Average is Never Good: A Call to Embrace New Viewability Best Practices
By Eric Wheeler, CEO of 33Across
Billions of dollars are at stake for publishers' advertising revenue. The reason? They are at the mercy of the "domain average" used to establish viewability scores.
Media buyers typically measure ad views based on the average viewability of a website domain. That means they monitor a website but don't account for how key issues – such as load times, ad placement, website layout and historical averages – can distort the metrics, leaving the publisher with a watered down score and the advertiser with a less than optimal experience.
Using a domain average is not an effective way to measure viewability.
The good news is that advancements in viewability are being made. For example, several vendors such as OnScroll have emerged that won't even load an ad unit unless they know it will appear in view, while others differentiate through demonstrated higher viewability metrics. Still, the antiquated domain average metric remains the norm. The truth is that in-view ads are worth a lot more than the out-of-view ads.
As an industry, online ad buys should use the viewability of individual units, not the historical average of all ads on websites.
A Closer Look at the Problem
Today, viewability rates are calculated by the average load percentage of all ad slot positions (top, side and bottom) over a period of time.
For most viewability measurement companies, the evaluation process looks something like this: ad slot number one loaded 90 percent, while ad slot number three only loaded 20 percent of the time. When these two numbers are averaged, this website receives a 55 percent average viewability rate. However, slot one is way more valuable than slot three. Publishers that are engineering high viewability into their sites may not be getting enough credit for this effort and other publishers may be propping up slot threes with higher than actual average scores.
Why are we evaluating the entire domain versus the individual ad? It may be easier, but the easiest path isn't always the best path. What des this average score mean to publishers and advertisers? Not much.
Solutions: A Lens into the Future of Viewability
As the industry demands higher viewability standards, publishers will need to ensure they can deliver. Here are three ideas for improving on the viewability benchmark today:
+ Make the shift from buying via domain average measurement to buying via ad unit measurement. It delivers a much truer ad view metric that benefits the industry as a whole. It benefits advertisers with more specificity for better targeting and measurement. For publishers, it boosts credibility through viewability. By supporting a pricing model that drills down into the ad unit level as opposed to the domain average, publishers will attract more and higher level advertisers.
+ Find good viewability partners. Publishers should work with viewability vendors like Integral Ad Science that make publishers more attractive by scoring the overall and specific ad unit viewability score for their site.
+ Redesign for Web and mobile. Publishers need to redesign their sites to fully monetize them. The good news is that this is already starting to happen. AdAge recently reported that facing pressure from advertisers, dozens of media companies, such as Conde Nast, are re-evaluating where display ads appear to ensure viewability.
Publishers should think about new placements, not just the standard placements in the same content blocks. They should be creative with how ads are inserted into the natural flow of content as viewed on various devices.
Let's clear up a popular myth: People often perceive their ads will perform better if they are above the fold. This is not necessarily true; it depends on what the site visitors are doing. Scrolling down is actually a sign of an engaged audience, so test basic assumptions during this redesign phase. Above-the-fold metrics will quickly cede to currently "in view" and we'll all be better for it. An ad that's never seen is not worth much.