Social media sites want it.

Insurance companies and drug makers want it.

Fast-food chains, sporting goods companies and car manufacturers want it, too.

What is it? Your attention.

“That is really the key currency in these markets, but the gauge being used – the Nielsen measure – is imperfect,” said Jura Liaukonyte, the Dake Family Associate Professor in the Charles H. Dyson School of Applied Economics and Management, in the Cornell SC Johnson College of Business

Liaukonyte, Matthew McGranaghan, M.S. ’14, Ph.D. ’20, now an assistant professor at the Alfred Lerner College of Business and Economics at the University of Delaware, and Kenneth Wilbur, professor of marketing and analytics at the University of California, San Diego, tracked ad viewership using tools that, instead of just monitoring the television, measured actual viewer presence in the room, and focal attention on the screen. The research analyzed 4 million ad exposures over the course of a year.

Their findings – including the fact that nearly a third of TV ads play to empty rooms, and that viewers are four times more likely to leave the room than change the channel – are detailed in “How Viewer Tuning, Presence and Attention Respond to Ad Content and Predict Brand Search Lift,” which published Feb. 9 in Marketing Science.

Paying thousands of dollars to advertise on television is a huge proposition – never more so than for the Super Bowl, for which 30-second TV spots this year will cost advertisers as much as $6.5 million. These advertisers expect at least a little bang for their substantial buck, but how do they know what kind of bang their ad is generating?

Advertising pricing relies on measuring how many TVs are tuned in to a specific channel, and not whether people are actually watching the TVs.

For this research, the co-authors worked with TVision Insights, a TV performance metrics company that developed innovative technology to passively monitor who’s in the room and whether they’re actually looking at what’s on the TV screen, while respecting viewer privacy.

“TVision is providing long-needed updates to the metrics we use to inform TV ad buying,” McGranaghan said. “Advertisers care about people watching, not necessarily that the TV is on.”

“We wanted to quantify whether the current industry standard is doing a good job predicting what advertisers care about,” Liaukonyte said.

The metric referred to in the paper’s title – “brand search lift” – refers to viewers immediately heading to the internet to find out more about a product after seeing its ad on TV. The novel viewer presence and attention data predict online search response, while standard industry metrics do not.

“Analyzing actual attention in real-time shows that TV ads can drive brand search, a key intermediate metric of advertising effectiveness,” McGranaghan said.

Among other results, the team found that ad viewing behaviors vary depending on channel, time of day, program genre, age and gender. For example, older viewers are more likely to avoid ads by changing channels; younger viewers are more likely to avoid ads by leaving the room or diverting their visual attention – likely due to multitasking with a second screen.

Additionally, ads for recreational products – beer and video games, for example – do the best at retaining viewers, the researchers said. Among the worst at keeping eyes on the screen are prescription drug ads, particularly those for serious conditions.

The Super Bowl, of course, is a different animal from every other show in the TV ad realm, the researchers said. While TV ad viewership during commercial breaks generally tracks downward most of the time, McGranaghan said, their data shows it actually increases during the NFL’s big game.

“If you are not a football fan,” McGranaghan said, “you are likely watching for these very expensive, polished, and entertaining ads.”

Said Liaukonyte: “It has become sort of like the Oscars for the advertising industry.”