More Than 56% Of Ad Impressions Are Not Seen, Google Says
- by Laurie Sullivan @lauriesullivan, December 4, 2014
As marketers become more aware that Web site visitors never see many of the ads served on publisher sites, verifying that someone actually saw the online advertisement becomes a greater concern.
Since an ad served does not necessarily equal an ad viewed, Google conducted a study of its display advertising platforms Google and DoubleClick, which revealed that 56.1% of all impressions Google
measured in a recent study were not seen, and the average publisher's viewability rate came in at 50.2%.
Google conducted research using its Active View technology to learn about the current state of viewable impressions on the Web. It reveals five factors that affect the extent to which ads are seen.
A viewable impression occurs when 50% of an ad's pixels are on screen for one second, per the Media Rating Council and the Interactive Advertising Bureau. The data used in Google's study is based on display ads in desktop and mobile browsers, and does not include mobile in-app or video ads. The data was collected in July 2014 and October 2014.
The research finds that the most viewable position is right above the fold, not at the top of the page. The most viewable ad sizes are vertical units, and page position is not always the best indicator of viewability.
Not all above-the-fold impressions are viewable, while many below-the-fold impressions are seen. Median viewability for above-the-fold ad units is 68%, whereas median viewability for below-the-fold ad units is 40%.
Viewability varies across industries. The highest viewability rate typically belongs to the sites with more captive engagement.
Finance has a viewability rating of 43.6%, travel, 42.6%; home and garden, 42.5%; beauty and fitness, 42.3%; law and government, 42.2%; world localities, 42.0%; auto and vehicles, 41.9%; book and literature. 41.5%; and shopping, 40.8%. Reference sites hold the highest at 51.9%; compared with online communications at 48.9%; and games at 48.4%. From there it drops to people and society at 43.6%.
While the viewabiity rate ranges across content, verticals and industries, the content that holds a user's attention has the highest viewability rating, per Google.
Less than 12% of publisher sites have viewability rates of 30% or less in Google's and DoubleClick's display network, per the findings.
From experience with Google that has nearly gone hostile, Google AdWords, not AdSense, told me 3 years ago in January to take down 2 of the 3 banners Above the Fold or I would pay a penalty. I refused. Note Sweepstakes Today was one of the original websites invited by AdSense in a test market about 8 years ago to put 3 banners Above the Fold. Google stated that they would double our revenue which they honored. The redesign cost to our website was about $40,000. We earned this back within a year so it was win-win deal. NOW is a different story. They retaliated for not removing the 2 banner by taking away quality ads and a average amount of revenue of 400 percent. Less than we were making than before the first change. This is why many publisher are going to the RTB who don't have ad placement restrictions. So this story's cherry picked info is nothing more than a smoke screen by Google to cover up for their own mistakes. Google in the end will lose the good will and money from thousands of publishers if they continue down this path.
It's good to have such information out there for all to see, but one wonders whether Google realizes the negative impression that these revelations are causing, especially among branding advertisers, who are just now waking up to the "viewability" problem.
Although this is interesting data, it's still just one of many KPI's out there. What's interesting to me is with the internet, there is so much data out there that we can have these conversations, where as in the "old days", I couldn't complain to a newspaper that I wouldn't pay their rate base unless they guaranteed me everyone read page D19 of the paper when my ad appeared. Because nobody knew! That all said - like I mentioned - viewability is really only one thing you should be looking at with your ads. If your rates are lower, take that into account and weight the publishers viewability against that. If you're collecting data on your targeting (which you SHOULD be doing), weight the value of the data collection... Remember, whether the user scrolls to the bottom of the page and physically sees your ad or not is inconsequential in data collection. Once that ad loads and places the cookie - you now have data on the consumer whether they saw the ad or not - and that holds value as you get deeper into behavioral and remarketing strategies, or just as simple as whether to know if the audience on that particular site is even the right audience or not. So yes, I want all of our advertisers ads front and center and everyone noticing them, but if they aren't viewable, I also know how to salvage the situation where in the days of no metrics, I was merely guessing.
This would be particularly damning data... if "ads being seen" was any sort of success metric for an advertising campaign.
Instead, this is yet just another reason why advertisers and agencies alike need to continue to drive home the fact that the success metrics for a campaign are never diagnostic metrics such as this, but KPIs that tie back to how a business makes money - revenue, sales, leads, ad impressions, and so on.
Does that mean that you don't look at diagnostic metrics such as impressions, clicks, CTR, CPC, ranking, or "views?" Of course not. These are metrics that the people actually running the campaign look at every day as a way to evaluate and optimize a given campaign - but they were never designed to be the end game and none of them should be considered a "KPI" (because they are not).
That said, what are we even comparing this number to in the first place? The headline that brought me here (from the newsletter) claimed "Google: Majority of online ads are never viewed" which is neither factual or helpful. What's the view rate for other media options? Print? TV? OOH? Do we know if 56% is bad or the best anyone has ever had in the history of advertising media?
The ad business really wants to be pissed about this number, but the other reason they would be was because they've been poorly managing their campaign all along and this is "proof" of why things don't work (rather than them admitted that they suck at advertising).
Jeff, you ask about comparability with other media, which is a fair question. The problem is that you probably won't like the answer. Take TV, for example. When an advertiser pays for a 15- or 30-second announcement it almost always runs from start to finish on the viewer's screen---an occasional technical foul up being the exception. Therefore the TV norm is 100% viewability in the ad's entirety. The same goes for radio and print media. The ad always appears in its entirety. When Google reports that 50% of the Internet ads it tracked are "viewable" it's talking about a totally different standard and, frankly, one that would never be acceptable to a TV, radio or magazine advertiser. Just think about it. To be "viewable" an online display ad need only have half of its content on the user's screen for one second. That's absurd. Of course, unknown numbers of TV, radio and print media audiences avoid or skip over ads. The same is true for online users. The difference is that virtually all TV viewers, radio listeners or magazine readers could have noted the ads directed at them and, in the case of TV or radio, they could have seen or heard a complete ad message if they wanted to. This is not true for online users who don't even have this opportunity in a shocking percentage of cases. As a guess, I would estimate that a typical online ad "viewability" factor---if calibrated in a directly comparable manner to "legacy media"---that is 100% viewability for the entire message----is probably in the 10-20% range. If that is the case, Imagine what this does to intermedia CPM comparisons.
I agree Ed. To take the magazine and newspaper examples you raised one step further, it would be like being charged based on print runs as opposed to audited paid copies. You'd be paying for pulped copies that never made it onto a truck. Or with television saying ... well we broadcast the signal ... it's not our fault the TV was off.