Editors' pick: Originally published Sept. 23.
Ad buyers and marketers are frustrated that the social media company has revealed that it's been overestimating the average viewing time for its videos for two years, The Wall Street Journalreported late Thursday. Facebook has told some ad-buying agencies were told that the average time spent had been overestimated by 60% to 80%, according to the report.
Facebook said that it first found the error about a month ago and fixed it and informed customers. The social media giantexplained that its "average duration of video" metric should have been calculated by dividing the total time spent watching a video by the total number of people who played a video. Instead, however, the metric only included people who had viewed the video for three seconds or more, exaggerating the ad performance for marketers and ad buyers.
On Friday, the company apologized in a post on Facebook.
The company stressed that the miscalculation does not affect the way it billed advertisers, and that it's introducing new metrics that will replace the current measurement method. These new metrics will be video average watch time, which refers to the time spent watching divided by the total number of plays, and video percentage watched, which measures the percentage of a video viewers watch per session.
The legacy metrics will be gone in early October while the corrected metrics have been available since late August.
While ad buyers are reportedly upset by the error, the overestimate isn't as problematic as it appears on first glance, however, said Tigress Financial Partners chief information officer Ivan Feinseth.
For one, ad pricing online varies from ad pricing on more traditional media platforms such as radio, television or print media, he explained. While traditional media avenues charge advertisers largely by the size of the audience -- circulation of a magazine, for instance -- Internet ads rely on more than just the pure size of audience because it looks at how viewers interact with videos.
"[Internet ad billing is] also based on how people watch the ad, then click on the ad," Feinseth added.
While Facebook hasn't officially quantified the extent to which ad views were overestimated, the company is quickly and proactively working to address the missteps, he noted.
Shares of Facebook are down about 2% Friday morning to $127.39. Its stock is up about 21.8% year-to-date.
For Mark Zuckerberg's social media giant, ad revenue growth and video ad momentum in particular have been among key factors driving the company's stellar stock performance and bullishness around its growth prospects.
Facebook reported a 63% year-over-year increase in ad revenue to $6.2 billion in its second quarter results in July, attributing the growth to user and monetization increases. Mobile ad revenue has emerged as a particularly bright spot, having scored an 80% annual increase last quarter.
Wells Fargo analyst Peter Stabler also wrote in a Friday note that there are more important metrics and issues regarding Facebook's video ad measurement.
"Having reviewed multiple decks of actual campaign data and spoken with contacts on the measurement issue, we haven't encountered analysis that has dwelled on 'average view time'," he noted, adding that cost-per-completed-view is a more important metric because it allows agencies and marketers to analyze video ad performance.