Is 100% Viewability 100% Necessary? Maybe Not
- by Tobi Elkin @tobielkin, May 9, 2016
Goodway Group, a managed-services programmatic media firm, on Monday released “Applying Viewability Intelligently for Measurably Improved Performance,” a white paper whose findings run counter to the industry’s 100% viewability standard.
“The research clearly shows the benefits a viewable impression has over one that isn’t viewed. No one doubted that,” said Jay Friedman, Partner and COO of Goodway Group, in an email to Real-Time Daily. “What the paper also shows is, the amount of ‘gray viewability’ is frustrating, [which complicates] any ‘case closed’ answer on whether or not 100% viewability is beneficial, reasonable and necessary.”
Friedman said that “gray” impressions are measured differently by different accredited vendors. Rather than providing a binary yes/no, the viewability outcome is a “maybe.”
“It’s for this reason we’ve concluded that requiring 100% measured in-view impressions by a single vendor is likely pushing the benefit of viewability measurement past the point of diminishing returns for marketers,” Friedman told Real-Time Daily.
Testing viewability’s impact on ad performance by analyzing more than one billion delivered impressions, Goodway’s research identified four key findings:
1. Using two viewability technology companies, Goodway determined that consumers who are served ads that are measured as viewable convert at an 8% to 9% greater rate than those who are served ads that are not measured as viewable.
2. A large gray area exists in viewability: impressions that are measured as being in-view for some amount of time are not enough to meet the Interactive Advertising Bureau (IAB) standard.
3. Contrary to intuition, sites that have nearly 100% viewable inventory are the worst-performing sites in the programmatic ecosystem.
4. Viewability’s lift on performance can’t be compared with other dimensions’ lift, such as behavioral data or site/context. The research found that viewability is, instead, the basis for enabling all other performance to catalyze a campaign.
Goodway maintains that “viewability” has become a buzzword “and a default for marketers,” but his research found that when campaigns are built solely around viewability, they often miss the mark--to the detriment of advertisers.The white paper shows that 100% viewability doesn’t automatically lead to conversions. “The programmatic ecosystem requires that multiple factors, including viewability, be considered to achieve optimal campaign results,” Friedman said.
The report recommends adopting a multidimensional
approach when executing campaigns: place high-quality creative in effective media vehicles, accept a margin of error when it comes to viewability, and cherry-pick viewable inventory in the private
marketplace.
Maybe this is correct for static print-style "display" ads---though I tend to doubt it--- but it has no bearing on visual, story telling ads. It doesn't matter wether the way digital ads are placed mitigates against 100% "viewability". Fine. The advertiser should pay only for those "exposures" where a 15-second or even a longer commercial might have been seen in its entirety---that's from start to finish. Even that is being generous as we don't know how quickly the user began to watch the commercial---- even if it was fully shown. If digital ad sellers think they can peddle TV's branding advertisers anything less than 100% viewability they are sadly mistaken.
Is this still a work in progress? Perhaps the industry needs more time and measurement of different points along a continuum to reach agreement on what optimal viewability is, for video and display, and under what conditions, similar to determining optimal frequency for ads.
Henry, I agree in principle, however the basic issue is that there are several "industries" involved on the buying side---search, promotional, classifieds, direct response and branding TV-style. The latter represents a small share of digital ad spending and therefore is in a "minority" position. Even if the other segments get together and decide that the IAB's definition of "viewability" is valid---and they are willing to pay in full, or use some agreed upon discounting scale for dealing with ads which are truncated or not shown in their entirety, that is not going to fly with most TV advertisers who wish to tell consumers a short story---say limiting themselves to 15 seconds. Getting a few words out, with a few frames shown on the users' screens just doesn't cut it for such campaigns. It's like being asked to pay $12 to go see a big Hollywood movie, only to find that only the middle or only the first half of the film appears on the big screen. How many moviegoers would be happy with that? How many would settle for getting only half of their money back?
The problem is that different viewability vendors have different success rates in measuring impression-level inventory. We have seen plenty of instances where Moat says an impression is not measurable (and attributes 0% viewability to the impression) but AOL or IAS says it is...or vice versa. Individual publishers may or may not have installed code on their sites to support propoer vendor measurement, and each vendor has a different methodology anyway, which leads to varying results. If an agency client wants to optimize on other KPIs like completion rates or clicks, viewability is one solid factor but can sometimes lead to sub-optimal results -- like the article says, some of those "unmeasurable" impressions can actually convert quite well on KPIs that matter to the client. It depends on what the client really wants. I agree with the author that too much focus on viewability has diminishing returns, and there is some benefit in working with more than one vendor -- perhaps take a weighted average of 2 or 3 solutions and then remove the inventory layer that consistently under-performs across multiple measurement solutions.
This whole conversation is so interesting because all I hear is how measurable digital online advertising is for clients and the industry hasn't even agreed on a standard for what's considered viewable.