Viewability: was the ad even on screen?

For years, advertisers paid for every ad that was served, even if it loaded at the bottom of a page nobody scrolled to, or in a browser tab nobody opened. Viewability asks a simple question: did the ad have a chance to be seen? An impression that meets the standard is called a viewable impression.

A billboard on a highway is useful. The same billboard stored face-down in a warehouse was still "delivered" to the city, but nobody could see it. Viewability separates billboards on the highway from billboards in the warehouse.

The standard, in numbers

The Media Rating Council (MRC), working with the IAB, published viewable impression guidelines for desktop in 2014 and updated them in 2015, followed by mobile guidelines in 2016. The core rules are:

These are deliberately low bars. They define an opportunity to see, not proof of attention. Some large buyers, most famously GroupM, have negotiated stricter standards such as 100% of pixels in view. The viewability rate of a campaign is simply viewable impressions divided by measured impressions.

How viewability is measured

1. A script travels with the adOn the web, a ad verification vendor's JavaScript tag runs alongside the creative and checks its position relative to the visible part of the screen.
2. In apps, a shared SDKThe Open Measurement SDK from IAB Tech Lab lets one integration inside an app report viewability to many vendors consistently.
3. Timers and geometryThe code tracks how much of the ad is on screen, whether the tab or app is in focus, and for how long, then reports whether the threshold was met.
4. Measured vs unmeasuredSome impressions cannot be measured, for example when the tag is blocked. Reports should show the measurable rate separately, because a high viewability rate on a small measured share can mislead.

What lifts or sinks viewability

• Placement: ads near content people actually read, and sticky units that stay on screen, score higher than ads at the very bottom of a page.
• Lazy loading: loading an ad only when the reader is about to reach it raises viewability and speeds up pages.
• Page speed: a slow ad may finish loading after the reader has scrolled past it.
• Format: video that auto-plays out of view, or ad refresh in background tabs, drags rates down.

Viewability beyond the browser

On connected TV, viewability works differently. A TV ad fills the screen, so pixel geometry says little; what matters is whether the ad actually played, whether the TV was on, and whether anyone was in the room, which measurement cannot know directly. With server-side ad insertion, the ad server may not even be able to run a measurement script. That is why CTV buyers lean on completion data, device-level signals and fraud checks rather than a classic viewability rate. In audio and digital out-of-home, the idea is replaced by audibility and by estimated audiences passing a screen.

A worked example: a campaign serves 10 million impressions, of which 9 million are measurable. If 6.3 million are viewable, the viewability rate is 70% of measured impressions. If 5% of all impressions were also flagged as invalid, the honest figure to optimise against is viewable, valid impressions only.

Buying on viewability

Many buyers now pay only for viewable impressions, a model often priced as vCPM (cost per thousand viewable impressions), or set a minimum viewability target in their insertion order. That shifts risk to the seller and rewards good page design. Publishers in markets from the UK to India and Australia have redesigned layouts around it, often cutting the number of ad slots but raising the price of each.