Definition
An ImpressionOne display of an ad or search listing, as counted by the platform. is counted when an ad is served. Whether it could have been seen is a separate question, answered by a measurement script running alongside the ad. The industry standard, maintained by the Media Rating Council, defines a viewable impression as one where a minimum share of the ad's pixels was within the visible area of the browser or app for a minimum continuous time, with separate thresholds for display and video. The exact values are set out in the council's guidelines, which are the source to read before relying on any particular number.
Viewable impressions ÷ Measured impressions × 100(%)
- Measured impressions are those where the script was able to run and decide whether the standard was met. Some environments block or cannot host the script, so measured is a subset of served.
- Viewable impressions are measured impressions where the answer was yes.
- Unmeasured impressions are left out of both counts. They are not assumed viewable or unviewable; they are unknown.
Why the denominator is measured, not served
Dividing by measured impressions keeps the rate honest about what was actually checked. It also creates a trap: a seller that measures only a small share of its inventory can report a high viewability rate on the part it measured, while most of what you bought went unchecked. Measurement coverage, the share of served impressions that were measured, must be read beside the rate every time.
How to interpret it
Higher is better, but only once three things are known. First, the coverage behind the figure: a rate on well-measured inventory is evidence, a rate on thinly measured inventory is a sample. Second, the format and placement: large formats and placements near the content people came for tend to be seen more than small units at the edge of a page, so rates are compared within one format. Third, who measured it: an accredited verification vendor and a seller's own report are not the same evidence, which is why the media auditing and verification guide treats independent measurement as part of the buy.
The most practical reading converts viewability into price. A CPMCost per 1,000 impressions. buys a thousand served impressions; dividing the same spend by viewable impressions only gives a viewable CPM, which is always higher. Two placements with similar CPMs and different viewability are not the same price. The impression calculator tutorial for programmatic deals shows how an all-in viewable CPM translates into delivery.
An outdoor furniture brand recounts two exchanges
Quillon Outdoor Furniture buys display on two exchanges at about £4.00 CPM, spending £2,000 on each for 500,000 impressions. Exchange A measures 450,000 of them and 315,000 are viewable: a 70% viewability rate on 90% coverage. Exchange B measures 100,000 and 80,000 are viewable: an 80% rate on 20% coverage. B's headline rate is higher, but only 80,000 impressions are known to be viewable, against 315,000 on A. Cost per thousand known viewable impressions is about £6.35 on A and £25.00 on B. The brand requires measurement on every deal and moves the B budget to placements that can be verified.
What moves it
- Placement on the page. Slots that load where the reader is looking, or that stay in view as the page scrolls, are seen more often than slots far down a long page.
- Ad format and size. Larger units and formats that fill more of the screen meet the pixel threshold more readily.
- Page and app design. Slow loading means the ad may render after the reader has scrolled past; infinite scroll and cluttered layouts reduce time in view.
- Measurement coverage. Changes in which environments can host the script change the denominator and so the rate, without any change in what people saw.
- Invalid traffic. Bots can fake a viewport and register views; filtering IVTAd impressions and clicks not caused by genuine human interest: bots, fraud and accidents. can lower a reported rate while improving the real one.
- Buying controls. Pre-bid viewability targeting and deals that guarantee viewable delivery change which impressions are bought at all. How those controls are set in one platform is covered in DV360 inventory and deals.
Related metrics
ImpressionOne display of an ad or search listing, as counted by the platform. counts are the raw material viewability filters, and CPMCost per 1,000 impressions. becomes viewable CPM once the filter is applied. IVTAd impressions and clicks not caused by genuine human interest: bots, fraud and accidents. and Brand SafetyControls that keep ads away from content that would harm the advertiser's reputation. are the other two checks verification vendors run on the same impressions: one asks whether a real person was there, the other whether the surrounding content was acceptable. A clean, safe, viewable impression is the minimum a buyer should pay for. Beyond viewability sit attention measures, which try to estimate whether the ad was actually looked at; the guide to viewability and attention explains where one stops and the other begins, and brand safety and viewability fundamentals shows how the checks combine in a programmatic buy. For awareness campaigns, the DV360 awareness guide treats viewable reach rather than served impressions as the delivery target.
What good looks like
The site cites no benchmark beyond the standard itself. The council's threshold defines what counts as viewable; it does not say what share of impressions should meet it, and rates differ by format, placement, device and how the inventory is sold. Judge the figure against your own deals over time, always beside coverage, and through the price it implies: the cost per thousand known viewable impressions is the number that lets two placements be compared fairly.
Common mistakes
- Reading the rate without coverage, so thinly measured inventory looks better than well-measured inventory.
- Dividing by served impressions in one report and measured impressions in another.
- Treating viewable as seen. The standard is an opportunity to see; attention is a further question.
- Comparing display and video rates that are judged against different thresholds.
- Accepting a seller's own measurement where an independent vendor's is available.
- Optimising to cheap viewable impressions in placements that meet the standard and do nothing for the brand.
