MIAMI BEACH, Fla. — At the Possible conference, where optimism flows as freely as rosé and every panel promises to reinvent advertising by Tuesday, Tim Vanderhook, co-founder and chief executive of Viant Technology, delivered a reality check. Some of your TV ads aren’t just underperforming. They are performing to absolutely no one.
“We see 20 to 30% of the ad impressions, no one’s actually in the room,” Vanderhook said, casually detonating a statistic that should make every media buyer reconsider their life choices.
This revelation comes courtesy of TVision Insights, which Viant acquired and integrated into its demand-side platform. The system doesn’t just confirm that an ad aired. It checks if a human being was present and, more uncomfortably, whether that human was looking at the TV or doomscrolling instead.
The deal combines TVision’s attention data with Viant’s existing identity and content layers. The result is a system that aims to measure performance across linear TV, connected TV and even the walled gardens, without relying on each platform to grade its own homework. It also introduces an “attention-adjusted CPM,” which prices inventory based on real human engagement rather than the comforting fiction of served impressions.
Somewhere between sessions and iced coffee refills, some might call that “a vibe shift.”
From eyeballs to actual eyes
Vanderhook framed the combination of TVision’s person-level attention data and Iris TV’s content-level signals as a turning point. Or, less grandly, a long-overdue upgrade from educated guessing to something resembling reality.
“To have all these signals integrated into a DSP is where it starts to change planning, buying and measurement,” he said in an interview with Beet.TV contributor David Kaplan.
The pitch is simple. If an ad runs during high-attention, co-viewed content, it should cost more. If it runs while viewers are in the kitchen or arguing with their phones, maybe not so much.
“If the attention is really high and the co-viewing is really high, that content owner is gonna see higher CPMs,” Vanderhook explained, adding that the inverse is also true.
In other words, not all impressions are created equal. Some are premium. Some are basically wallpaper.
CTV finally tries to prove it works
Connected TV has spent years promising performance, often delivering very nice charts about reach. Vanderhook argues Viant’s identity infrastructure changes that equation.
“One of the great advantages of CTV is addressability,” he said.
With household-level IDs, advertisers can target specific audiences and, crucially, measure outcomes in a way linear TV never could. The example he offered was telling. If data shows golfers are more likely to engage with a product, campaigns can focus on that audience across CTV.
It sounds obvious. It also sounds like something the industry has been circling for a decade.
AI runs the bids, humans keep the anxiety
No Possible discussion would be complete without an AI segment, and Vanderhook did not disappoint. He described a spectrum of adoption that ranges from “we trust the machine” to “legal would like a word.”
“AI bidding has the highest adoption, over 90% customer adoption,” he said, noting that machines are better at adjusting prices in real time than humans who insist on sleeping.
Planning, however, is another story. Adoption sits closer to 30%, largely because media plans still span multiple channels that AI does not fully wrangle yet. And fully autonomous campaigns remain more aspiration than reality.
“That final piece of letting AI run the whole campaign… we have the least amount of adoption there, and it’s really a question of trust,” Vanderhook said.
Translation: marketers love automation until it spends their budget in creative ways.
Winners will have something you don’t
On consolidation, Vanderhook offered a blunt assessment that cut through the usual industry jargon.
“What do you have that’s exclusively unique to you?” he said, describing the defining question for the next generation of ad tech.
The answer, he argued, is either exclusive data or exclusive inventory. The walled gardens already have both. Independent players, including Viant, are racing to secure the former through acquisitions like TVision and Iris.
Still, Vanderhook suggested the next frontier might not be data at all. It might be something older and less programmable.
“At a certain point… we need to go back to the foundational principles of advertising,” he said, pointing to creative as the next area ripe for reinvention.
Yes, after all the algorithms, identifiers, and attention metrics, the industry may rediscover that ads need to persuade people.
Revolutionary.
You’re watching Beet.TV coverage of POSSIBLE 2026. For more videos from this series, please visit this page.
