2012 was the year of ad viewability, and with the hype came a lot of questions. Who is measuring viewability most accurately and how? Can it really be used for targeting in real-time bidding or only as post-campaign measurement? And will improving ad viewability really increase marketers' ROI?
Even the Media Rating Council (MRC), the body behind the cross ecosystem movement toward viewable impressions standards, raised its own concerns, by suggesting that these questions and more must be answered before the industry can fully accept new standards for viewability.
Here are four key misconceptions plaguing the industry's understanding of viewability:
Recently we've noticed a lot of discussion focusing around sample sizes for measuring viewability. Some vendors claim to measure viewability on a very high percentage of the impressions they have access to -- i.e., their sample size for measuring viewability is very large. These vendors seem to imply that this makes their measurement more accurate than that of other vendors.
The thing is, viewability is not reported for each impression -- it's reported on an aggregate level (campaign, placement, publisher, etc.). We have a 90 percent sample size, but we also know that size doesn't matter when it comes to statistical accuracy as long as the sample is large enough and unbiased. This means that when dealing with millions of impressions, 10 to 15 percent sample sizes are equally as accurate as 90 percent. This is how the industry has been doing other types of measurements (i.e., audiences) for years.
It's also important to remember that, at the end of the day, sample sizes are irrelevant if the measurement methodology isn't as accurate as possible. In other words, focus on "how" the vendor measures viewability vs. "how big" they claim to be measuring.
In its recent Viewable Impression Advisory, the MRC warned:
"Cross domain iFrames pose a large obstacle to measuring viewability. While partial solutions exist and others are under development, no vendor known to MRC today can determine viewability in all cross domain iFrame situations."
It's easy to see why the MRC is worried about cross-domain iFrames, as they restrict access to crucial parent data information such as URL, coordinates, text, and so on. Nearly 90 percent of commercial display ads live within cross-domain iFrames. It's a complicated problem to solve, and many players in the industry found themselves behind the 8-ball as viewability moved to the forefront of the conversation in 2012.
But, the MRC got this one wrong. Solutions for determining viewability through iFrames do exist. Ours, for example, was built from the ground up over three years ago with the ability to see across 98 percent of the commercial internet. Advertisers and agencies evaluating viewability vendors should make a point to ask tough questions related to iFrames, as this will be a key factor in determining the quality of their solution.
Using viewability measurements for post-campaign reporting and analysis does not, in itself, guarantee significant ROI improvement. Measurement and analysis simply don't scale enough.
The industry's growing understanding around viewability issues has exposed the urgent need to find an accurate and feasible methodology for measuring ROI. Properly measured ROI should only account for the conversion events that have been definitively caused by a given campaign. Clearly, an unviewable impression couldn't possibly be the cause of any action by a consumer. That said, existing last-click, last-touch, rule-based, or algorithmic attribution models still don't distinguish between viewable and unviewable impressions in defining what led to a conversion event.
These models, therefore, lose track of any causal relationship between impressions and conversions. The recent introduction of the "viewable conversions" approach -- which simply discounts all unviewable impressions for the purpose of conversion attribution -- still does not help much in finding causal conversions and, hence, in measuring a display campaign's effect on overall ROI.
The future of viewability is in the industry's move towards programmatic buying -- in other words, using viewability as a predictive metric to target and optimize the media that marketers bid on in an RTB setting.
According to the current convention adopted by IAB and MRC, a viewable ad impression is defined as an ad that is 50 percent in view for more than one second. While it makes sense that any ad placement not satisfying the above definition is very unlikely to be viewed, it's less clear whether this definition should universally apply to all web pages.
Other factors can play an important role in an ad's viewability. For example, despite an ad being on a consumer's screen for the desired amount of time, there can be other distractions competing for his or her attention, such as a video or 5 to 10 other ads. Being in-view really only means that an ad has the potential to be viewed.
While it seems to already have been discussed at length, technically, viewability is still in its infancy. The industry is working hard to establish best practices that will help us get there. More importantly, we must recognize viewability as a key stepping-stone for this industry, not the final solution. There are many other aspects of ad environment and exposure quality for marketers to consider when trying to engage a potential customer and create resonance with their campaign.
David Hahn is SVP of product management for Integral Ad Science.
On Twitter? Follow iMedia Connection at @iMediaTweet.
Not a People Connection member?
Full Summit Calendar | Request Invite
1 5 ad technologies that will be dead in 5 years
2 6 signs your agency is dying
3 The best social media campaigns of 2013
4 The most meaningless (and hilarious) job titles on LinkedIn
5 8 types of problem clients (and how to handle them)