Looking Forward To A World Ruled By The Viewable Impression

Data-Driven Thinking“Data-Driven Thinking” is a column written by members of the media community and containing fresh ideas on the digital revolution in media.

Joanna O’Connell is Senior Analyst at Forrester Research.

Ad impressions will drop by 50% or more. CPM’s will increase commensurately. ComScore 500 publishers will finally get the respect they deserve and recapture market share from their junky ad network rivals. Consumers will start noticing – dare I say liking? – display ads. Display will no longer be the red headed stepchild in the shadow of direct response rival search or brand rival video. That’s what we can look forward to in a world ruled by the viewable impression standard.

Why am I on my high horse about viewable impressions? And why do I think it’s such a big deal for the industry? Well, we in the display advertising world have a real problem, and it’s been plaguing us since the earliest days of the medium: how to think about, understand, and make sense of the true value of a display ad impression. Our initial answer, back in the late 1990’s, was to say, “Let’s just measure what we can – clicks!” and our addiction to the click through rate (CTR) as the key measure of display advertising’s success – or more often failure – was born. We should all be kicking ourselves for that, and I know many of us are: We took a brand new marketing vehicle, ripe with promise and untainted by the rules of marketing channels that came before it, and turned it into a cheap, bottom-of-the-barrel direct response mechanism.

Who did this serve? No one.

It didn’t serve publishers, who were – and still are – under constant pressure to prove that the display banners they sell in their premium, proprietary environments have intrinsic value.

It didn’t serve advertisers, who were trained to be reticent that there IS any intrinsic value to ad exposure in and of itself. (Full disclosure: working in display planning and buying for as long as I did, it was hard to sit idly by and hear the value of display be constantly questioned.)

That’s why my interest is so piqued by the growing industry call for a “viewable impression” standard. Witness the IAB’s “Making Measurement Make Sense” initiative, where it’s tenet no. 1; comScore’s new Validated Campaign Essentials offering (post AdXpose acquisition); and Google’s very recent announcement around Active View, its own spin on the viewable impression.

I concede that we’re far from having THE answer when it comes to defining, let alone evaluating, a viewable impression (see the ongoing, heated debate among players like C3 Metrics and Yieldex on this subject), and certainly, I also concede that there arguably remains significant technical work to be done to get to numbers we can all believe.

However, I view it as a big win for marketers and publishers alike that we’re finally having a healthy dialogue about the value of an impression that is seen. So why do I see this as a great thing in the long term?

Publishers who believe in their product have little to lose in committing to selling only those ad placements which users have overtly, and measurably, been exposed to. It simply supports the promise of the premium publisher: that high quality content and a sticky environment creates loyal – and valuable – audiences, just the kind of folks marketers want to get in front of with their messages. Forbes clearly buys this argument, recently publicly announcing adoption of the comScore toolset to power its “validated Brand Increase Guarantee” program.

Marketers can finally stop debating whether or not there’s any value to ad exposure, and start focusing on quantifying that value relative to their business objectives, for example by determining the right view conversion window to use given their product’s purchase cycle – or focusing on the contributing impact that those ad impressions had on later conversions through down-funnel channels like search – rather than completely writing off the possibility of any impact.

So who DOES lose long term in a viewable impression future? Cookie bombers, attribution stealers and other “junk car” display salesmen (you know who you are). Consider yourselves on notice.

Follow Joanna O’Connell (@joannaoconnell), Forrester Research (@forrester) and AdExchanger (@adexchanger) on Twitter.

Enjoying this content?

Sign up to be an AdExchanger Member today and get unlimited access to articles like this, plus proprietary data and research, conference discounts, on-demand access to event content, and more!

Join Today!


  1. To say that switching to viewable impressions will drive CPMs higher seems wildly speculative to me. I believe we will simply see a reduction of impressions with little to no change in price.

    Viewable impressions are antithetical to what happens in all other media. Magazines get credit for all impressions in a magazine per circulated magazine. This industry is used to proxies, and viewable impressions push us further and further toward some more granular accountability standard that is once again trying to set a higher bar for online without any evidence that it will change anything. I also think this is not something tied to what large brands would cite as a reason they aren’t moving budget online.

    The primary problems to solve are creative formats and efficiency of media buying. Not limiting how we count impressions. Instead we should fix those two problems and focus on other value we can provide as a strong driver of brand adoption.

    Viewable impression gives me a “meh” kind of reaction. I predict a bunch of effort without significant payoff.

    Eric Picard
    CEO, Rare Crowds

    • Jeremy

      so just because magazines lack the ability to measure we should stop all efforts to take advantage of this technology online?

      regardless of how you determine the effect of your ad spend; everyone can agree that unseen ads have zero effect. There is no analysis required.

      Companies that do things the right way (i.e. pubs who don’t create BTF placements, or networks who buy quality inventory) will ultimately be rewarded because advertisers will insist on “in-view” technology.

      My concern is that agencies have invested heavily in “programmatic buying” through their trading desks and will resist broad adoption of in-view technology. After all, the exchanges have virtually no above the fold inventory. If you don’t believe start asking your DSP about it.

      • Matt Greitzer

        I don’t think trading desks will resist. There are plenty of ways to target viewabilty through programmatic buying. In fact, it’s probably easier and more cost effective for a programmatic buyer to price in the required premiums/discounts for viewable vs. non-viewable impressions than for a planner to account for this in an upfront buy. You might get a publisher to agree to 100% viewability, but they are going to charge a huge premium to do so.

        By the way, can we ditch the “above-the-fold” and “below-the-fold” labels when talking about viewability? Below the fold impressions can be viewable, too. (Consider scrolling down on a page).

      • Jeremy

        No one doubts that targeting viewable inventory through programmatic buying is possible. The problem is the complete and total lack of supply. That inventory just doesn’t filter into the exchanges at scale. Ad Networks will always have an advantage over a trading platform in this regard because they go direct to the pub and can offer to buy the viewable inventory at a price that is just above the yield the pub is getting for that inventory in the exchange.

      • An ad impression is defined as the opportunity for a consumer to see an ad, not that the ad was seen by that consumer. My only complaint is that as an industry we keep setting the bar higher for online than advertisers have requested. We’re shooting ourselves in the foot over and over again.

        I think if the buyer wants to implement a verification service and negotiate with the publisher/network to only pay based on that verification services’ numbers – that’s their business to do. But I’m not a fan of this trend toward setting higher bars than are needed. Viewable ads are not the area that are keeping dollars from being spent – it’s never been raised in front of me by a marketer despite hundreds of conversations asking about moving budgets from traditional to online.

      • Winifred Hooper


        It’s SSPs/Exchanges that most commonly have direct relationships with publishers (just like ad networks), not trading platforms. Also the inventory they have access to (compared to ad networks) is directly dependent on the clearing CPM. In fact it’s likely in this current landscape that they’re seeing more premium (viewable) inventory than the ad networks.

        In fact, through private relations (not just private exchanges) trading desks form with publishers (via the SSPs/Exchanges) they can guarantee spend and are more likely to access more premium (viewable) inventory than if they went through an ad network given in this instance the publisher is able to set their own floor for each buyer (absent of the ad network). And of course this doesnt even begin to consider the controls and transparency publishers are able to unlock when they for private programmatic relationships with buyers (over running through ad networks).

        Let’s face it
        1- these viewable impressions sure as hell aren’t going to be given to the ad nets
        2- it’s up to buyers (in bringing the necessary spend and floor CPMs to the table) to push publishers to unlock the inventory to be bought programmatically, as otherwise its straight back to site direct buys

        Of course, to Joanne’s point, having the tech to properly classify in a standardized fashion, and then accurately bid on all of these viewable impressions remains an entirely different story.

        -Winifred Hooper
        I’ve read ALL the reports

      • Jeremy

        If I’m willing to pay in bulk $.01 CPM higher for in-view inventory than what the exchange/SSP is currently yielding; what pub wouldn’t sell to me?

      • Winifred Hooper

        Yes, but chances are if the ad net is willing to pay a $0.01 higher CPM for guaranteed in-view impressions (which assumes they have buyer demand against it), than that same buyer would likely pay an incremental $0.02 CPM for that same in-view impression with the additional transparency & control of knowing the exact domain it ran on, plus being able to overlay whatever 1st or 3rd party audience segment they wanted and buy it programmatically via a consolidated platform.

        I have a very tough time believing ad networks’ big comeback is going to be through the window (pun intended) of viewable impressions

    • Eric,

      The magazine analogy simply does not hold water.

      Advertisers happily pay for all magazine ad impressions because there is a guaranteed “opportunity to see” each impression. In other words, the ad has been printed and is extant on a page.

      There is a tacit understanding that the reader may or may not view each page; in fact, that is why front of book placements sell for higher CPM’s than those on later pages. What matter is, there is a chance for the ad to be seen.

      Conversely, online ad impressions that appear “off-screen” haven’t been “printed” and in fact do not exist. Only when they appear within the viewable area of a browser is there an “opportunity to see”. Advertisers should expect only to pay for impressions that provide said opportunity.

      Finally, there is a huge disparity in inventory quality and supply between magazines and Web content. I suspect online advertisers would swallow an unviewable impression rate of ~15% if it was assured all impressions were running on quality, branded content. But today the rate can be as high as 90% of impressions served out of view, on no-name sites buried beneath layers of ad traders. This problem simply does not exist in traditional media because it’s nowhere near as fragmented, and it’s 100% transparent.

      A viewable impression standard goes a long way towards solving for this problem, and helping clear the path for brand dollars, *especially* as streaming TV begins to overtake cable/traditional TV over the next decade.

      Put yet another way, an impression served out of view is like a 30 second spot that shows only dead air. Brands simply won’t go for that.



      • Kirby –

        A. All I’m saying is that we’re spending too much time trying to fix the wrong things. If magazines were a viable economic model without ‘view’ ads, clearly we have a good case in point.

        B. Your analogy of ‘the ads haven’t been printed’ if the user didn’t view them is flat out wrong. If the page is rendered, it’s been “printed”. The current IAB impression guidelines make this clear – and if you’re counting on the page load, then the page has been rendered and is viewable by the user if they scroll down.

        C. CPMs reflect the lower value of a banner verses a full page magazine ad. I don’t buy your 90% number. I could imagine a high percentage when looking at specific inventory sources(remnant exchange or ad network only), but I seriously doubt that 90% number overall, and certainly if you look at guaranteed inventory bought from a publisher that number doesn’t fly at all.

        D. An impression served out of view is EXACTLY like a magazine page that wasn’t opened. It is NOT like a 30 second spot that shows dead air, which implies a user viewing a blank screen with no content on it.

        If you’re still in Seattle, let’s grab a coffee or beer and argue it out in person. 🙂


  2. Eric, I agree that a viewable impression is not the holy grail of display measurement, but we’re never going to be able to divorce ourselves from the granular levels of accountability that only display can provide. I firmly believe that engagement should be the standard as it solves the viewability challenge (you can’t engage with an ad you can’t see), solves the creative format challenge because publishers and agencies will be inspired to offer formats that drive interaction and it solves the media buying inefficiencies by optimizing to the only metric that’s been proven to influence consumer behavior.

  3. I agree with Jeremy – saying that it’s not the right answer because it’s not how it works in other media (which is debatable) is like saying, “hey, the status quo is awesome! if it ain’t broke, don’t fix it!” That’s antithetical to digital media – we are innovators. We change every day, often for the better, occasionally for the worse. What we all need to be asking ourselves every day is, how do we create a healthy, sustainable ecosystem that supports the needs of both publishers and advertisers? Moving to a model where pubs are rewarded for inventory and advertisers are too, that’s a win-win. It hurts the crappy players, the shady guys… yes. But I’m cool with that.

  4. chainguy81

    If an ad placement is consistently not ‘viewed’ by a user, then the impression should be worth less. When placing a bid on an impression, the buyer is taking a chance on the ad being clicked on or the ad later resulting in a conversion event. The point is, the buyer will not be over paying for the impression. If DSPs could create a backend goal based on the probability to view an ad (so sellers could actual sell on this pricing model), this would be added value, but I do not see viewable impressions becoming the standard in a programmatic buying world.

  5. The impression today is not a standard. It can be easily inflated: double the ad units on a page, double supply, and gamed: I continue to hear countless reports from the demand side about excessive quantities of cookie-matched impressions generated in invisible iframes or through other less than ethical means. If the currency that we all trade on is to gain more share from other forms of media, the impression has got to start to resemble something a little more reliable.

    Also, let’s not forget that the issue is not just cutting the overabundance of supply by exposing impressions that aren’t in view. Every measurement model we use to today instantly gets better if we can prove that a user actually saw a client’s message when subsequently attributing conversions or lift to that impression. Much investment has gone into working around this problem, rather than just fixing what’s fundamentally wrong with much display measurement today – the impression.

    • Winifred Hooper

      Also we haven’t even begun talking about the pop-unders yet!

  6. Kit Merinda

    This all points forward towards thinking less about viewable impressions, and more about viewer actions. As things evolve, the difference between an online impression and a magazine ad is that there can be a large variety of ways for a viewer to interact with the ad (in fact, choosing the types of interactions, based on audience and goals, will become a fundamental part of designing an online campaign).

    Once you have those interactions structured, it’s another important step — for both advertiser _and_ publisher — to then quantify the results of those interactions, by thinking carefully about where in the funnel of customer acquisition each interaction is taking place, and weighting and measuring how valuable that interaction is to the whole process (you can begin to see why construction of such a campaign would have to be quite sophisticated and data-driven in design).

    This turns what used to be a static impression into the first phase of customer engagement, and as such, could turn online display into an absolutely powerful tool in the whole process, far beyond what thinking about it “just” as advertising accomplishes right now.

  7. First of all, are all below the fold banner placements are considered non viewable impressions? Below the fold placements are certain not worth as much as above the fold but not all of these are non viewable. Thus the plot thickens. What happened to those below the fold impression that are served in an Iframe. How you know that this impressions was seen by a user scrolling down the page? As I know it, such scripting techniques experiences difficulties when such script are served in an Iframe. Not all banner placements are served with Java Script tags. Lots of web sites out there still uses Iframe to deliver their banners. Just one Iframe call within the mix, may not be the last call, will messed up the script ability to tell if that impression is now within the user’s desktop screen viewable area.

  8. Have any of you actually seen a report with so called viewable impressions? While I totally get the benefit, the technology currently is immature. ~70% of the impressions have no information at all!

    We’re pretty deep on serving up super-fast tags and we don’t know of a easy way to increase the number of reads.

  9. Comparing impressions to magazine advertising is a poor proxy. First, the print business is shrinking while the online world is growing, but more importantly, there is an ecosystem of companies trying to measure effectiveness of the print advertising world. The reality is we operate in a segment of the advertising industry that is measurable and optimizable. We might prefer the “soft” reporting in the print world, but that is not reality. If one could prove that a particular page of a magazine was never seen by any reader, the advertisers would run from that page. The print business can not supply that data, but that does not stop advertisers from wishing they had it.

    For digital advertising, the cat is out of the bag. View-ability is part of the future landscape. SSPs, exchanges, and DSPs will adapt and incorporate view-ability data soon enough so programatic buying will allow for view-ability metrics. The challenge for the publishers will be to understand the value of this metric and how to influence view-ability.

    Publishers already struggle with the variety of goals and metrics used to measure “success” today and this will add more complexity. However, for publishers that leverage this complexity to their own advantage, they will win.