AmEx: We Don’t Agree With Industry Viewability Standards

visible manThe divide between publishers and advertisers around viewability sharpened during a Tuesday morning panel “Inside The Mind Of The Advertiser,” hosted by analytics provider Integral Ad Science.

This wasn’t much of a surprise considering one panelist was Ari Bluman, GroupM’s chief digital investment officer for North America. GroupM has taken a notoriously hardline stance around viewability, insisting on only paying for ads where all pixels are in view.

“An ad that’s not seen is worth zero,” Bluman said. “Not less. It’s worth zero.”

Many on the sell side feel this position is too draconian. Matt Prohaska, a New York Times vet and CEO of Prohaska Consulting, told AdExchanger in a previous interview that even non-viewable ads have value since they can still cookie a user.

Bluman said this might be worthwhile for some media buyers, but not for an agency as large as GroupM, which he said has all the data it needs.

Rachel Herskovitz, global media manager at American Express, echoed Bluman’s sentiment around non-viewable inventory.

“We have a zero-tolerance policy too, and we’re working through what’s possible with that zero-tolerance policy today,” she said. She expects publisher partners to be on the same page in terms of rigorously monitoring ad quality.

“We’ll be thinking about how that works when we go to market and buy, and it’s very important for our publisher partners to have zero tolerance as well,” she said.

Herskovitz also acknowledged multiple hurdles around buying viewable impressions or avoiding fraudulent impressions.

Last March, the MRC gave the okay for companies to transact online display around the basis of viewability, defining it as an ad that’s 50% in-view for one second.

But Herskovitz said the standard – often considered a jumping-off point meant to catalyze further discussion around viewability – is inadequate.

“I don’t agree with MRC standards,” she said. “Their definition of viewability doesn’t make sense. No one spends a second viewing an ad. If that becomes our standard, we’ll build technology around that and I think that’s a mistake.”

Yet she agreed there needs to be a consensus standard accepted across the advertising industry.

“If every brand and agency has a different standard, we’re all screwed,” she said, adding it’s still unclear what the right standard is.

And even if both sell-side and buy-side entities unite around a single viewability standard, there are different measuring methodologies, just as there are multiple vendors accredited by the MRC.

Herskovitz said she was aware of the differences, stating that her top criteria is transparency.

“If we’re not measuring viewability the same way a publisher is measuring viewability, they need to be transparent about it, then we can talk about it and solve it,” she said. “We want to work with as many partners as we can, but we can’t do that unless there’s transparency. That’s the biggest baseline for us.”

But there’s another point of contention between publishers and advertisers: Publishers want to charge more for viewable inventory. This policy is a non-starter for Herskovitz.

“Why should we pay more for something that needs to be seen?” she said. “That doesn’t make sense.” She harkened back to the way TV advertising is sold: If a television ad airs, the advertiser pays.

Of course, the online ad economy is more complicated. Carol Chung, SVP of media technology at Digitas – which works with American Express – said that depending on campaign goals, inventory with low viewability can perform well. (“Allegedly,” Bluman interjected.)

Chung’s point was that the desire to pay more for highly viewable inventory and less for inventory with low viewability is too simplistic.

“You can’t have a conversation about pricing and just talk about viewability,” she said. “To set accurate pricing, you need to go all the way through to the end, and there are multiple levers to pull. There’s the performance, plus the pricing, plus the viewability, plus whatever frequency factors you have.”

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. The only folks who think 100% viewed ads are bad are bad publishers and middle men whose businesses depend on obfuscation and tonnage. Advertisers and quality pubs have no issues here. Let’s move away from a currency that promotes fraud and quantity over quality and move to one that promotes efficacy, measurable impact and consumer experience.

  2. Please show me any report that shows a campaign delivered 100% viewability. Doesn’t exist.

  3. I find a bit ridiculous the unwillingness for buyers to shoulder any of the costs of adopting these viewability standards via pricing changes. There are large fixed and variable costs for publishers to implement new technology and new processes to deliver against a viewability standard and thereby deliver new value to buyers. What’s next? Will buyers demand new technologies to seek out and eliminate every possible impression that won’t yield a return on THEIR branding or performance initiatives? On the publishers dime?

    Secondly, “all pixels”? How about variable pricing based on the length of time viewed factored by percentage of pixels viewed? Maybe factor in the brightness on the user’s screen? Or number of people looking over the user’s shoulder? I’m only half kidding, but my point is that Herskovitz can’t pick out a couple of variables to take a hardline stance in her own favor on and then pretend to be reasonable by wanting convergence on standards and better transparency.

    Agreed with the latter part of Adam’s comments – the current standards are contentious because they just don’t do a good job of aligning pricing with value; they are too simplistic. There are so many more measureable factors that could be incorporated to create a pricing structure that buyers and sellers could both endorse.

  4. On TV? imaging saying that to TV folks: I’m not paying you for the guy (even worse, “panel guy”) that switched channel or went to the bathroom or FFW’d the ad or is looking at his phone/table/laptop…. LOL

  5. Agree with another commenter, we are focusing on the wrong thing. Purchase decisions need to be made around the campaign goals and performance. I can certainly understand a advertisers challenge here, but there are too many flaws in this entire system to focus on just one.

  6. The only ones who are really going to win from the new viewability measures are the tech vendors. We’re adding an additional layer of complexity which will, necessarily, increase the cost and effort to serve ads. For good honest publishers it’s a bust because the biggest burden lands on them, and because the Agencies (as evidenced by the sentiments in the article) will use this as a hammer to reduce their CPMs. For agencies it’s a fail because they become obsessed with a new measure that ISN”T PERFORMANCE. Just like click rate and then ad verification before, this is another poorly conceived measure that has nothing to do with the effectiveness of the ad spend. All online ad deals should be bought based on time or traffic percentage. E.g. you get the home page for an hour on Tuesday morning, and 25% of the back sections for the week. No more reconciliation, no more impression fraud, no more click fraud, and all the technology becomes much much simpler.

  7. Viewability is simply one alternative measure of a digital. There is over emphasis in the industry on its broad adoption.

    The empirical data I’ve seen supports Ms. Chung’s proposition that viewability matters for some campaigns (and others not so much). It depends on the campaign goal and cost-benefit analysis for the client. I think Mr. Blumen is doing his clients a disservice by taking such an extreme position without the data to back it up (allegedly).

    I would add that while a zero visible ad certainly can add value by “setting a cookie”, this is a dated way of thinking about the overall problem and solution. After all, why bother placing an ad – just buy 1×1 pixels to set cookies across the web and get that last click or impression. Better yet, implement fractional attribution and you’ll really see the value of cookie bombing.

  8. Hi Ryan,

    Thanks for continuing the discussion on viewability. I wanted to share some clarifying points to help untangle some of the confusion around viewability and in response to Ms. Herskovitz’s comment about the definition of the viewable impression. The viewable impression was never intended to – and does not – serve as a measure of user engagement or an ad’s effectiveness.

    Rather, it’s intended to serve as a measure that indicates an “opportunity to see” was established with the ad’s delivery. The viewable impression for display ad.’s one-second time requirement was determined after extensive study of hundreds of millions of data points in various environments and contexts. This allowed us to identify norms for when individuals begin to recognize and act upon digital ads, things that can only occur subsequent to the establishment of the opportunity to see the ad. Based on this research, we arrived at one second as the minimum time required to qualify as a viewable display impression precisely because it proved to be just prior to the time when people began to recognize and act upon the ads.

    Viewable impressions represent a starting point in the development of other elements of our long-range plans to enhance digital measurement, on which MRC is working in collaboration with the Making Measurement Make Sense (3MS) initiative principals. For example, the viewable impression will now serve as a baseline building block for other metrics that will more fully incorporate duration (such as digital GRPs), as well as metrics that will better lend themselves as measures of ad engagement and effectiveness.

    Thank you,
    George Ivie, CEO, Media Rating Council

  9. @Josh, their issue isn’t that the ads delivered weren’t viewable, it is whether they pay for them or not.

    I can understand where buyers are coming from and if they don’t want to pay for those non-viewable impressions, it’s their money, don’t pay for it. However, I don’t think this should be looked at as a ‘discount’ on your current media buy either and should be viewed rather as a factor that produces an eCPM.

  10. For Publishers – The issue is around the GUARANTEE of 100% viewabilty. That isn’t to suggest that an advertiser couldn’t optimize away from sites with poor viewability. They should. Publishers need to deliver value and it is difficult to do that if a majority of the ads aren’t even seen. But just as there are different price thresholds for cpm, cpc, cpa, and cpe shouldn’t there be a separate price for CPvM? This isn’t being against viewability it is around the guarantee of 100% and the increased management of that guarantee. Also missing for this discussion is the roughly 20% of ads that viewability measurement tools claim aren’t viewable but are in fact in view due to iframes or other elements not currently managed consistently by the various viewability vendors.

  11. Albert Einstein said it best, “…not everything that can be counted counts….and not everything that counts can be counted”. This is exactly what viewability is. I am not saying publisher should not be held responsible for ad quality. However, viewability should not be a campaign KPI. Because 1) 100% viewability is technically impossible 2)even an ad is 100% viewable, it does not mean it generate positive impact on campaign performance, engagement or sales. Over emphasize on viewability may in turn, hurts performance.

  12. Very interesting debate – my personal view is that today CPMs factor the fact that 40-50% of the ads are not viewed so if buyers decide to target higher viewable impressions they should be ready to pay more and this shouldn’t impact their performance as logically they will need less impressions. That might be a candid view but as an industry we need to work hand in hand and not use new metrics just to pressure even more one side of the ecosystem…

  13. Hulk Smash Viewability

    Advertisers doesn’t want to pay for a non-viewable impression? Fine.

    However, we, as the Publisher need to maintain our revenue to stay in business in order to provide advertisers with the brand safe content that they want, so if there’s less inventory, then the price will go up. So yes, you will be paying a higher CPM for less impressions so we can back out to the same exact eCPM we were getting before all this nonsense.

    And…Don’t ever run another Print advertisement again & we’ll see what you’re really not willing to pay for. By the way, how was the conversion rate on that Bus Stop Ad? Did anyone click on it? Were you able to re-target them? No? Hmmm.

    Also, you can forget about running your 3rd party JavaScript and 4th party Rich Media tags. That’s over. Welcome back to sending us jpg’s and we’ll be billing off the 1st party ad server reporting. What? You don’t trust Google?

    While we’re at it, no more 970×250, Skins, Wallpapers, Push downs, Portraits, or 300×600 ads. We’re going to a 25x25px ad size model. If all you care about is a percentage and a time measurement, then the size of the ad is irrelevant. If it doesn’t perform, well, that’s not our fault, fire your creative director and make a better ad.

    Now I will also be rotating your ads after every second. Heck, I might get 50 ads to deliver in one page view session because I met the stupid minimum threshold. Thanks auto-refresh!

    ComScore? Nope. Not going to allow you to use them to measure viewability. They actually admit to forecasting the likelihood of viewable impressions for 25% of their measurable methodology. My cousin Nancy has free time and a computer, she can do a much better job if you’re looking for someone to measure viewability (she only forecasts 10%).