A Brief Guide to Viewability For Digital Advertisers

On January 5, 2005, an astronomer from the California Institute of Technology, saw an object in the sky, way out beyond our ninth planet, Pluto. This body turned out to be even larger than Pluto and was given the name Eris. Discovering Eris forever changed the way we view our solar system because it resulted in a re-categorizing of what we call a planet. As of 2006, Pluto and Eris — due to their comparatively small size and distance from other planets more local to our solar system — are placed in a new, secondary category called “dwarf planets.”

In the world of digital advertising, we are at a similar crossroads with ad “impressions.” Given our advancing technology and better understanding of consumer behavior, the advertising industry is now reconsidering how we define an impression in much the same way that astronomers re-evaluated the definition of a planet. We now have two types of impressions: served and viewed. Served impressions are served via technical means to a web browser, but a consumer may have never viewed them, whereas viewed impressions suggest that the ad has been both served and viewed by a consumer. Given this re-categorization, perhaps the time has come, as it did for Pluto, that an impression is no longer considered a dependable form of ad response measurement.

This topic is still brand new (and a bit unclear) to most people and there are more questions than answers, though the Interactive Advertising Bureau (IAB) seems to be working to sort things out. To help un-muddy the waters, we’ve put together this brief guide to viewability for digital advertisers.

What’s the big deal?

As it turns out, the big deal is this philosophical conundrum: Do advertisers pay for ads that are never seen by consumers? Frequently, the answer could be “yes.” Digital advertising, like radio, TV, and print, is loosely based on an ad’s “opportunity to be seen” by consumers. Just as someone listening to the radio may turn down the volume during an ad or someone watching television may go to the bathroom during a commercial, there is no truly scientific and completely accurate way to count each individual’s presence and attentiveness for the duration of an ad. Advertisers are buying the ads’ opportunity to be seen or heard by an estimated audience. Impressions are similar. Some ads may be featured at the very bottom of the web page, and are never seen by the consumer, much less read or properly understood. Previously, impressions indicated that the ad loaded into the page — nothing more. So, this is a big deal. In digital advertising where everything can presumably be measured, why are so many paid impressions not being viewed by consumers?

How are viewable ads measured now?

IAB considers an ad to be viewable on a desktop computer if 50 percent of the ads’ pixels are in view for a minimum of 1 second. For video, 50 percent is required for 2 seconds. For larger display ad units, 30 percent is required for at least 1 second.

Jeff Burkett, a product strategist at The Washington Post, wrote for Digiday about his views on what he calls “dark media,” where he calls out ad measurements firms, claiming that they miss out on 30 percent of viewable impressions through failed measurement. He refers to this 30 percent as “dark media” and claims it amounts to fraud by the measurement firm and lost revenue by the publisher who ran the ad. The MRC backs him up (indirectly) by saying that non-measured ads should not be considered non-viewable simply because they were not captured. Burkett told Digiday that it’s a “flaw in some vendor’s technology.”

What’s the IAB saying?

It is the Internet Advertising Bureau’s responsibility to be an advocate for both the publisher and the advertiser, and also to find some common ground by which to accurately measure an ad’s performance. For Randall Rothenberg, president and CEO of the IAB, a Town Hall in late 2014 was his opportunity to make a definitive statement on the organization’s future course of action in regards to viewability. He said, “100 percent (viewability) is currently unreasonable. Why? Because different ad units, browsers, ad placements, vendors, and measurement methodologies yield wildly different viewability numbers.” The IAB went on to suggest that 70 percent viewability was more reasonable given inconsistencies in measurement (perhaps indicative of Burkett’s 30 percent dark media claim). Basically, some ad formats are not consistently measurable. However, because accurate measurement makes digital advertising more compelling than any other form, time can only tell how quickly we can find an agreeable technology solution. No doubt it will happen…some say it already has.

How is technology being utilized?

Some say we have the technology today (perhaps we always have) to measure 100 percent of ad impressions. A commonly used programming language for the delivery of ads is Javascript, which “lazy loads” ads into browsers once the page scroll hits the ad, thus, preventing an ad from being loaded into a browser (web page) premature to its viewing.

Google’s Active View is a leading technology that claims to be able to measure 100 percent of ads, giving publishers the ability to charge appropriately for viewed and non-viewed ads. Moat is another that is measuring “in-view impressions,” meaning that the ad must be in-view of the consumer for 1 second or more, in alignment with the IAB’s new standards. With accurate measurement of an ad’s performance, publishers can appropriately charge for the ad without the mystery of paying for non-viewable ads or even worse, non-human viewed ads.

Impact of website and mobile design

Of course, there are those who blame web and mobile design formats that create ad space not mapped to users viewing patterns. Creating ad space below the fold at the bottom of the page doesn’t necessarily drive the type of viewable impressions that advertisers want.

Where do we go from here?

Late last year, the IAB put forth eight recommendations for both publisher and advertiser to find common ground in measurement of impressions and actual viewability. Although they were heavily publicized and praised, they are worth repeating here in full:

IAB principles for 2015 for the road to viewability:

  1. All billing should continue to be based on the number of served impressions during a campaign, and these should be separated into two categories: measured and non-measured.
  2. Given the limitations of current technology, and the publisher-observed variances in measurement of 30–40 percent, it is recommended that in this year of transition, measured impressions be held to a 70 percent viewability threshold.
  3. If a campaign does not achieve the 70 percent viewability threshold for measured impressions, publishers will make good with additional viewable impressions until the threshold is met. Such a guarantee ensures that all paid measurable ad impressions will be viewable at a threshold that both exceeds the minimum standard and falls within observed variances. To illustrate how the 70 percent threshold will work, assume: a campaign delivers 10 Million served impressions; of those served impressions, 8 Million were measurable; of the measured impressions, 5 Million (62.5 percent) were viewable. In this case, the publisher would need to deliver 600,000 additional viewable impressions to reach the 70 percent threshold and make good. The agency will be billed for 10 Million impressions assuming full make-good.
  4. All make-goods should be in the form of additional viewable impressions, not cash, and should be delivered in a reasonable time frame. Make-good impressions should be both viewable and generally consistent with inventory that was purchased in the original campaign. Determination of threshold achievement is based on total campaign impressions, not by each line item. In other words, some line items may not achieve threshold, but others can compensate.
  5. For large format ads, defined as 242,500 pixels or over, a viewable impression is counted if 30 percent of the pixels of the ad are viewable for a minimum of one continuous second.
  6. All transactions between buyers and sellers should use MRC accredited vendors only.
  7. A buyer and a seller should agree on a single measurement vendor ahead of time. The industry aspires to variances of no more than 10 percent between viewability measures provided by different vendors. All stakeholders must avoid costly, labor-intensive, error-prone manual processes of reconciling different sets of viewability numbers, hence the benefits of agreeing on a single vendor.

Lori Goldberg is the CEO of Silverlight Digital.