Skip to content
Apr 25 / Nathan Levi

Hooray for the viewable impression!

What with the Olympics and the Diamond Jubilee soon upon us, you might think there couldn’t be much more to celebrate about. You’d be wrong. The viewable impression metric has arrived! (cue the erection of bunting and the arrival of a deviled egg platter). For those of you who are not digital media geeks some explanation might be required before you burst into rapturous applause and scoff down your coronation chicken.

The metric that we use to trade display media is the impression. An impression happens when an online advert is served to a website and is ‘seen’ or viewed by the user. Or so you might think. The advert doesn’t have to appear on the website (it may load after the user has vacated to another website) and it doesn’t need to appear ‘in view’ (meaning that it might load on the page but the user has scrolled down or up and has missed it). Both these scenarios are pretty common. So common in fact that according to some reports up to 68% of display adverts are never seen by the user. In effect most of the impressions that advertisers buy are never even seen, let alone clicked on. This is hardly a surprise considering the volume of ads that are served and the ease of which it is to ignore them.

So what I hear you cry! A huge number of TV ads are never seen too. Think of how many people leave their living rooms to go to the loo or put the kettle on during Corry (not me, I’m more of a Glee man). Whilst this might be true, advertisers are paying because the ad is available to be viewed. This doesn’t translate online as many ads are not available to be viewed, and are often served so far down a web page that they are hard to find, let alone see. This is the main reason that brand advertisers are not investing more money in display. Not only does no one click on the damned things (except in error), but most ads aren’t seen either.

Publishers like Facebook are desperate for brand dollars. After all 63% of all media is spent on branding. The drive to justify the brand impact of display media is critical to the future growth of all content providers online, especially Facebook (who have managed to dwindle us with an entirely unmeasurable metric called ‘engagement’ for too long). We are all chasing this quest to find out if online display really packs a punch, we are searching for the truth and won’t rest until we know.

The viewable impression, which seeks to measure impressions that have actually been viewed is regarded as one answer to solving this issue. If we know that our ads have actually been seen then surely we will be one step closer to understanding the broadcast impact (or limitations as the case may be) of online media. Some detractors of the viewable impression have argued that this will simply force publishers to raise their rates. Well I say so what! Surely advertisers will be happy to pay for ads that are actually seen. Furthermore, with all this talk of attribution, surely the most important missing piece is to understand the impact of ads that are actually viewed and their potential role in the purchase funnel. All attribution models right now account for impressions whether they have been seen or not which is a complete waste of time. Giving credit to a view which has not really been ‘viewed’ is like awarding Barack Obama the Nobel peace prize (yes, that happened too).

There are of course many technical challenges to measuring whether an ad is in view as most ads are served through iframes. This is set to change however, and I think we’ll see more publishers allowing their ads to be tracked as in-view. We may then be closer to that ever unknown truth about display media and it’s real impact as a branding vehicle and a broadcast medium. We may even find out the unimaginable, and that might not be a bad thing in this opaque world we call advertising.




Hooray for the viewable impression! Viewable Impression (CPMV) viewable impression online media digital marketing Brand Awareness Advertising

One Comment

leave a comment
  1. Frank Walter / May 13 2012

    I trialed comScore’s VCE product for their APAC case study recently. What we found: viewable impressions varied wildly by publisher and creative size with webmail inventory performing worst; 70% or more is a good benchmark for a network; there’s gold below the fold – some of this inventory had high view rates; our DSP did much better than most of the networks.

    What I’d like to see next:
    1. Viewable impressions measured by ad servers not 3rd parties – this would make it much more straightforward to improve attribution models.
    2. Viewable impressions data synced with DSPs – in any auction the best informed bidder will make the wisest bid.

    Which leads me to ask: why didn’t Google buy AdXpose instead of comScore?

Leave a Comment