People Aren't Watching Facebook Videos as Much as Facebook Said

The platform massively overstated how many minutes users watched each video.

Facebook founder Mark Zuckerberg at a meeting with Jack Ma, the CEO of Alibaba, in Beijing in May. (Reuters)

Over the past year, both media companies and advertisers have invested an enormous amount of money, labor, and time in Facebook’s video products.

You can see evidence throughout the industry. It’s not just the auto-playing Donald Trump campaign ads that clog your News Feed (or, at least, which clog mine). In August, BuzzFeed restructured its organization so that video oversees most non-news divisions. In April, Mashable laid off about 24 writers and editors so it could make a “strategic shift” to video. Media analysts (including myself) began comparing where audiences spend their time to where advertisers spent their dollars. We suggested that soon mobile video could erode the entire television ad market.

As of this morning, mobile still reigns. American consumers spend about 25 percent of their time on mobile phones (and a huge amount of that time on Facebook), but advertisers only spend about 12 percent of their budgets there.

But any confidence that Facebook will soon devour the entire TV market was a little shaken Thursday by news that Facebook has been (accidentally, it says) juicing its metrics. The Wall Street Journal is reporting that Facebook misstated a key video metric—the average amount of time people spent watching a video—by 60 to 80 percent over the last two years.

Large ad firms like Publicis, which spent $7 billion on Facebook ads last year, are annoyed. Publicis is demanding that Facebook submit to third-party verification of its ad metrics, according to the Journal. And while that change would be unusual for platform-based video ads, it is standard in other sectors: Nielsen Media Research estimates every television show’s audience, for instance, and not a network like NBC or CBS.

Facebook has since apologized for its overestimation. It’s also clarified maybe the most important variable in this conversation: that the miscalculation did not affect billing. (Of course, some advertisers may have decided to invest in Facebook video because of the strength of this one particular metric.)

“We have also reviewed our other video metrics on the dashboard and have found that this has no impact on [other] video numbers we have shared in the past, such as the number of video views,"said a statement from David Fischer, Facebook’s vice president of sales. “We want our clients to know that this miscalculation has not and will not going forward have an impact on billing or how media mix models value their Facebook video investments.”

The roots of this miscalculation go back to one of the founding assumptions of Facebook video. The platform counts every instance of a video playing for longer than three seconds as a “view,” saying that that is long enough to count as “intent to watch.” Professional YouTube video makers like Hank Green loathe this approach. Last year, Green accused Facebook of lying to juice video metrics, arguing that a “view” should only be registered when someone watches most or all of a video (which is how YouTube counts them). He also said it cheated by boosting native Facebook videos into people’s News Feeds well above embedded YouTube videos or Vines.

This views metric turned out to be Facebook’s undoing. Prior to last month, the company calculated “average time viewed” using its views metric: It divided (total number of minutes played by everyone) by (total number of views). As such, it artificially got credit for the couple seconds of video playing before the three-second cutoff in the numerator but it never accounted for those people in the denominator. It has fixed this in the new metric by dividing (total minutes) by (total number of plays).

Suffice it to say that this is all very embarrassing for Facebook, and somewhat disconcerting for the organizations that have retooled themselves to better produce video for the platform. It doesn't mean that it's futile to invest in Facebook video, which is still an enormous platform in terms of getting people's attention. But it does point to how perilously concentrated the shortform online video market is. Major companies have made longterm plans to cater to an organization that’s both selling a product and measuring its effects. Facebook creates the market for Facebook video, then it sells you access to that market. It encompasses so many people—and so many people who use it habitually—that marketers can’t afford to care.

Over the past two years, Facebook has contorted its entire user experience around the demands of video. Log on and it will show you live videos, pre-filmed video with rolling captions, and auto-generated videos constituted entirely of old pictures of your friends. It says that its users love video, even if they can’t scroll through videos to preview them like they would images or an article. It even pays media companies to produce Facebook Live videos and then it boosts them in its feed.

Today’s news doesn’t doom Facebook’s push to video. But it does make me wonder if the company has been sloppy with its thinking about video elsewhere. The downside of measuring your own monopoly is that you can miss negative feedback until it’s too late. When it comes to video ads, Facebook is its own designer, judge, analyst—and executioner.