Facebook has overestimated how much video individuals have looked throughout the previous two years, the firm has conceded.
One promoter recommended that, now and again, video seeing measurements had been overestimated by up to 80%.
Facebook's examination are a critical instrument for promoters, who use them to work out the amount of their video substance is being viewed.
The informal organization said the blunder had been altered and had not changed how much sponsors paid the site.
"Unsatisfactory" conduct
The mistake influenced a Facebook metric called "normal term of video saw", which should tell distributers for to what extent, all things considered, individuals had viewed a video.
In any case, the metric did exclude viewers who had looked for under three seconds in the tally.
Reducing the shorter perspectives - including individuals who had disregarded a video in their news sustain - swelled the normal review times for every video.
In an announcement, Facebook said: "We as of late found a blunder in the way we compute one of our video measurements.
"This blunder has been altered, it didn't affect charging, and we have advised our accomplices both through our item dashboards and by means of offers and distributer outreach," it included.
The video-watching metric has now been renamed to all the more precisely reflect what it gauges, the organization said.
The metric is presently called "normal watch time" and Facebook began utilizing this to accumulate measurements on video utilization before the end of last month.
The announcement likewise said the video-seeing metric was one among numerous that advertisement firms use to work out if their substance is being viewed.
The Wall Street Journal cited promotion purchasing firm Publicis as saying that Facebook's distorting seemed to be "unsuitable".
Publicis said it demonstrated the requirement for outsider check of insights accumulated by Facebook.
The informal community has beforehand been censured for considering a video being "saw" following three seconds.