Hacker News new | past | comments | ask | show | jobs | submit login
85 percent of Facebook video is watched without sound (digiday.com)
274 points by prostoalex on May 31, 2016 | hide | past | favorite | 133 comments

Videos on facebook auto-play in silence as you scroll through your feed. It's only once you click on them that sound is enabled.

Given the 85% figure - I would not be surprised if the "watched" metrics simply means it loads the video and begins playing it.

That means potentially only 15% of all facebook videos are actually engaged on.

We need more data on how long someone needs to view the video to consider it "watched." The way I understand it [1], is that facebook counts any video as watched so long as it shows - much like adwords impressions. Only if it goes equal to or past 10 seconds is it counted differently.

I read this as indicating that facebook video ads aren't particularly engaging.

[1] https://www.facebook.com/business/help/1582420952009573

"I read this as indicating that facebook video ads aren't particularly engaging."

I wonder if most of everything isn't particularly engaging for many people. Think about how much you skim most articles, comments, photo series, how you use your phone while watching TV, etc. Watch someone use Instagram and often they'll look at a photo for a fraction of a second before moving on or giving out a double-tap.

"how you use your phone while watching TV"

I find myself doing this all the time, even when the show is engaging. I'll automatically open Reddit and start scrolling through the front page, and before I know it I missed 5 minutes of the show. I do it on autopilot, and it annoys and scares me.

Oh yes. I do that a lot with HackerNews, actually. The show goes to one screen, HN goes to the other. I usually start skimming the site when the show gets "slower", only to realize few minutes later that I missed an important part of the action. An annoying habit I try to fight off.

I think it stems from the need to have the brain occupied to some level, and the show reaching close, but not to, that level. I feel the need to add something to cross the boundary. Eating while watching helps (it seems to require just enough cognitive effort to reach the comfort level), but it has its own bad side effects (like scheduling meals to the shows).

Why I quit reddit and uninstalled all the quick games on my phone. After a while I lost the compulsion but it's an interesting experiment to show yourself just how often you're opening your phone for drivel.

A while back I instituted a personal rule that I am not allowed to use my phone or laptop while watching TV or a movie. My rationale is that if I am uninterested enough in the show to need additional distraction, the show is not a good use of my time and I should just turn it off. It was really hard to get out of the habit, but it's worked pretty well for me. I find that I understand and enjoy the shows I'm watching more, and I also watch less TV because I more consciously notice when I'm getting bored, and I go do something else.

You might find this podcast[0] (and his[1] other writings on the subject very interesting.

Cal Newport discusses "Deep Work" and how deep, concentrated, uninterrupted focus is 1) becoming more rare these days with so many feed, notifications, devices, etc. and 2) absolutely essential to being productive and successful in almost any field in today's society / economy.

It's more interesting than I'm describing. Cal's insights into learning, research, lifestyle, studying were very influential on my study habits and mindset during college. Check him out.

[0] https://www.33voices.com/interviews/deep-work

[1] http://www.amazon.com/Cal-Newport/e/B001IGNR0U

These days, I only watch TV (usually through Amazon Prime) on my iPad while using the treadmill or stationary bike. That consumes enough of my attention—and makes it enough of a pain to use my hands—that it prevents the urge to do a third activity while watching. (Plus it's healthy!)

I will do it while watching a movie, except I'll often pause the movie to check a couple of sites (that I've already checked within the last 30 minutes) and before long it's taken me 6 hours across 2-3 nights to watch a 2 hour movie.

It's pretty stupid behaviour. Procrastinating even while relaxing.

I realised this is why I've been enjoying shows with subtitles recently (some anime titles on netflix) because it forces you to watch the show without the constant phone checking.

I read it different. From personal experience alone I watch a lot of videos on FB fully without ever clicking to listen the audio. And several content producers are adapting to this by creating video where audio is optional. With subtitles or sometimes it is nothing more than a "animated slide presentation". But also videos like highlights of NBA games.

I watch those food prep videos without sound every time.

I do that a lot too, often I don't care to listen to the sound or am in a place where it would be inconvenient to listen.

I'd like to point out that this is but a single data point.

You can say that any time anyone says anything (and on HN someone usually does), but if you think about it for more than half a second, actually it is more than a single data point if multiple content producers are optimizing video for silent viewing.

Content creators usually tweak their content to increase their metrics, whether those metrics translate in more engagement is different.

> You can say that any time anyone says anything (and on HN someone usually does)

As well they should, personal anecdotes rarely contribute much to the conversation.

It's a conversation-demolishing verbal tic, along with incantations about causation and correlation, straw people, ad homunculi and the like. Avoiding these helps conversations stay conversations rather than tedious protocol.

Some of us find conversations that devolve into these things to be incredibly frustrating to be part of. People love to share anecdotes about themselves but rarely care much for others peoples' anecdotes.

Before sharing that you prefer pepsi over coke you should ask yourself whether you care at all that some stranger on the internet prefers coke over pepsi. Assuming the answer is 'no', then why bother sharing these sorts of things yourself.

Facebook considers videos viewed after 3 seconds, and YouTube after 30. Instagram after 3, but it counts unique users and not views. Twitter is when someone clicks, and Vine is when someone finishes one loop of the video.


I'm one of the 85% – I typically watch vids in silence when I'm commuting, especially if I'm already listening to music

Same, but on Twitter. The political interview tweets I see often add subtitles because they understand this phenomenon.

I think this is significant. If facebook starts interrupting my music I would use it far less.

5 seconds is their metric for a view

in their ads interface it breaks it down by completion percentage.

I thought it was three seconds, did they change it?

My mistake you're right it's 3 seconds


HN won't let me correct my original post because I didn't notice this until now



And if you have set auto load on, they basically pre load the first 3 seconds and that counts as a view if I'm not mistaken.


Seems like 5 seconds is a good threshold to determine if someone is going to stay or move along - I assume they rigorously got to this number.

I suppose the next question would be can you correlate the enabling of sound to that 5 second metric, and if so, what is the difference in engagement?

I only wonder further because dialing in ad engagement, especially as it relates to seconds of interactions, is a really interesting way to determine value of these ad plays.

5 seconds seems to short unless paired with page/mouse position, I regularly notice that videos will start playing when they are 2/3rds on the page and I'm read or looking at a picture above it only to scroll right past the video as soon as I register it's of no interest to me.

Advertisers already know this which is why more and more FB videos include subtitles and (the smarter advertisers) will communicate whatever message they're trying to land in the first few seconds of the video.

So even if not 'engaged' with the video still does its job.

Yup – and, as a result, at least one ad tech company has emerged to meet advertisers' need for video caption copywriting, design, and automated testing. And they have customers.

They address the engagement issue in the article, but not in a particularly convincing way : "Internal studies conducted by the agency showed that KPIs like brand lift and intent to purchase were not affected by whether the viewer watched the video with the sound on or off" I wonder if they are actually measuring engagement over a complete feed stream, or just some subset of videos that were allowed to play longer than 10s (guaranteeing a baseline level of engagement).

I view it as much of facebook usage is at work or other public places where people don't want sound. :-)

Most people have headsets.

I think headsets at work is a tech thing. Most of the world doesn't get to wear them all the time.

I almost exclusively use FB on my phone. Anecdotally, I see similar behavior from most of my friends. Yes, many people have headphones for their phones. But my guess is that they are not plugged in most of the time people use their phone, and they're unlikely to plug them in for a brief video.

Most people? Hardly. Of the minority of people that have headsets, a minority thereof are willing to pull them out of their bag to listen to some advertisements.

Also video player is one huge click-machine. Every major player allows you to pause by clicking on video player window, Facebook instead opens player in new "window" with suggested content. I bet this accounts for a lot plays as well. Same way Google was inflating Bing visits by adding dedicated, extremely sensitive search button to Windows mobiles.

The article states 3 seconds is a view. If I'm on desktop and stop to read a post above or below an autoplay video then it would likely count as a view. I'm sure some people do watch video without sound but I agree the 85% figure is likely misleading.

Often times when I'm on Facebook I'm killing time somewhere it is inappropriate to have audio playing...

So, they can be just gif, then.

You can turn off auto-play of videos here:


Edit: Apparently some people don't see that setting. These are the settings I see: http://i.imgur.com/XexAp8k.png

(Aside: Anyone recommend a better image host than imgur?)

For me, visiting on desktop only allows me to choose between "SD only" and "HD when available". There is no option for "disable autoplay".

Congrats, you're in special Facebook mind control "feature" testing group.

Not sure why you're being downvoted. The option to turn Facebook sounds off on desktop was not there for me when they rolled it out (it was for others). I stopped keeping a Facebook tab open.

> Not sure why you're being downvoted.

Is it really about "mind control"?

(I wouldn't downvote it (and can't anyway), but that might be why others did.)

I think it is. I would bet money that Facebook employs or consults at least one psychologist to help them design their app for maximum addiction. An example from a few weeks ago is how "trending" stories are handpicked by humans.

newspapers also handpick trending stories. Sometimes they even write articles!

Is A/B testing really so insidious?

If the purpose of the test is to subtly herd your users towards a result which is good for business but bad for the users, then yes.

If it confuses users who don't understand why it's different for their friends, it can be perceived as such.

It's unclear how much A/B testing has to do with this. If I use a mobile user agent (e.g. Nexus 6 via the Chrome Dev Tools) I get the option to enable or disable autoplay.

I don't see it either now, but I guess I'm lucky because I turned it off as soon as Facebook introduced this "feature", so now I really don't have to think about it.

The default settings for mobile was so weird. Auto play videos, and "non-HD" photo and video uploads. I reversed them as I prefer to conserve bandwidth used for my stuff by not playing random clips...

"Non-HD" photos and videos on Facebook are compressed to hell and back... Must be JPEG at something like quality 60%. There's been recommendations to just upload at a high resolution, which works fairly well when viewed on HiDPI displays as the resolution hides the artifacts, but still terrible otherwise. IIRC they had a trial with WebP but abandoned it? It's too bad we don't have good low bandwidth options there, because of course I can see their side of it too. Should honestly have been a problem solved eons ago by now.

What settings do you see on that tab?

For me, I can only change the default quality, there's nothing to configure autoplay behaviour.

Screenshot of what I can see: http://i.imgur.com/0F2Cjyv.png.

What do you mean by "better" ? I'm thinking of creating an image host.

One that gives me the direct image link right away, specially easy to use on mobile.

I didn't even think to look for this setting on LinkedIn -- it exists there too. Click your avatar > Privacy & Settings and it is on the Basics page a little ways down.

I think imgur remains the best no-nonsense plain image host, although maybe I'm still recovering from years of photobucket.

If you use github or gitlab sites for a personal website, you could probably use your own repo as a miscellaneous image host-- as long as it's within the bounds of their TOS.

> I think imgur remains the best no-nonsense plain image host

Not anymore. They started clean and bloat free and now have a ton of junk on their desktop pages, push you to the mobile app with banners and overlays on their mobile pages, and forcibly redirect (esp mobile) direct-links to the full-page HTML versions [1].

This behavior is believed to be one of the reasons Reddit (which is basically the reason Imgur exists) is now switching to self-hosting images.

[1]: https://www.reddit.com/r/changelog/comments/4kuk2j/reddit_ch...

I only see the option to toggle the quality.

I get these options on the desktop web version of Facebook:

Video Default Quality


You can still change the quality of a video you are watching by clicking the HD icon in the video player.

Auto-Play Videos


These settings only apply when you use the Facebook website. [Follow this guide](https://www.facebook.com/help/633446180035470) to change auto-play videos in your Facebook app.

Thanks very much, didn't notice that.

> 85 percent of Facebook video is watched without sound

Probably for the better because I suspect a large percentage of watchings is happening in the office without the boss knowing.

Sadly, we are living in an era where everybody has a television set on their desk, and is watching the equivalent of America's funniest homevideos all day. Apparently without sound.

As someone who mainly browse on the phone, I have gotten used to browsing with sound off.

The issue is not that I am hiding from my boss (I have a computer and headphones at work).

The issue is that I don't want everyone in the same room as I am to hear my youtube videos.

I don't need sound on a cat video. I'm quite content watching the kitty in silence. Except for that 1 minute video where the cat meows to "happy birthday". Or the budgie-and-cat video, because the bird sounds cool and it helps to see how patient (or lazy) that cat is: https://www.youtube.com/watch?v=QmfnnKSMiVk

I wholeheartedly agree. Facebook video is an offer, you look at it or you don't, perfectly content in the knowledge that you won't miss something important. Audio would remove that choice.

For me, the headline reads "15% of Facebook video is watched with sound". An amazing number, what are those 15%?

This is quite frustrating to read. You cannot write about this stuff without carefully defining the terms.

What does a video being watched mean? It could be 10s, 20s, scrolled to the end, ...

What does sound enabled mean? It could mean sound clicked, but right at the end of the video, it could mean 10s with sounds, 20s, etc...

Take articles that do not carefully define engagement metrics with a pinch of salt.

While autoplay has a major hand in this statistic, users know they can get sound for the video if they want, and it looks like users are willingly watching entire videos without sound (autoplay ends if you continue scrolling). I think it's analogous to the rise in popularity of gif(v)s.

The moving pictures in the Harry Potter world's newspapers seem oddly prophetic now.

My feed is comprised of a lot of snowboard videos. I often watch these videos without sound because it's not necessary to hear the music.

I'm watching those videos in public most of the time - the last thing I want is to be that asshole who plays the sound out loud in public. I could bring headphones with me some of those times, but I don't really find a compelling reason to listen to anything on my cellphone.

No kidding, it autoplays every fucking video regardless of if you're on wifi or not.

Paper has an option to disable autoplay. I'd be surprised if the main app didn't.

It does.

It can be disabled and you should use is. Saves a ton of data and speeds up for device.

Unless Facebook is actively trying to slant the number of video views, auto-play should just be disabled. It really doesn't even need to be an option to turn it back on, because I doubt that many would want to.

It should be disabled by default. Absolutely the most obnoxious thing about FB.

..but that dark pattern helps achieve magnificent number OP mention :)

Now that ever more sites are transitioning to HTML5 videos, we need easy-to-use and ubiquitous browser options or addons to disable the autoplay of such videos as well as GIFs. Not everyone everywhere is on cheap broadband or fiber. Am rather surprised none of the browser vendors nor any addon maker has implemented such a functionality yet, although I'm most likely overlooking something here.

On Firefox set media.autoplay.enabled to false under about:config. As of version 41 this blocks play() unless you've interacted with the video. https://bugzilla.mozilla.org/show_bug.cgi?id=659285

You can also set plugins.click_to_play to true to do the same for Flash.

I always keep the sound off, mainly because I read facebook in bed next to my significant other when I can't sleep. I don't want to wake her up blaring some cat video.

Content creators have known this for a long time, and it's become a self-fulfilling prophecy/positive feedback loop.

Most videos I see on Facebook now have subtitles, because most people watch them without sound. And now because most videos have subtitles, I don't have to click on a video that catches my interest.

Considering the awful compressed audio that goes with ads, it's a good thing.

I could stand ads and TV if it weren't for the blasting audio.

It really is bad. I can't upload my band's music videos to Facebook video because the compression is so distorting...

When creating a new compression spec, my team always sacrificed a slight bit of the video bit rate to ensure the audio bit rate was high enough to not be annoying. For example, doubling 64k to 128k makes significant improvement to audio, but does not really help/hurt the video. Take a low bit rate video and have one version with decent audio and another with bad audio, and people will say that the video with bad audio is worse even though the video quality did not change.

I'm not talking about that compression, I'm talking about dynamic range compression.

Also worth noting that content creators have caught on to this already and are tailoring videos be without sound. There's defiantly a facebook style of video emerging that would look out of place on more traditional video hosts.

Yep, the last few movie trailers I've seen on there have been subtitled. I assumed this was the reason. If this sees us getting more and better quality subs for video content then I'm all for it.

I almost always watch YouTube videos without sound. For me putting the headphones on and fully concentrating on the sound and image of a video is too big of a "commitment" when I don't even know what the video is about. It's simply more comfortable and effortless to skim through it without sound unless I really know I want to hear what's going on. I wouldn't be surprised if a lot of people did this, specially at work. Heck, I would even speculate that it's probably fueled the rise of macro memes and animated gifs.

Who knew there would be a silent film[1] renaissance in 2016?

[1]: https://en.wikipedia.org/wiki/Silent_film

I think that 85% is pretty reasonable, most content creators provide subtitles now, and the audio is often superfluous (I'm looking at you BuzzFeed Tasty clips!).

85% of all videos I watch are without sound. BECAUSE I'M LISTENING TO SOMETHING ELSE!

What an utter waste of resources. I don't even want to know how much power these videos consume on server side and client side

I've fielded calls from several older relatives who thought their computers were broken because videos on Facebook had no sound. They saw the video playing and had no idea they would need to take further action if they wanted sound.

95% of the videos I watch on youtube are without video - I just use playlists for music.

When you train (involuntarily) your users of low quality and spam/ads, they will be using it in the most unobtrusive way and the way you hoped they wouldn't.

My FB videos don't auto-play. There is a play button displayed in a still. I'm glad it doesn't auto-play, but am curious why I don't observe this.

Maybe you've turned this feature off in the settings -> videos tab? Some people see it, some people don't. I don't see that option now, but I remember turning it off as soon as Facebook introduced autoplaying videos.

Mine never used to, and would always bug out when I clicked play due to having flashbock enabled in firefox. Then I disabled flashblock and now all of my feed videos autoplay (and don't bug out).

For some reason I read your entire post in a voice sounding equal parts surprised and annoyed at Facebook not optimizing for the Flashblock add-on.

However, I'm going to assume that's just my narrow, PEBKAC'd mind.

The average conversion from facebook videos are like 0.001%.

Just fyi.

Sounds like they could optimise the costs by encoding two versions of the video - one without the sound that would play by default, and the other when you click on the actual video.

The CDN invoice probably isn't the smallest of sums you can imagine :).

They wouldn't want a few second pause while the audio is buffered.

A tidbit that I found interesting recently is that their video stream is different from the audio stream source.

says a comment (by yelnatz) above

The size of the audio track is probably around 2% of the size of the whole video. I don't think they'll bother :).

2% of a bazillions of dollars is still like nearly a bazillion dollars.

They'd have to either strip audio out of the video stream on the fly, which will probably force them to use less "dumb" CDNs and incur a processing time cost (which, like most today's on-line businesses, they happily offload to their users now), or they'd have to keep two copies of the same video on their servers, which would make it cost 2x "a bazillions of dollars".

You are very probably right the simple solution makes sense. It is an interesting optimization to be able to make though.

Videos starts playing when automatically, Most annoying feature of facebook.

And a lot of the videos Facebook shows are stolen: https://www.youtube.com/watch?v=t7tA3NNKF0Q

I wouldn't be surprised if most people don't even realise that they watch a video and not a gif. People are not known to understand interfaces intuitively.

A tidbit that I found interesting recently is that their video stream is different from the audio stream source.

Kind of a no-brainer optimization if you think about it.

Not really surprising. I rarely watch videos that people share on facebook or reddit, however if I'll watch gifs all day on the same sites.

Judging exclusively by the number of videos I see with captions, I imagine this statistic was at least intuitively known if not specifically.

Talkies are just a fad anyway ;-)

Mostly due to Autoplay... Having no concerns on wasting bandwith sure helps

same with pornhub I guess ;)

I'm judging you now!

(Auto)Played != Watched

I reckon this is also true for Instagram

Auto-play: what a waste of bandwidth.

"watched" :)

Video on Facebook is shit

or eyes! lol

You know why we turn the sound off? Have you heard how whiny and self righteous millennials are?

Who actually uses Facebook?

The whole world....

Denying facts doesn't make you look intelligent.

Definitely not 'the whole world'. Almost nobody I know uses Facebook any any serious sense. Facebook 'login' - yes. Sometimes to contact someone. Some lazy browsing. But nobody uses it in terms of the 'Facebook experience' as we come to know it.

I suggest that the majority of FB 'daily active' users amounts to logins. I check FB once every few months and yet surely, I'm commenting or logging in somewhere under FB pretenses.

I suggest FB is dying as we know it, though clinging on in peripheral usage and it's new messaging app.

Actually I have been hearing lots of “trendy” Gen-Zs say this, but the funny thing is all of them post on Facebook regularly, just not as much as their other networks.

Facebook is becoming like email. Everybody uses it, but the “cool” people use something else too.

Because bots don't have ears.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact