providing value both to the consumer and the industry, where new smart data-driven services can help dealers do things like reconfigure cars to users’ desires while also tightly integrating with all the inventory and VIN information about the cars on their lot, in transit and at other lots. For users, especially the younger generation that has embraced the web for activities like researching and buying cars – and whose enthusiasm can influence their elders – TrueCar “aims to do more with social, with rock-solid mobile solutions and to highly contextualize it all with improving enhancement of data,” says Dunn. “That’s the lifeblood of the industry.”