A Viewable Impression is an ad served where at least 50% of the pixels were on-screen for at least 1 second (display) or 2 seconds (video) — the IAB-standard threshold separating served from actually-seen.
A viewable impression is an ad served where at least 50% of the creative pixels were on-screen for a continuous second (display) or two seconds (video) — the threshold the IAB and Media Rating Council set to separate ads that had a chance of being seen from those that technically fired but never reached anyone's eyes.
The mechanism is straightforward. A measurement tag runs client-side JavaScript to track what percentage of the ad unit is inside the active viewport. The moment the ad clears 50% visibility, a timer starts. If it stays above that threshold for the required duration, the impression counts as viewable. Miss either condition — not enough pixels, timer not reached — and it remains a served impression only.
What makes this matter in practice is the gap. On open-web display, roughly 25–40% of served impressions never clear the threshold. That means a campaign reporting 4M impressions might have delivered viewable reach closer to 2.6M. When we look at in-market campaign data on AdLibrary, CPM calculations made on served numbers routinely overstate effective reach by 30% or more.
The 2025–2026 measurement landscape adds complexity. Advantage+ placements and Andromeda's cross-surface delivery mean your creative can land in environments where IAB measurement tags have reduced fidelity — in-app, CTV, and short-form video don't always support the same viewability instrumentation as desktop web. Meridian and Robyn MMM frameworks increasingly weight viewable impressions (not served) as the base signal for reach curves, so viewability errors compound into model error downstream.
For deeper context on how AI tools are changing measurement, see AI analytics tools for marketing in 2026. For creative implications — including how ad format affects viewability by default — AI image generation for ads covers format choices that affect on-screen time.
Treat viewable CPM, not served CPM, as your apples-to-apples comparison unit across vendors.
Most paid-media reporting blurs "served" with "viewed." In my experience auditing campaign data, the gap is bigger than teams expect — typical display campaigns see 25–40% of served impressions never cross the viewability threshold. CPM-based decisions made on served numbers overstate real reach by 30% or more. That's not a rounding error; it's a pricing and allocation mistake that compounds across every vendor comparison you make.
