Come Super Bowl Sunday, one Boston-startup will use facial recognition technology embedded in hundreds of area televisions to track how participants respond to the commercials, a primary draw for many viewers tuning in this Sunday night to watch the Denver Broncos take on the Carolina Panthers.
It’s how TVision Insights plans to help advertisers better understand their audience on one of the most-watched nights in television.
TV spots are notoriously expensive during the Super Bowl. This year, NBC is reportedly charging companies $5 million for 30-seconds of airtime. Companies shell out to get their products in front of the game’s mammoth audience, which last year approached 115 million people in the most-watched single broadcast in U.S. history.
“The way that it’s viewed is unique, attention during commercials is a lot higher than it usually is,” said TVision chief revenue officer Dan Schiffman. He said that through an analysis of viewer attention, “we’re able to make deductions and get more effective, creative advertising that’s geared toward the right audience.”
TVision works through a sensor housed in a small box beneath a user’s TV set. The sensor tracks how many people are in the room, when their eyes glance to and from the screen, and when they smile or frown.
Come game time, the company will be monitoring 600 Boston-area individuals who agreed to let TVision install its hardware in exchange for compensation from the company. Participants’ age, gender, ethnicity, and education level, as well as household income and political party affiliation, are considered as TVision determines what sort of people respond to certain ads.
Schiffman says his company does not store any images its sensor capture and does not keep any information that could potentially identify participants.
TVision’s data can be analyzed at the individual person level, but Schiffman said this technology could someday allow broadcasters to beam personalized commercials into different household in real time.
“It’s happening faster than you think,” said Schiffman.
Schiffman’s prediction for what kind of ads fans will see this Sunday? Ads that have a plot, a conflict, and a cast of characters. In other words, he said, “Ads that don’t seem like ads.”
Correction: A previous version of this article incorrectly stated the technology TVision uses to gather its data, and the manner in which that data can be analyzed.
