Did Facebook’s faulty data push publishers to make terrible decisions on video? | Hacker News Reader

Did Facebook’s faulty data push publishers to make terrible decisions on video?

niemanlab.org

207 pointslaurex106 comments

I have unlimited 4G data but I never watch videos. For one I don't want sound blaring out of my phone. Second I can actually read faster than your video. Information density.

Our ancestors invented the script for a reason and its not going anywhere.

> Our ancestors invented the script for a reason

Despite agreeing with the conclusion, I don't really think this is a reasonable statement. Our ancestors domesticated horses for a reason, that doesn't mean new technologies that weren't available then can't entirely supplant that need.

For videos to work, I would assume that you'd need the audio to play as well, in most cases anyway. The thing is, I don't know anyone who doesn't mute their phone (or computer) all the time, and old un-mute when they really want to watch, and hear, a video.

It's certainly Facebooks fault for giving out wrong statistics, but I also feel like the advertising industry have failed to do extremely basic research.

I don't. I keep the ringer volume muted, but the media volume turned up - this usually serves me pretty well.

...until I hit an autoplaying video somewhere, but I usually just learn to avoid sites that do that.

Most Facebook videos have captions burned in at this point.

If the ads are in any way related to conversions, the data comes home to roost very, very quickly.

Ad spends are measured and results are tracked.

The value of those ads is a function of how far the needle is moved.

It's only when the needle can only be loosely measured can this go awry. For example, in a brand awareness campaign, it's sometimes hard to measure effect.

If FB was reporting '80% viewed' when it was '20%' then it's basically fraud.

What some guy says at a conference about 'the future of ads' is almost irrelevant, people are saying things all the time, and even if companies were to allocated more budget there ... they should see the results for themselves and adjust.

Yeah, time viewed is really only a vanity metric that only someone building a campaign with no direct result at all would look at.

The kind of big business campaign just blowing money on branding is probably getting an even worse return on TV.

Time viewed is not a vanity metric for a brand awareness campaign (fyi those a 'real things'), or even an indirect campaign say for a movie launch where a 3 second view is not the same as a 25 second view.

If you're launching a big film like Avengers, a 25 second view is a like a 'soft lead' and you can correlate that to how many tickets you're going to sell.

Brand awareness is much more intangible, but they are in the end real things.

But yes, the softer metrics are massively abused by agencies, and in-house marketing teams trying to justify spend, hires etc..

Weirdly, so much of the 'conversion' stuff is utter crap, like ads for 'photos of Jet Fighters' or 'actresses with and without makeup' where they jam 40 ads on the screen each time. It's this kind of crap that really ruins the internet.

I wonder if, when setting CPMs , the advertisers apply blindly the returns from TV advertising to online.

Brand awareness campaigns are measured (and billed) by impressions, not by "average length of video viewed," which was the metric that Facebook lied about.

No, brand awareness campaigns can by run using either images or videos (or any other medium), and if the campaign is video based, then '% video completion' is a fundamentally critical metric.

It's simplistic to lay publishing's woes so squarely at the feet of Facebook's video stats. This was contributing factor, certainly, but the executives who pivoted to video weren't abandoning 7% margins on text in search of 15% margins — if they were, they'd likely have left the text operation intact and built video atop a successful business. Editorial operations were already struggling! In some cases, the promise of video may have spurred additional investment in publishers that wouldn't have been able to get it with just text. While video is more expensive to produce, the CPMs are also higher, which creates a bigger cashflow. It seems likely that this was a factor in Buzzfeed's NBCUniversal rounds.

Many orgs had to let a bunch of people from text to free up resources for video, video being so much more expensive than text/audio.

Hold up - they didn’t “have to” let anyone go, they chose to because they wanted to hire other people.

No. They didn't "had to" produce more video content. But once they decided to do just that, they had to let people go because producing video is more expensive. That's what the GP is saying.

This may be true, or it may be that the "pivot to video" was the excuse they used for something they were going to do anyway.

Spot on. News publishers have been suffering from declining ads revenue for years. They needed to pivot to something anyhow.

In other words, everyone had the same bad data: The investors, the marketers, and the publishers all thought video was a winner.

...Everyone except Facebook, who was making money, and sat on the data for a year.

Don’t ad buyers have internal mechanisms to track conversion rates and revenue uplift independent of what the publisher tells them?

Facebook was hardly the first player in the video game, with YouTube, Hulu and a few other big brands relying on this model exclusively.

Keep in mind that buyers are everyone from Big Corp down to Mom n Pop. The smaller the shop, the more you must rely on the metrics you are told. And even small places may value awareness and/or indirect conversion. In fact, a local, interest based ad may be totally focused on awareness, hoping to increase the chance that you walk into the store the next time you drive by.

but why are CPMs higher? i mean do they actually get higher return of investment or is it all based on faith again?

It wasn't just Facebook metrics that made people jump on the video bandwagon. The whole video ad stack from player to ad server to header bidding wrapper to bandwidth is full of ways to spend 4-6 figures per month on fixed overhead and variable cost to run a player at scale, let alone production. Most pubs have been financially burned on video by an industry that can't quite understand why it isn't working. All the while, marketers continue to have the largest and most impactful influence ever with TV.

I know why it isn't working: human hands are made for interacting with things. While sitting in front of a TV, we humans tend to believe what we're told, and we're typically prone, not on guard. Who wants to hold an ad in their hands? Who doesn't feel burned when a video ad interrupts their flow of their digital (fingeratively) pleasure trail?

Now imagine what kind of influence can be had over another when they are in a fully immersive digital (binary) experience. Its like handing over your entire limbic system to the machine.

Malevolent influence at scale is remarkably accepted by humans, it is our past. I hope our future moves towards respecting the sovereignty of individual human consciousness.

Laws, attitudes, and behaviors need to change. We all who have been Branded have had a piece of our identity conformed to serve the ambition of another. I think the average American brain walking through the grocery store is analogous to a chairlift pole covered in stickers, a wall covered in layers of gum. Any exactly how do you clean away an Impression?

I'm a recovering internet marketer. I've felt lousy for a long time for having been so cavalier with influence. I've been trying to see what better looks like for a long time. What do you think better looks like? Feels like?

I believe people are just applying whatever worked on TV blindly, but on the TV you are buying a part of the flow, while on the web you are buying an interruption to the flow which should discount its cost right out the door. And while there are many ways with which text ads can fit in a document layout, there is not many ways to lay out an ad in a video. I find the product placements that many youtubers do is more appealing, and a better way to promote a brand rather than "this ad will close in 20 seconds"

Weren't people spending more time on video on Facebook due to Facebook prioritizing video and pages turning still images into videos of still images?

Anyone else know what I'm referring to here? You see an image someone shared, but it's a video, so you click it, and it's still a still image, because page admins were trying to game the system.

What? Really? I don’t use FB but sounds like something anyone would do to game the system

Yerp. Here are a few links regarding the issue:

https://www.quora.com/Why-do-people-post-short-videos-of-sti...

https://www.socialmediatoday.com/social-networks/facebook-cr...

thanks, people always trying to outsmart others...

How is this different from Volkswagen? They inflated a metric by 600+% according to the legal documents on the article, for 2+ years. And kept all the revenue from the 'mistake' while crushing competitors that lost business for those two years.

I’m not sure if you are suggesting FB should face similar consequences as VW or that FB’s conduct is excusable because others do similar things.

Some differences are that VW plead guilty and paid a 2.8B $ fine, the CEO resigned, executives have been facing criminal charges, and a slew of international cases are still outstanding. Another difference is that VW’s deceit had direct human health consequences.

From the the linked post in the article (https://www.wsj.com/articles/facebook-overestimated-key-vide...):

> Ad buying agency Publicis Media was told by Facebook that the earlier counting method likely overestimated average time spent watching videos by between 60% and 80%, according to a late August letter Publicis Media sent to clients that was reviewed by The Wall Street Journal.

From the FB post (https://www.facebook.com/business/news/facebook-video-metric...):

> The metric should have reflected the total time spent watching a video divided by the total number of people who played the video. But it didn’t – it reflected the total time spent watching a video divided by only the number of “views” of a video [...].

How does this work?

They explain that a "view" is when the user watches the video for more than 3 seconds.

View time is probably highly heavily skewed towards plays that last less than 3 seconds. (also because facebook autoplays videos without any pause time)

If 60% of viewers play the video for 0-3 seconds, and only 40% view the rest of the video, you're grossly overstation the average view time.

Removing users with less than 3 seconds seems 100% intentional (to avoid exposing the fact that most viewers are not really viewers, they immediately dismiss the video), not an error.

Right on.

In other words, fb measures were correct for "people watching videos for 3 seconds or more". But they didn't: 1. Qualify their numbers with that. 2. Disclose how many people were NOT watching videos.

Without these two critical details, the numbers represented "all Facebook users".

No, they were still incorrect for "people watching videos for 3 seconds or more." For that to be correct, the total view time (numerator in the figure in the article) would have to be view time only for those users who watched >3 seconds.

Thanks, this makes sense. If you think of this in terms of distraction, it's huge.

> “The best way to tell stories, in this world where so much information is coming at us, actually is video,” Mendelsohn continued. “It commands so much more information in a much quicker period. So actually, the trend helps us to digest more of the information, in a quicker way.”

This opinion is hard to swallow. This is not . at . all true. If the "information" you're trying to convey is colors and bokeh, sure...

A lot of the comments below this are rebutting this point by mentioning that videos are better for learning, or complex tasks. This seems obvious to me. Unless I grossly misunderstand how most people use facebook though, they're not there to watch videos about learning or car repair. They're there to consume content as a distraction. It's right there in the beginning of the quote: "the best way to tell stories..."

In that context, this comment is absolutely spot on - we're not talking about a tutorial video showing how to replace the headers on your BMW. We're talking about a video with a talking head that reads a "story" to you. "Can you believe what the president did today?" (cue outrage).

"It commands so much more information in a much quicker period," is a nice way of saying "video distracts people much more effectively than text."

>A lot of the comments below this are rebutting this point by mentioning that videos are better for learning, or complex tasks. This seems obvious to me.

Video is better for learning? I've participated in many online courses and I have noticed consistently that for a lecturer to convey the amount of information that could be obtained in 1 hour of reading they would need at least 3 hours of lecturing.

The only reason I listen to lectures in addition to reading is for metadata. IE, what do they emphasize. And typically I need to do this at 2x speed or else I fall asleep.

Taking a step back, the idea that learning is about fast data transfer is probably not going to hold up to scrutiny at all.

Learning needs to be data transfer at the speed with which your brain learns, which isn't going to be as fast as it can process data through an existing conception of the world.

You need to consider complimentary data incoming through multiple senses, and data which is better represented in moving picture form.

Id argue it 1) primarily depends on the topic 2) also primarily depends on a specific persons brain 9/10 I hate video results.

For 1 if I want information I don't need logos/animations, I don't need "hello fellow YouTubers", etc.

Event the "repair car" example. Suppose need to make some action to repair a vacuum pump. Unless I just don't know where it is in the engine bay (or some other location/orientation information) what value is looking at anything?

Reading is way quicker than watching a video. Speech is just too slow.

If an informational video doesn’t have a transcript or a way of speeding it up I’m not interested. At work we have a subscription to Lynda.com and while the content is okay, the only way the videos are bearable is to watch them at 2x speed.

Yeah people in videos talk slow. They also do the whole introduction thing.

Videos are good for ads. And videos are interesting I guess if you actually CARE about the guy who talks in it. For information and learning: text.

Agreed, there’s generally a lot of filler and intros for about 30 seconds to a minute of genuinely useful information.

I guess it’s the same as most pop businesses books: there’s about a blog post’s worth of material, padded out with case studies, anecdotes and other fluff to reach the page count.

It really depends. If you're conveying a dance routine for example, text is horribly inefficient, a short dance instruction video can be much better.

I have the opposite for news for example. Essentially, whenever people read text out loud which I could have read myself, I typically prefer reading. Moreso because I can skim through text, and move back and forth to make connections, re-read a part after a new insight or to check if I'd remember it clearly, with video this is quite difficult and tiresome.

Although I do enjoy audio/video played at high speeds, it's much more effortless to listen to or watch double speed video than it is to speed-read. Particularly because most written text is quite information dense and not meant to be sped through, whereas most video is more conversational, less dense, and can be sped up to a point where you can take in a lot of information in an easygoing manner in a short time.

Then there's a unique section of high-quality production material. Here I would say, video can beat most formats, it's really efficient and makes a long-time impression. For example, a $5m documentary with video material for every anecdote, beautiful animated graphs to illustrate every metaphor, every causal-mechanism explored etc, can beat 100 pages of dry text. But it requires tons of effort and resources. On the other end of the spectrum is simple video material you see on television (point camera at talking heads), which is really inefficient (safe for being able to speed it up, but even then... whenever people are reading out loud news or an opinion, I'd rather just read it myself then watch someone do it for me.) But this format is really cheap to produce, it's just a camera and talking heads, but inefficient, I dislike it myself.

“Writing about music is like dancing about architecture.”

- Martin Mull (https://quoteinvestigator.com/2010/11/08/writing-about-music...)

This seems inherently contradictory.

Secondly, maybe dance has something to say about architecture. Have some imagination. Judging how others appreciate just makes you look ignorant. This is not a new theme to art criticism.

A few months ago I would have agreed wholeheartedly. I recently had to replace the alternator on my 2008 solara and after finding instructions in text form unclear I found a video on YouTube that made the process extremely easy. Thing is, there were a few tiny details I was able to glean from the video that made the difference being very frustrated and very successful.

It's certainly better at some things, but in general, I think it's a worse platform for conveying information.

- find content within a video is difficult (I'd rather search text) - you're at the mercy of the presenter's pace (you can adjust speed, but that's only part of the story) - sometimes listening isn't an option - sometimes presenters rely too much on visuals, so you have to watch as well - video can't be as easily edited as text, so most people just put subtitles to point out errors (this doesn't help if you're not watching) - less easy to link to additional information than text

In general, unless someone is demonstrating a step-by-step, graphical process, I prefer text.

For complex instructions like vehicle repair, videos can be amazing -- especially with good focus and lighting.

For letting me know what Trump tweeted yesterday, I don't need a five minute video with commentary. Even for protests and marches, a few pictures and a paragraph will pretty much do. Most news I'm happy to just read about.

Agreed. I guess video vs text is like most tools. Use the one that’s the best fit for the job.

i think this belongs to the " colors and bokeh" category. It all boils down to: how can you give the same information faster. For a car repair, a video is much quicker way to show what goes wher. The vast majority of videos on youtube, though, are elongated transcripts read aloud.

> This opinion is hard to swallow. This is not . at . all true.

I used to think so too. Then I saw how my child learns stuff with far greater command of underlying concepts and connections through video and realized it's probably just a function of how I learned to learn or am used to growing up without essentially videos on any topic available on demand.

I have observed that the learning my child acquires through video is actually not "book learning" (the term exists for a reason) but much more nuanced, vicarious and realistic in a way text simply isn't. If fact, my child's goto for understanding something "let us search for a video on YouTube that will teach us how to do X". From an information theory standpoint this also makes sense.

So let's allow for the possibility that our opinions are simply shaped by our experiences of growing up without readily accessible video and reality for our children may be different.

Purely speculation, but I think a well made video/lecture is the best way to learn concepts and reading a concise document is the best way to get information. Videos have richer tools with visual aids and tone, but text is searchable and fast if you know what to look for.

Not all the time. At least when it comes to subjects involving physical things as opposed to software or theoretical subjects, video greatly expands the amount of raw information captured at the expense of total density. It takes quite a bit of control out of the author's hands, so to speak, by capturing far more than the words they explicitly choose to write. When it comes to fixing or building things, for example, this means that video has the potential to capture a lot more incidental information that the author might never think to include like the speed and strength of movements as well as tactile responses that are nearly impossible to describe in a consistent manner.

I think text is better when you already know what you're looking for and video is better when you are still trying to develop an intuition for the subject.

I would argue that for a deep understanding of a system, the best medium is in fact a simulation game that incentivizes you to optimize your way through that system.

For me, I was taught how to learn from reading the course material, and then follow that up with a lecture. Eventually, I could even do away with the lecture. However, in the early 90s, there was a series I saw on PBS that was a visual course on physics. (I’m thinking a WGBH production, maybe WQED). After reading the material followed by the lecture, these videos really made things clear with their visual reinforcement that the classroom could never do.

Like everything in life, YMMV, and everyone learns differently.

Almost nothing especially random video content on ads makes complex use of audio visual aids that would enhance comprehension.

Its simply the same stuff you would put in an article communicated at a fraction of the speed with less editing for brevity and prose because it's much harder to reorder video or reshoot instead of swapping a few words.

It is probably worthwhile primarily because it makes a stronger emotional connection which is to say its a more effective vehicle to deceive and manipulate you.

If you think videos are better for learning value obtained per minute I would be interested in reading the proof and can certainly keep an open mind but I'm not inclined to value an anecdote.

Does he actually believe this? It sounds like a marketing fluff piece that some intern had to make up.

I mean the guy is Facebook VP for Europe and probably knows very well that business reports (for example) are not digested "in a quicker way" if you make a 3 hour video out of them.

Apparently "information" is a euphemism for "attention".

Nicola Mendelsohn is a she. https://en.wikipedia.org/wiki/Nicola_Mendelsohn

Oh, thanks for clearing that up! Although the name can be either female or male (and is probably even more likely to be female) and the article doesn't indicate the gender I just assumed that it was a guy.

I guess it shows that my perception on the industry is still very much influenced by stereotypes; no matter how much I would like to think that I'm immune to that.

Probably an unpopular opinion here: I personally find that I understand and remember things a lot better if I learn them from video (as opposed to reading text alone). Anyone else like this?

I'm amused seeing these binary words-are-better/video-is-better anecdotes.

It depends on the context, the people creating the content and the people consuming the content. Sometimes video is better, sometimes text is better. It just depends.

When I wanted to learn about (perhaps ironically) bookbinding, seeing videos about it was great. However, for quick reference back at how to (for example) tie a kettle stitch knot, a book is much quicker. Video is not as convenient for skimming or using in a random access manner, and search is definitely worse.

For learning something related to coding, I far prefer text to video, although that may be partly because it is easier to copy and paste into an editor so I can mess with it myself.

I could totally believe that, I'm not anti-video. I definitely prefer it for some contexts. On the other hand, I find it painfully slow for more abstract concepts and I can generally learn far more from a book than from a documentary.

Good video can be better. Most video is crap. The problem is that crap video is much harder to correct than crap text. You can skim a slow article. Much harder to skim a video. Even AI can extract the most important information from a news article. Much more difficult for video.

Video is great ... as an art medium.

I’ve been in and around the commercial spot industry for most of my employable life. I’ve seen the rise & fall of production work due to the lower cost to entry to professional quality gear and the bottom falling out in ‘08. TV ad sales have never gone away though. The production budgets fell, but not the ad buys. There now exists video productions solely for online, and it will never see TV. From my conversations with these agencies, online ad click through rates are still a drop in the bucket from their TV presence. How they justify this I don’t know, and I know that’s supposed to be a huge selling point of online advertising.

Does that mean video is more effective than text?? Depends on your agency I guess, or maybe depends on the campaign.

How is the ROI or whatever the TV equivalent of click through rates measured for TV commercials? Is the measurement different between ads that focus on selling something vs those that focus on maintaining brand awareness (like Coca Cola and the big car makers)?

That's what I've never understood. I'm in the camera department, so I'm just concerned with with acquiring the picture. I've chatted with some of the ad peeps on set, but they are mainly concerned with getting the spot produced. I've never properly chatted with the analytics folks. I'm assuming in the old days, it was Nielsen ratings. Today, I'd assume your digital cable box is providing all of the analytics ever needed. Still don't know how comparing eyeballs viewing an ad translates to people spending dollars in a store.

I have had direct conversations with small market clients. The client hired a new agency, and that agency hired us to shoot the spots. The agency came up with a decent small market campaign. The next time it came time to shoot the next series of spots, the client mentioned the spike in business from the ads after they started airing. Yes, the agency also had radio and online campaign, but it was the TV ads that resulted in the spike in calls. ROI must be pretty good. In a single day shoot, we can shoot 6 spots for this client. They then air 1 spot per month.

I'd imagine the smaller market clients can see a significant ROI vs mega-clients shooting brand awareness campaigns. If you've ever seen Mad Men and how they put agencies on annual retainers, then you'll get what I mean. The Madison Avenue crowd probably have tons of research, but that's so far out of my circles of interest.

My theory is that all is faith-based. Because the amount spent on advertising is relatively small, advertisers are happy to spend whatever the publisher asks, and their management will still be happy. As long as both advertisers and publishers believe they provide value to each other (and they definitely do, they just don't know how much) the model is not questioned and everybody 's happy.

Online banner advertising is often similar , with the major difference that there is a limited 24 hours in a day in which the TV budget gets allocated, while there are vastly more publishers on the web.

TV and YouTube/Facebook video ads are totally different. First, you can target audiences on YouTube/Facebook you can't target on TV. For example, people who are interested in project management software. Second, YouTube and Facebook ad results are much easier to track than TV. There's absolutely no doubt in my mind that as far as video ads go, YouTube/Facebook is the way of the future, not the cable TV. In fact, most likely TV ad option is just going to be an extra option inside YouTube/Facebook in not so distant future.

> In one, Facebook had inflated Average Duration of Video Viewed from 2.0 seconds to 17.5 seconds (an increase of 775%); in the other, Facebook had inflated Average Duration of Video Viewed from 2.4 seconds to 17.3 seconds (an increase of 621%).

How does Facebook measure the success of video viewership? From my experience, Facebook forces you to view the whole ad. This contrasts with YouTube: You are forced to watch for a few seconds and can opt-out later.

This gives an important signal: The viewer was not interested and thus bailed out. He didn't watch the whole video. This can't happen in Facebook: You are forced to watch the whole video-ad if you want to carry on watching your original video.

Another complaint: Auto-Play. If I'm watching a video, Facebook will auto-jump to another video in a ridiculously short-amount of time. There is not opt-out of this behavior. You are forced to watch the next video; and potentially the next ad.

Again, this contrasts with YouTube: Auto-Play is opt-in. The time before the player jumps to the next video is significantly longer. The time before the player jump in Facebook is so short that I almost always fail to prevent it.

So dark-patterned Facebook these days. It is really annoying, scary and unfriendly.

>“The best way to tell stories, in this world where so much information is coming at us, actually is video,” Mendelsohn continued. “It commands so much more information in a much quicker period. So actually, the trend helps us to digest more of the information, in a quicker way.”

How bad of a reading disorder do you have to have to think you can extract information more efficiently from video than text?

Nobody is mentioning video ads. I would gladly consume more video on my phone if there wasn't an adroll.

I can use adblock on my computer, but it's more difficult on the phone (DNS66, etc) so less likely I will have adblocking on the phone.

As publishers moved more to video to capitalize on video plays, so did ads come with them, thus video became less appealing for me.

That's how it played out for me the past 2 years, and probably you too.

Why is it more difficult on a phone?

it is exactly the same process as a desktop. First install a decent browser (not the crap one shipped by the OEM), for example, firefox for Android. Then install your favorite adblocker extensions (which should be uBlock Origin). Done. No step 3!

This sound bite from WNYC Marketplace [1] reminded me of how shakey creative industry has always been.

"Who cares how long people watch videos on facebook? News organizations...thinking video was where the ad money was." [...]

Ben Bradford (reporting). [...] Just months after he was hired, Gonzales was let go [as a video producer for Mic].

Gabe. And we started seeing the same happen everywere. Any you know I think at that point you realize the jig is up, right?

[1]: [Marketplace Business: A look back at the "pivot to video" after a new lawsuit alleges Facebook misled about video viewership](https://www.marketplace.org/2018/10/18/economy/facebook-pivo...)

> What does seem clear now is that Facebook’s executives’ statements about video should not have been a factor in news publishers’ decisions to lay off their editorial staffs. But it’s hard not to conclude that publishers heard that rhapsodizing about the future and assumed that Facebook knew better than they did, that Facebook’s data must be more accurate than their own data was, that Facebook was perceiving something that they could not. That their own eyes were wrong.

If you're making massive, sweeping changes to the structure of your business and product offering based on Facebook PR, you deserve to go out of business.

What FB said could have even been true and these organizations could have staffed up on all video only to later become victims of an algorithm change or some arbitrary FB rule in the future.

Did internet ad fraud con advertisers? Direct clicks and conversions are such lowest common denominator crap. Suitable for ponzi schemes and snake oil sales. Tesla advertises amazingly well online. Their conversion rate is clearly zero. Not close to zero, not "they don't care" It is zero, nothing, not one single conversion.

Across the breadth range of companies built on advertising from coca cola to chanel as far as I can see they don't use internet ads. These companies have not gone away. Google, Facebook, ebay etc etc all have poster ads in my town. I guess targeted internet ads didn't work for their purposes.

What you are describing isn’t a con. A coke billboard fills a different purpose than a coupon or a transactional online ad.

What is a con is telling advertisers that I watched something for 5 minutes, when in fact I watched for 5 seconds. Imagine if you were an advertiser who paid for a product placement in videos. Your campaign was a total waste of time.

> Across the breadth range of companies built on advertising from coca cola to chanel as far as I can see they don't use internet ads.

"Coca-Cola has just hired its first chief digital marketing officer" (2016) https://www.thedrum.com/news/2016/12/12/coca-cola-has-just-h...

"Chanel has been investing in other areas of digital, boasting more display and video ad impressions in the past year than almost any other brand in the study." https://www.l2inc.com/daily-insights/chanel-flaunts-digital-...

There's a distinction that needs to be made between brand advertising and performance advertising. Brand advertising is most of what we've had in the past (most TV ads excluding infomercials, most Newspaper ads - not including classifieds of course).

Why is a Rolex "worth" $10-20k? Why is an Apple product priced at 2x the price of an equivalent? In large, because they spend millions every year advertising their brand - with no measurable conversion mechanism. In my opinion, this is why online advertising is so obnoxious - it's mostly low effort performance advertising which tends to be much less enjoyable from an aesthetic level.

The moral of the story: if what the "person/organization who ought to know about these things" is telling you doesn't make sense or doesn't match with your own experience, consider the possibility that they might be just plain wrong. Especially if it would be wrong in a way that makes them look better.

Absolutely this. We're actively building a system to better monitor effectiveness of social content and it's amazing hearing the relief from people when they're seeing numbers for the first time that look like actual KPIs versus platforms spewing out self serving vanity metrics.

FB and much of the current social marketing scene in general have been handing off best practices that feel highly questionable at best.

What about the equally important take away that you can trust Facebook to lie at any chance it suits them?

Oh, I think that was already clear before this. :)

The upvotes say people found it noteworthy.

I'm curious why Facebook chose to share the wrong metrics. What was the intent?

"why Facebook chose to share the wrong metrics"

Why did you choose to include that bug into your code?

Srsly, people don't choose to make mistakes (but sometimes they choose to do cover up).

Also, there is no such things as "Facebook". Facebook didn't choose, someone inside Facebook either make mistake, or deliberate decision. "Facebook" is not single entity with single agenda and plan, but complex company, with different people optimizing for different things (i.e. security/privacy and growth and ads). There is shared high-level vision, but not every decision is made at single place.

State a falsehood with the intent to deceive is implied by "lie." I'm curious what intent people are ascribing to this action.

You say it was a mistake, at the same time you give many reasons why it likely wasn't a mistake: many people vet the public-facing communications of such a large company.

They make more ad money on video, especially when their measurements overestimate engagement.

As a bonus, it pulls money away from YouTube.

Inaccurate, the misreported metrics didn't impact billing[0].

[0] https://www.businessinsider.com/facebook-exaggerating-view-c...

Oh yes they say that, but the increased video ad placement the metrics supported did.

You're glossing over the bigger problem. The "person/organization who ought to know about these things" not only has possession of all the data about user behavior on their platforms, they also control the means by which content producers access users.

You're understating the monopolistic power that Facebook's network effects have produced in the online content space.

It's only recently that third party analytics companies came into the space with an incentive and enough data to make meaningful market-wide forecasts.

The excuse for all the trackers involved in online advertising is exactly because not a single advertiser trust any publisher.

Your comment is still very valid for small clients on self-serve systems, though.

I think the other lesson is to be wary of sales pitches being sold as an expert recommendation.

From the linked FB post (https://www.facebook.com/business/news/facebook-video-metric...):

> This error should not stand in the way of our ultimate goal, which is to do what’s in the best interest of our partners and their business growth.

Hooo wait, I thought it was about connecting the world!

> “All the while, Facebook continued to reap the benefits from the inflated numbers”

I don’t see how Facebook benefited from the data being wrong. Facebook wins by having publishers to produce formats and content that users actually want to see.

Facebook was launching it's own competing news thing wasn't it? eg clearing away some of the potential competitors

This seems like that lawsuit against Google adsenes while back. the settlement, if any, will be very low and will go mostly to lawyers. like google, Facebook is very, very good at winning its cases and paying very little otherwise.

This reads like a carefully disguised defend, deflect, and dilute piece crafted against the publishers on belhalf of Facebook. Are we sure it is not paid for by Facebook or some entity aligned with Facebook?

This truly feels like an article that doesn’t understand the landscape of digital advertising. It is naive to believe that publishers are taking queues from Facebook on how to publish their content. All publishers ( including Facebook ) take their content queues based on what advertisers want. This is their source of revenue, and the only way to move the bottom line as a publisher. Video advertising is far more lucrative than banners. In order to create a consistent user experience, you cannot have video ads alongside static content. While the miscalculation is certain metrics might have made Facebooks inventory look a little better, it 100% did not have an industry changing influence as indicated here.

I knew it! But seriously I’ve heard about this “ trend” for years but anecdotally most people I know don’t consume random videos on social networks.

Vids from friends?

That seems to be a different story.

So Facebook repeatedly gives publishers incorrect info, and some of those publishers go broke.

Meanwhile, Facebook is getting it's own (news) publishing operation off the ground. Which could probably do with having less competitors.

That could be viewed in a rather specific light. "Operating in Bad Faith" might be the nicest way to put it? ;)

Reply on news.ycombinator.com