Digital billboards that marry image recognition of car models with other data streams such as mobile phone data will soon display advertising tailored to individual vehicles.

Share story

WASHINGTON — Here is what’s around the corner: Smart digital billboards will detect the make, model and year of oncoming vehicles and project ads tailored to the motorist.

Roadside cameras will read license plates, and powerful computers will make snap judgments based on likely home address, age, race and income level to pitch products or services through the billboards.

Once ads flash up on roadside digital screens, the sales pitch may not stop. Any mobile phones in a passing vehicle may light up with a reinforcing message linked to the ad.

A series of factors is reshaping the quintessential experience of the road trip or job commute. Smart billboards are here, gracing the sides of bus shelters, urban interstates and pedestrian walkways. And as the digital billboards grow in size and number — rotating ads, customizing them to passing traffic and earning far more income — old-fashioned billboards face an existential moment.

Throw in artificial intelligence and powerful computers, and the roadside experience is on the cusp of change. Digital electronic billboards actually stare at us — and make judgments about who we are and how we might spend our money.

A big unknown: the impact of self-driving automobiles on both old-style “dumb” billboards and their smarter progeny.

“Often your car is a proxy for demographics. We get several ad agencies who say, ‘I want to advertise to affluent men over $100,000 (in annual salary) with XYZ education.’ Often driving a BMW or an Audi is a proxy for that,” said Kevin Foreman, general manager of geoanalytics at INRIX, a Kirkland company that gathers and sells real-time traffic information.

To determine make, model and year of cars on the road, startup companies marry powerful computing, roadside sensors or cameras and pinpoint advertising.

One of them is Synaps Labs. Its co-founder and chief executive, Alex Pustov, said the company installs roadside cameras roughly 600 to 650 feet in front of electronic billboards. The cameras feed images of oncoming cars through a cellular signal to a computer.

Packed in the computer’s memory are some 2,000 different images of each of 1,600 makes and models of cars, he said.

“Initially, it was labor intensive. We needed to collect millions of images,” Pustov said. “We manually created libraries of car makes and models.”

It only takes a second or so to transmit and digest the image and channel back a targeted ad that a driver would see for eight or nine seconds, Pustov said.

When multiple lanes are filled with traffic, the computer can determine broad groups of targets, say, owners of older automobiles, and flash ads accordingly.

“Most car companies want to advertise to seven- to 12-year-old cars. They don’t want to advertise to a 1- to 2-year-old car,” Foreman said. “Ford spending money on you when you’ve just bought a new Ford is lousy. But me, I have a 12-year-old Ford. I’m a great candidate.”

Smart billboards can also target motorists on the highway or pedestrians passing bus shelters by picking up cellular or mobile signatures, Wi-Fi signals or even beaconing given off by certain apps.

The billboard sector, or what the industry prefers to call “out of home” advertising, comprises $7.5 billion of the $185 billion annual U.S. advertising market, said Andrew R. Sriubas, chief commercial officer at Outfront Media, one of the nation’s big three outdoor advertisers.

Industry experts are cautious to note that the data harvesting is anonymous, hoping not to evoke the creepy billboards of the 2002 movie Minority Report in which a protagonist finds signage addressed to him directly.

“It doesn’t have to know who you are. It needs to know what you are. It says I see phone ID 453ABCD. I happen to know that phone number is associated with a millennial Hispanic female, therefore send it this ad,” Sriubas said.

Moreover, the data industry collects vast information about the whereabouts of mobile users by the apps on their smartphones, which share global positioning system, or GPS, signals every 15 seconds.

“When you click ‘I allow’ on your favorite mobile app, if they’re a partner of ours … you most likely are anonymously sending us your GPS point heading,” Foreman said.

That is partly why INRIX says it can anonymously track the GPS signals of over 300 million drivers in 65 countries. Moreover, one in four cars coming onto the road today emit their own GPS signals.

Smart billboards can consider other factors for targeting, such as time of day, weather conditions and upcoming events. A digital sign catering to pedestrians can also make judgments.

“It can detect gaits. So it understands male versus female, it understands kids versus adults,” Sriubas said. “There’s a bunch of very sophisticated algorithms that it can understand.”

Using that data, he added, “I know you’re male or female. I know you’re within a certain age category, 30 to 40. I know that you live in this location and you work in that location.”

Smart ads for cars may riff off the seasonal weather or time of day.

“For a Jaguar campaign that we did,” Pustov said, referring to a smart campaign in his native Moscow, “when it was snowing they were showing a Jaguar ad that demonstrated the car is very comfortable in the snow.”

When weather was better, and little traffic was on the road, “They were saying it’s a very powerful car.”

The rollout of digital smart billboards is far from uniform. U.S. municipalities, counties and states have different restrictions on placement, brightness and frequency of rotation of outdoor ads, and whether video — which can distract drivers — is permitted.

“There are significant safety issues. There are significant privacy issues that are still out there. Usually these kinds of issues get worked out,” said Dan Jaffe, head of government relations for the Association of National Advertisers, a marketing-industry trade group.