Optimizing for Viewability? Know the Risks

“Data Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.

Today’s column is written by Andrew Pancer, COO of Media6Degrees.

Viewability - it’s a word of the moment. I’m pro-viewability (it’s hard to not be!) but there are issues with establishing it as a new metric. Although viewability captures one component of ad effectiveness, when used in isolation, it can lead to a step back for marketers rather than a step forward.  If we aren’t careful, the viewability movement could get ambushed.

Here’s what I mean: As an industry, we have come light years in a short period of time on several evaluative metrics, such as brand safety, view-through conversions and brand lift. Viewability is important as well, but when used as an optimizing criterion, it can have a perverse effect. I could tell you that I’m putting your display ad on a publisher site where it is 100 percent viewable. But just because the ad is viewable doesn’t mean that I’ve chosen the right publisher, or that the environment is a brand safe, or that the ad has impact amidst the clutter on the page. When aggregated across good and bad sites, viewability scores can be manipulated and misunderstood.

Ask yourself this:  Would you rather have your ads run with 80 percent viewability on quality, brand safe sites, or with 100 percent viewability on sites that include non-brand safe environments? While the major viewability vendors produce detailed information about placements, they also provide summary reports that rank vendors based on viewability scores. Many may skip the detail in favor of the summary report, and optimize accordingly. Viewability then trumps brand safety and site quality -- a sub-optimal outcome for the marketer and a setback for the industry.

If quality pages are in demand today, viewable ad space on quality pages is at a particular premium.  When purchasing viewable media, marketers must ensure that they are buying space on quality pages.  I cannot stress this emphatically enough.  We’ve found in a series of studies of exchange inventory that viewability of available impressions is often inversely related to page quality.  Case in point:  Some link farms feature the best viewability on the exchanges, but they’re brand-debasing and ineffective places to advertise.

According to a recent study by eMarketer, the Internet’s media ad spending share is projected to increase by 8.2 percent in the next five years in the U.S. That’s an additional 8.2 percent of marketing dollars that could potentially be going towards supporting poor quality pages when programmatic buying is employed. That’s an additional 8.2 percent of the budget that advertisers may be reluctant to allocate towards digital, unless the industry becomes vigilant about the quality of pages in our networks and exchanges. Whether you are the publisher, the brand or the marketer -- ultimately you (and your ads) are all victims of low-quality, fraudulent and dangerous pages with illicit content.  Do you want your brand’s ad associated with that, even if it is highly viewable?

Follow Andrew Pancer (@apancer) and AdExchanger (@adexchanger) on Twitter.

9 Comments

  1. David Simon

    Well said. When you've got premium guaranteed buys only reaching 80% "viewable" metrics, you have to ask yourself if the metric, as a measure of quality publishing environment, may be flawed. Some of the best vendors in the space have begun to build more advanced measurement metrics that can measure and make targetable the data that distinguishes "premium publishing environments." Marketers should begin testing those data points to understand the impact to ROI, rather than settling on a simple, yet broken metric, like "viewability."

    Reply
  2. With respect to Andrew, this critique on viewability is a red herring. Viewability is not being viewed by the "Making Measurement Make Sense" initiative (3MS); its association sponsors (IAB, ANA, and 4As); the Media Rating Council; or the ~60 marketers, agencies, and publishers comprising 3MS's leadership as a grand unified standard for the optimization of advertising campaigns. Rather, it is a long-missing baseline - a minimal currency by which brand advertising can be transacted efficiently and effectively in digital environments and, importantly, planned, bought and understood in cross-platform environments with other brand-advertising media, all of which employ some version of viewability as a baseline standard.

    There is no question other forms of analysis can and should be applied to create a good media plan, and to measure its effects, the impact of the advertising content, and the return on the investment in the campaign. Viewability has never been proposed as a measure of a "quality publishing environment," so to call it a "broken metric" reflects a misunderstanding of the initiative's purpose and scope.

    Reply
    • Randy, Thank you for the comments. With all due respect I think you are misunderstanding my issue. At m6d we are big fans of the initiative and are actively implementing it into our daily operations. BUT...numerous clients of ours have shared what they have been receiving from their viewability vendors. And they are planning to use some of these reports to make optimizations regarding who stays on the plan and who does not. My issue is that these reports are very high level and do not tell the whole story. This leads to a concern that incentives will lead people to optimize towards viewable impressions regardless of quality. And in my opinion this would move the industry backwards.

      Reply
  3. Good point, Randall. I'd also like to add that while many exchanges may have brand-debasing and ineffective places to advertise, it is important not to generalize across the industry nor give marketers the impression that 3rd party link farms are the only options they have to work with. For example, Facebook pages are brand safe by providing protection against an ad appearing beside pornography, violence, etc. Advertisers working with an FBX 1st party retargeting partner like Triggit can also monitor performance by page type and optimize against them with transparent data (a feature that not all FBX partners are able to execute yet). And by the way, all ads on Facebook are viewable since they up and scroll down with the user. Here's some more thoughts about why savvy marketers shouldn't be repeating the past mistakes of programmatic buying across exchanges: http://blog.triggit.com/facebook-exchange-unlike-traditional-rtb/

    Reply
  4. Come on Andrew. There's no viewability bogeyman here. If any network, platform or other person/system is optimizing exclusively to viewability I'd spend more time questioning their optimization ability. It's just one of many (many!) things to optimize for.

    Reply
  5. In support of Andrew's comment we have seen precisely what he describes. An RTB client bidding on the most viewable inventory using Adsafe's powerful pre-bid data, massively increased their viewable impressions, but some illegal download sites turn out to have a lot of good viewability! As soon as the client started added targeting of inventory rated brandsafe by Adsafe, along with the viewability targeting, the bad sites disappeared and their clients were getting more ads seen, in better environments...

    Reply
  6. Paul, ExchangeWire

    Andrew, I agree, optimising campaigns off of technology that doesn't track 30% of the internet (webkit browsers: Chrome, Safari) is plain insane. Use it as a guide, fused with other data.

    Reply
  7. Kirby Winfield

    Guys, if an ad is not viewable, it cannot be effective. If it is viewable, it can be. That doesn't mean it is. Of course one should (and we do) look at page content, quality, time in view, dwell time, and other influencers to optimize.

    Viewability is simply a prerequisite to all other optimization metrics.

    Reply
    • Alejandro Correa

      Hi Kirby,

      It is true that an ad that is not seen is not effective. What is not so clear is whether the same holds true for placements where the majority of the ads (or a high percentage) don't come into view.

      For example, a placement located at the bottom of a page may have low viewability, but users who scroll down to where it can be seen may be so qualified that their conversion rates exceed those of less qualified/more viewable placements.

      This, of course, is perfectly consistent with your larger point regarding viewability being 1 aspect of optimization...

      Reply

Add a comment

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>