Instead of calculating average time spent by dividing total view time by total impressions, the metric reported by Facebook was actually total time divided by impressions that delivered at least a 3 second view.
So the reported average view-time only showed results for users who were already engaged, making average engagement look far better than it actually was on a per impression basis.
Is this the end of the world? Will advertisers flee from Facebook? Time to unload your FB stock? Hardly.
Will this likely be forgotten in a week? Probably.
As Facebook has pointed out, it’s extremely important to note that no client was overbilled as a result of the error, and now that the error has been discovered, Facebook is being totally transparent about the issue.
The holding companies, unlikely to start amplifying an issue that might generate client questions about budget stewardship, have been generally muted in their public response to the issue.
But it does raise some critical questions.
Like how did this go on for two years without advertisers, agencies or Facebook questioning the inflated results or auditing their own processes? I genuinely don’t think that Facebook juiced their metrics; mistakes happen, and you’ve got to believe that they’ve got bigger fish to fry.
But it’s undeniable that in a two-year period when Facebook has made significant inroads in video, much of their video business would have been built by showing competition-beating results. Media-buying decisions are often driven exclusively by results.
In a business with the scale of Facebook, one has to wonder: How many millions were diverted to Facebook at the expense of other platforms based on these flawed results? Of course, if you wonder about these things, wonder away, because you’ll never know.
So it’s tempting to conclude that the obvious losers here are the other platforms.
But it’s not just competing media suppliers that lost out here – it’s the advertiser.
Whether self-or agency-directed, more dollars flowed to Facebook than to other platforms which could have delivered superior, more accurately measured performance to brands.
And on a wider level, the whole industry loses when The Wall Street Journal covers the Facebook error with the headline “Doubts Rise About Digital Ads”
So while Facebook did not overbill anyone, it’s hard to imagine that working advertising dollars were not lost through the error.
Another point to consider is that while this particular issue doesn’t seem to be generating much more than a raised eyebrow, what would have happened if the error was actually in how a billable metric was calculated?
If the mistake had actually been around the number of billable video impressions, for two years, FB could be looking at a significant liability.
Which brings me to my final point and, full disclosure, I work for a third-party ad server:
This issue (and my hypothetical billable impression issue) may have been discovered and brought to Facebook’s attention within days, not years if Facebook allowed third-party ads serving or even tracking of their video.
(I know that Facebook allows limited third-party verification through Moat and Nielsen. But they’re not measuring the metric that was the source of this error.)
Which begs the question:
Is there anyone outside of Facebook who thinks it’s a bad idea to allow advertisers to use a third-party to verify the measurements that determine how millions of advertiser dollars are spent?