Is 100% viewability something to shoot for?

In 2016, GroupM proudly sent out a press release announcing they would only buy 100% viewable ads. Cool, cool. But ...

Did they NOT know that fraudsters were already EASILY falsifying the viewability measurements so that 100% of their ads were 100% viewable? Most instances of viewability falsification were not caught or prosecuted, but one was finally caught. March 6, 2018 Newsweek Media Group Websites Ran Malicious Code That Experts Say Is Used To Commit Ad Fraud Two executives were indicted, plead guilty, convicted, and fined. February 15, 2020 A Former Owner of Newsweek Pleads Guilty in a Fraud Scheme. There was no question they were doing the fraud by falsifying the measurements.

By committing to only buying 100% viewable ads, the agency group increased their risk of buying ad inventory that falsified viewability measurements. It also effectively rendered large swaths of good publishers' inventory unsellable in an instant. If you read the fine print of GroupM's press release, it says "starting from Jan. 1 this year, [ads purchased] must include a guarantee that digital ads must be on the screen and within view (not requiring the user to scroll down) for at least one second in order to count as “viewable.” This means the ads must START as viewable.

How does this harm real publishers? Take one real publisher for instance. They have a normal webpage with 3 ads above the fold and 1 ad at the bottom due to page layout. That 1 ad at the bottom is not initially "within view (not requiring the user to scroll down)." That means 1/4 of their inventory is no longer sellable according to GroupM's buying mandate. 25% of real publishers' inventory vaporized in an instant because of that proclamation. Good publishers don't change their page layouts that often, so all of a sudden they are left with only 75% sellable ad inventory, for example to sell to large advertisers whose digital budgets were controlled by GroupM. The 100% viewable proclamation didn't affect fraudulent sites' revenues because those sites were already falsifying their ads' viewability so all of their ads were sellable to GroupM.

For context, see how FouAnalytics reports viewability in the 6 charts above. The yellow line overlaid over the green volume bars tells you the hourly average viewability of your ads, along with an overall average for the time period, on the right side. All of the properly measured viewability numbers I see are lower than the viewability rates reported by legacy vendors. But note, they only measured 2% of the ads, but reported 84% viewable rate. More details here: https://www.dhirubhai.net/pulse/revisiting-viewability-dont-trust-always-verify-dr-augustine-fou

Also see the "viewability in practice" section in the following article -- https://www.dhirubhai.net/pulse/fouanalytics-pro-tip-viewability-measurements-the-independent/ -- to see how FouAnalytics measures viewability and shows you the supporting data on what was measured and how viewability was calculated in FouAnalytics.

For more context, since many advertisers are still inadvertently buying MFA sites. Those sites also exhibit very high viewability because they load ads all above the fold (think slideshow page with 6 - 10 ads surrounding the slide show on the same page). And crappy mobile apps that load ads in the background or in 1x1 or 0x0 pixel windows have 0% viewability. Be sure to check for these anomalies in your FouAnalytics in-ad dashboards.

So What?

Smart advertisers have moved away from optimizing towards higher viewability. Absolute viewability numbers are taken for what they are. Good publishers have 75% sitewide average viewability (e.g. 3/4 of their ads start above the fold and 1/4 of their ads start below the fold), and that is not a good or bad thing. It just is what it is. If, however, you have exceptionally low viewability, like 50% or less, according to FouAnalytics measurements, you should check on why this is. From what I have seen, very low viewability comes from types of fraud like pixel-stuffing (ads shown in 0x0 or 1x1 sized windows), ad stacking (more than 1 ad stacked underneath each other in the same ad slot), and popunders (ads shown in pages behind the browser), and often related to fraudulent mobile apps that load ads in the background when users are not using the app or the mobile device. Very low viewability rates will surface these issues so you can correct them (by blocking the sites and apps using those other forms of fraud).

Further reading: https://www.dhirubhai.net/in/augustinefou/recent-activity/newsletter/