Facebook’s False Video Metrics – Implications

Facebook video exggerations


Facebook stock dropped 1.63% Friday, following news that for more than two years the company had vastly overstated a key video ad metric.

In this post, I’ll discuss what happened, and the potential implications (i.e., why digital marketers should care).

The Problem

Facebook’s ad metric for “average duration of video viewed” failed to count video views of less than 3 seconds.

This is problematic for two reasons:

  1. Facebook videos are set to auto-play by default.

This means that casual viewers who are scrolling through their feed likely encounter videos they have no interest in watching, but may catch glimpses of for less than three seconds.

It’s essentially the difference between reporting “clicks” on a banner ad vs. “impressions.”

Here, the views of less than 3 seconds were impressions – not that they necessarily should have been counted as true “views,” but to not count them in the average duration is incredibly misleading, because:

2. Averages are particularly susceptible to extremes. So they are easily skewed by repeated low values.

If 90 percent of viewers of a 30 second video actually viewed it for just 2 seconds – and those 90% aren’t counted in the numbers Facebook reports, advertisers are definitionally over-paying for the distribution they think they are getting.

According to The Wall Street Journal, which broke this story on Friday, advertising giant Publicis Media was told by Facebook that the earlier metric likely overestimated the average time spent watching videos by up to 80 percent.

Video Wars

The revelation of Facebook’s grossly exaggerated reporting comes at a particularly bad time, as the platform is emphasizing video as it seeks to take market share away from Google’s YouTube.

In particular, Facebook has been banking on live video as its next revenue driver.

At Digital Summit DC last week, SEO experts reported that Facebook Live videos currently get 40 times the number of views on average as ordinary Facebook videos, since that’s the area that the social network believes it can best differentiate itself from YouTube. It is boosting those posts more than any others.

In fact, ABC News made history last week when it announced it will live-stream tonight’s first Presidential Debate between Donald Trump and Hillary Clinton via Facebook.

For Facebook to become a credible contender to YouTube (and even Twitter’s Periscope) in the video – let alone live video – space, it must be accurate in its metrics calculation and reporting.

Advertisers Respond

As The New York Times noted on Friday,

“Average duration of video viewed” is neither a public metric nor does it determine how much advertisers pay for video placement. But the figure is displayed prominently in partners’ dashboards, and it as helped assuage discomfort about the platform’s generous view-counting methods.”

Regardless, trust in Facebook’s metrics is paramount as all of those numbers are used by advertising and public relations professionals who need to justify their spending on Facebook ads.

Predictably, however, digital marketers and the larger advertising conglomerates (including Publicis Group and WWP) have said the mistake will have little effect on their Facebook ad expenditures.

The likely effects are two-fold:

  1. A greater push for 3rd-party measurement and access to Facebook’s metrics; and
  2. A potential shift toward greater ad spending on competing platforms like YouTube, Twitter, and Instagram.

My View

Many articles have come out over the weekend quoting representatives of boutique digital marketing agencies that downplayed the news.

This makes sense.

As a PR professional, it’s my job to try to sell clients on the need to not only have a Facebook page, but also to spend money to have us craft and distribute targeted Facebook ads for them.

And to a certain extent, Facebook, with its more than 1.3 billion users, is the only social marketing game in town.

Few advertisers are going to say anything negative against Facebook, as it will directly affect their ability to sell social marketing packages on the largest social media platform out there.

My personal concern is it took Facebook more than two years to discover the vast error (over-stating average view times by 80 percent?? REALLY??), or at least own up to it publicly, how do we know what other metrics of theirs we can trust?

The last thing I need right now is for an executive to read the negative news about Facebook and have one more legitimate objection to spending money there vs. a traditional print ad that will be both more expensive and less effective.

And yes, like the other digital marketing voices who were widely quoted this weekend, add my name to the list of folks who want to see 3rd-party audits of the self-reported viewing metrics of all of the major social media platforms.

Without greater transparency and trust, Facebook’s stock may see a far greater long-term decline than Friday’s mere blip.

Leave a comment

Your email address will not be published.