Follow Slashdot stories on Twitter

 



Forgot your password?
typodupeerror
×
Facebook Social Networks IT

Facebook Inflated Video Viewing Stats For Two Years (cnet.com) 49

Facebook has admitted inflating the average time people spend watching videos for two years by failing to count people who watched for less than three seconds. CNET reports: The metric was artificially inflated because it only counted videos as viewed if they had been seen for three or more seconds, not taking into account shorter views, the company revealed several weeks ago in a post on its advertiser help center web page. Facebook has been putting a greater emphasis on video in recent years, particularly live video. In March, Facebook began giving anyone with a phone and internet connection an easy way to broadcast live video to the 1.7 billion people who use its service every day.
This discussion has been archived. No new comments can be posted.

Facebook Inflated Video Viewing Stats For Two Years

Comments Filter:
  • Comment removed (Score:3, Interesting)

    by account_deleted ( 4530225 ) on Friday September 23, 2016 @09:46AM (#52946603)
    Comment removed based on user account deletion
    • by BarbaraHudson ( 3785311 ) <barbara.jane.hud ... minus physicist> on Friday September 23, 2016 @10:38AM (#52946953) Journal

      Why would you include people who watched a video for less than 3 seconds? Would you say you've read a book because you glanced at the cover? Or seen a movie because you saw a 30-second trailer (okay, that last one, too often the trailers contain all the interesting stuff, so you're wasting your time watching the movie, but you get what I mean(.

      As long as you don't include those people in the "total viewers" category, I see no problem.

      • Why would you include people who watched a video for less than 3 seconds? Would you say you've read a book because you glanced at the cover? Or seen a movie because you saw a 30-second trailer (okay, that last one, too often the trailers contain all the interesting stuff, so you're wasting your time watching the movie, but you get what I mean(.

        As long as you don't include those people in the "total viewers" category, I see no problem.

        It's Facebook being shady. They added the autoplay feature so that they could drive this metric and then excluded the people that intentionally skipped these videos they didn't want to autoplay to begin with.

      • by rhazz ( 2853871 )
        +1. The article is light on details but the fact that facebook just apologized rather than trying to explain it does suggest perhaps that the 3-second rule wasn't applied across the board. Given that facebook autoplays videos by default, as an advertiser I would definitely want to filter out people who scroll past the video, or at least categorize them separately.
      • Why would you include people who watched a video for less than 3 seconds? Would you say you've read a book because you glanced at the cover? Or seen a movie because you saw a 30-second trailer (okay, that last one, too often the trailers contain all the interesting stuff, so you're wasting your time watching the movie, but you get what I mean(.

        As long as you don't include those people in the "total viewers" category, I see no problem.

        Agreed and observed first-party. I don't have a Facebook account, so it doesn't matter in terms directly, but I will be given search terms for a video that I'm told to "look at the third related video under it". Go to YouTube, search short name to get to the video, click on the third related one below. It may have showed the video for more than 3 seconds, or less than 3. I wouldn't consider that "watching". It's "piggyback finding another video".

        I also (from looking over others' shoulders while they're

    • It's as bad as e-mail at this point.

      Huh?? How does a distributed and independent service like email relate to a top-down governed monolithic social network at all? "Email lied to me"...you mean people?

  • by Anonymous Coward

    It's only dishonest if they exclude sessions of less than 3 sessions from the "average duration" statistic but include them in the "number of views" statistic.

    If they exclude them from both, it's perfectly reasonable. Still good to disclose the methodology, of course.

  • by pahles ( 701275 ) on Friday September 23, 2016 @09:53AM (#52946675)
    They only measure the time the video is played, not how long it is viewed.
  • I'm not really surprised, since the videos are not monetized. Youtube has to pay content creators so if they inflate viewership, it costs them money. On Facebook, it allows them to show striking numbers, gain publicity, and try and leverage their platform.
  • by Opportunist ( 166417 ) on Friday September 23, 2016 @09:56AM (#52946699)

    Imagine they would count people "watching" videos for less than 3 seconds (read: people who click something, notice it's a video, go "fuck this shit, I ain't watching a video now!" and close it). Would that cause an uproar? You bet it would. "Bah, cheating, people aren't really watching that, it's just clickbait and they get lured there, people aren't really interested in the video, FB is only trying to say so to be relevant, people go to YouTube for videos..." and so on.

    I'd be the last defender of FB (as far as I am concerned, the day they finally croak should be called "privacy day"), but what exactly should they have done?

    • I think it would be interesting to know just how many out of all initiated video views are cancelled within 3 seconds. How many people are saying "A video? I'm not watching this!" is very relevant I would say.

      • by ndogg ( 158021 )

        I don't think it would be that interesting. Someone canceling the viewing of a video in less than three seconds seems like someone who accidentally clicked on something they didn't mean to.

        • I don't think it would be that interesting. Someone canceling the viewing of a video in less than three seconds seems like someone who accidentally clicked on something they didn't mean to.

          Another possible explanation is that Facebook once again turned on the option to automatically play videos in user's settings without asking them and the video starts playing as soon as they scroll it onto the screen.

    • Imagine they would count people "watching" videos for less than 3 seconds (read: people who click something, notice it's a video, go "fuck this shit, I ain't watching a video now!" and close it). Would that cause an uproar? You bet it would. "Bah, cheating, people aren't really watching that, it's just clickbait and they get lured there, people aren't really interested in the video, FB is only trying to say so to be relevant, people go to YouTube for videos..." and so on.

      I'd be the last defender of FB (as far as I am concerned, the day they finally croak should be called "privacy day"), but what exactly should they have done?

      The shady thing is that most of these ads get played automatically by Facebook. People didn't intentionally watch them to begin with. So what happens is that they count the people who may not be watching the ad anyway but don't count the people that intentionally skip the ad. If Facebook wants to be honest, they should only count ad plays that people intentionally watch - which my guess will be approximately 0. They're lying to their customers and telling them their ads are more successful than they act

  • by JoeyRox ( 2711699 ) on Friday September 23, 2016 @10:06AM (#52946753)
    Taken at face value the CNET summary would imply that they were actually correctly stating views since their methodology was excluding shorter views. The full explanation from Facebook makes it clearer:

    We had previously *defined* the Average Duration of Video Viewed as "total time spent watching a video divided by the total number of people who have played the video." But we erroneously had *calculated* the Average Duration of Video Viewed as "the total time spent watching a video divided by *only* the number of people who have viewed a video for three or more seconds."
    • depending on how you are browsing you can scroll through a page on FB and have a bunch of videos auto play this would seriously inflate the average viewing duration

    • Why are they using the mean? Isn't something like this precisely when you're supposed to use the median? Ok, that means they have to store a lot more data (time spent per each individual view, instead of just aggregate time viewed and aggregate number of viewers). But presumably they're already keeping track of every video every FB member has viewed, so this wouldn't be that much more data.
  • [not counting

    Gotta be sure to count autoplaying videos for number of videos viewed and time viewing videos, so long as you don't count the people who are quick enough killing the stupid annoying video that it would negatively impact the average view time.

  • Let me inflate the time I give a shit about fakebook...
  • Should that be counted?

    Some probably clicked accidentally, and it should not be.

    Some probably clicked, went "Ugh!" and closed it. Probably should be.

    On balance, it probably should not be. Non-scandal.

  • Huh.

    I just ignore them. And if I hear a company name, I put it in my Boycott list.

  • I'm sure someone from the public cares about this..
  • No, really, who is in Facebook? I created a Facebook account a few years ago, so I could use as a convenient way to access sites that require you register. I never ever log into my Facebook account directly; in fact, I have no clue about what it contains, if anything. The kicker is, most people in my professional and social circles seem to be using their Facebook accounts in a similar fashion. More damning, their kids seem to feel embarrassed about the possibility of having to use Facebook. Leaving aside th
  • It was WSJ that broke the story this morning, CNet is summarizing it.

    The original article is here.

    http://www.wsj.com/articles/fa... [wsj.com]

Keep up the good work! But please don't ask me to help.

Working...