For the latest authoritative marketing news, trends, and stats on social media stars and influencer marketing, subscribe to our industry digest newsletter!
With video advertisements from over 4 million marketers now constituting a large percentage of Facebook’s 100 million hours of videos viewed per day (Re/code), the potential reach and impact of Facebook video ads are undeniable. However, Facebook’s recent admission that it fudged one of its key engagement metrics (“average duration of videos viewed”) has left brands, marketing professionals, and notable publications all scrambling to make sense of what Facebook’s error means and, most importantly, how it will affect investment in Facebook video ads moving forward.
While Facebook’s confession that it had made a two-year-long mistake when calculating the average duration of video views was initially ill-received, it was the Wall Street Journal’s report that video view times may have been misrepresented by up to 80% which sent shockwaves through the marketing world.
By not counting any video that wasn’t seen for at least 3 seconds (and therefore didn’t qualify as a “view”) when calculating average video duration, Facebook shared data that suggested much higher engagement. When counting total views (including those for 1 & 2 seconds), the average view time on each video differs from the metric Facebook had previously presented for the last 2 years for up to an 80% miscalculation.
Essentially, Facebook calculated its “incorrect” average video view time by reducing the denominator of their equation (which should have counted every video view instead of only those 3 seconds or more) while keeping the numerator (the total time viewed) the same, an error that, when corrected, may drag down the overall average by 60% or more.
In response to the outpouring of negative press calling Facebook’s ad practices “deceptive,” among other things, Forbes actually came to Facebook’s defense. In an article entitled “The Media Got It Wrong: What Facebook’s Video Ads Issue Tells Us About Big Data Metrics,” author Kalev Leetaru claims that Facebook has always been transparent about how it defined “video views,” stating:
“If one simply looks back to Facebook’s original announcement of the video metrics dashboard, one will find that they prominently and clearly announced this definition [i.e. 3 seconds or more] from the very first release.”
The error, Leetaru says, is that Facebook failed to remind advertisers what counts as a “video view,” or possibly that Facebook’s metric for measuring views is not inconsistent with other video platforms—YouTube, for example, measures a “view” at 30 seconds of watch time. In either case, Leetaru argues that the advertising and media community, not Facebook, that is at fault for the recent uproar.
Related Post: How Facebook’s Overestimation Could Affect Advertisers
As one can see from the simple equation below, Facebook’s correct calculation should have divided the “total time spent watching a Facebook video” by the total “# of people who watched the video (or total # of views),” regardless of how long they watched each video for. Instead, Facebook divided the “total time spent watching a Facebook video” by only the “# of ‘views’ that lasted for three seconds or more,” a figure that would likely be much lower (especially given Facebook’s video autoplay feature) and therefore would inflate the average viewing time of each video.
For example, 600 of the 1,000 video views were less than 3 seconds and were therefore not calculated in the “average duration of video viewed” metric, the difference in the results—15 seconds vs. 6 seconds—represents a 60% discrepancy between what advertisers believed Facebook was measuring as opposed to what Facebook was actually measuring.
While Facebook’s error may or may not be accepted and furthers the movement for standardized measurement across all social media platforms, the important takeaway rests on the misrepresentation or misinterpretation of engagement on Facebook’s videos. Advertisers base their campaigns on engagement metrics such as average video view time and as Facebook’s error states, these times were off and significantly lower suggesting much less engagement. With much lower video viewing times, some advertisers may look to other more engaged platforms.