For roughly two decades, digital advertising worked because of one quiet assumption: the device is the audience. When someone scrolled a feed, the device was the person. When an ad loaded, the impression was the proof. The pixel could be trusted because it sat at the exact intersection of media and human attention.
That assumption built the modern advertising stack. It also built the modern measurement stack. Panels, identity graphs, ACR, attribution platforms, MMM models, set-top-box telemetry — all of them were architected on top of the same idea. Measure the device, and you've measured the audience.
The problem is that the most valuable media in 2026 doesn't live there anymore.
It lives in shared environments. Living rooms with three people on the couch and one on a phone. Sports bars with forty people watching the same game on six screens. NFL stadiums with seventy thousand attendees inside a single PA system. Retail floors with shoppers walking
past audio-enabled displays. YouTube playing on the television while two adults and a kid drift in and out of the room. In every one of those environments, the screen and the human are separate things. The household is not the person. The play is not the exposure. And the device, increasingly, is not the audience.
The measurement industry knows this. It has known it for years. The response has been to layer modeled multipliers on top of household data, to estimate co-viewing with vendor-specific assumptions, to count attendance as a proxy for sponsor exposure, to use foot traffic as a stand-in for in-store media impact. Each of these solutions is reasonable inside its own assumptions. None of them, on their own, can produce something the industry quietly needs and rarely names: a per-moment, room-level, human-verified record of exposure in a shared environment.
That's the blind spot. And it's worth pausing on how big it actually is.
U.S. connected TV advertising is projected to reach approximately $38 billion in 2026, growing roughly 14% year over year. In the 2026 upfronts, CTV ad commitments are projected to exceed primetime linear TV upfronts for the first time. The global sports sponsorship market is on track to move from approximately $70.2 billion in 2025 to roughly $74.6 billion in 2026. U.S. retail media is projected at $69.3 billion in 2026, with retail-media CTV alone growing 45% in 2025. ATSC 3.0 / NEXTGEN TV is now live in over 80 U.S. markets, reaching roughly 75% of the U.S. population.
Each of those markets transacts on measurement that cannot independently prove a real human was in the room when the moment mattered.
That isn't a criticism of the existing measurement layers. Panels are statistically powerful for national trend analysis. ACR is remarkable at recognizing what is on a screen. Set-top-box telemetry proves what was tuned. App analytics prove what was opened. Every one of those measurements is real. The point is that none of them was designed to measure presence — the actual co-location of a real person with a specific media moment in a shared environment. They measure everything around the moment. They were never built to measure the moment itself.
There are three reasons this gap is becoming harder to ignore in 2026.
The first is co-viewing. Roughly 47% of all linear and connected TV time, according to Nielsen, is consumed by more than one person at a time. Industry research has found that as much as 88% of CTV viewers report co-viewing experiences. Yet across the industry, co-viewing is currently estimated with modeled multipliers that vary by vendor, by program, and by platform. The same household, watching the same show, can be counted as 1.5 viewers on one platform and 1.2 on another. The industry has, in effect, standardized inconsistency. As CTV crosses the
size threshold where ad commitments exceed linear primetime, that inconsistency becomes a procurement problem.
The second is sponsorship. Sponsors are no longer satisfied with attendance numbers, logo seconds, and modeled media value. They want to know, at the per-moment level, whether the people in the building and the people watching at home and the people watching at the bar actually saw the moment they paid for. Rights holders are starting to need to prove it. Sponsorship measurement is one of the oldest categories in the industry, and one of the least solved.
The third is privacy. Persistent device IDs are eroding. Raw-audio listeners are increasingly unacceptable to platforms, regulators, and consumers. Cross-app fingerprinting is being deprecated by every major platform on a known timeline. The next generation of measurement infrastructure has to be privacy-native by default — not retrofitted. Most of the systems built for the previous decade weren't.
Add those three forces together and the picture is hard to argue with. Shared-environment media is the future of media. The measurement stack was not built for shared-environment media. And the gap is widening faster than the existing layers can close it.
We've spent the last few years building toward closing it. Not by competing with the measurement currencies, the ACR vendors, the attention platforms, or the clean rooms — all of which do important work — but by producing the one input none of them can produce on their own. A privacy-native, audit-grade, room-level verified record that a real human was acoustically co-present with a specific media moment.
We call that record a verified exposure event. We think it's the missing primitive at the bottom of the measurement stack. We'll have more to say about exactly what we've been building in the coming weeks.
For now, we wanted to start the conversation, because the conversation has been a long time coming.
The pixel proved the device. The next decade of media will need something that proves the human.
[global_section id="333"][global_section id="333"]
[global_section id="333"]