Yes it’s a negative feedback loop created by the very bad product decisions they made to optimize for local maxima of getting more ads revenue. You inject more and more ads followed by more and more suggested content, thereby reducing the in network content. The in network content doesn’t get as much of traction, therefore people stop posting. This in turn forces the feed to have more and more suggestions and a few stale in network posts from days ago. This drives engagement down from people who want to see in network content and they leave. So what is left is people who do engage with suggested content, pushing the product to make decisions to push even more suggested content. All of this continues till eventually fatigue sets in and a sudden rapid drop in engagement kicks in because your global maxima of a quality product was lost long time ago and your product dies.
No, this isn't a consequence of the status quo newsfeed algorithm. This is a consequence of Facebook making a purposeful strategic shift away from in-network content to suggested content. This strategic shift was highlighted in their last earnings call as a way to compete with TikTok, whose success is partly explained by their focus on suggested content. On the earnings call, they signaled that they were shifting more to a mix of the two content styles in the coming quarters, but the OP is likely enrolled in some sort of test that is giving them 98%.
See the below excerpt from Meta's Q2 earnings call:
"One of the main transformations in our business right now is that social feeds are going from being driven primarily by the people and accounts you follow to increasingly also being driven by AI recommending content that you'll find interesting from across Facebook or Instagram, even if you don't follow those creators. Social content from people you know is going to remain an important part of the experience and some of our most differentiated content. But increasingly, we'll also be able to supplement that with other interesting content from across our networks. Reels is one part of this trend that focuses on the growth of short-form video as a content format."
>Mid-2021, I stop using FB for anything beyond Marketplace.
The only things I find FB useful for are Marketplace and Messenger. Marketplace is an OK way to sell your junk locally (it basically took over from Craigslist because CL was taken over by scammers and CL never bothered to do anything about it), and Messenger is a convenient glorified phone book and chat app which has a video chat feature that actually works quite well through any browser and doesn't require a local client application like Zoom. It's also really convenient that it works through both my phone and through a browser on my PC, so it's easy to switch back and forth (much easier to type messages on a PC, for instance).
Discord has completely displaced the need for other social media in my experience. And they get hundreds of dollars per year from me. If this isn’t a winning model, I don’t know what is.
People focusing on the winning model are missing that none of the platforms fill identical social roles imo even if they copy features from eachother and that digitally immersed users seem to prefer it this way. I strongly doubt that TikTok would have been improved with the full feature set of Facebook even if it might benefit from integrative features for example. Discord in a similar fashion is where I go for real-ish time talk with friends or gamemates. Both of these are very different than Instagram which has largely become what my wall once was on Facebook but not to be confused for Facebook itself where I go to talk with people who aren't very cyber savvy and by extension join Groups targeted at non-cyber savvy demographics (like my HoA's group)
Discord is completely useless for socializing with anyone who isn't in the IT industry. My mom, for instance, is NOT going to use Discord, but I can get her to use Facebook Messenger.
Not just older folks, young people who don't game regularly don't bother to check it either. The UI makes it quite hard to find people on there, nobody uses real names, profile pictures are random game/anime characters...
A few things that make Discord unfeasible for general purpose messaging...
The need to host a server somewhere. It silos info, makes that info less permanent and less findable. I've gone to all iDevices at home, so can't self-host.
It also doesn't have great visibility outside the tech worker and gaming worlds. I'm happy to use it if somebody sends me a link, but it's not the default for most people I know (except for those that game). Slack is a bit more common, mostly because of broad adoption at workplaces.
> This is a consequence of Facebook making a purposeful strategic shift away from in-network content to suggested content.
In an ironic twist these changes has made it much easier for me to stop using Facebook and Instagram. My usage was never high to begin with but some people in my life only used those apps and I wanted to stay connected. Now every single time I visit either of those apps and I see suggested content I close the app and move on with my day.
Bad strategy, imo. Sounds like they should have tried to answer why people used their product in the first place. “A competitor does it” isn’t a valid excuse—its like suggesting The French Laundry switch to fast food because McDonald’s is a competitor.
I think your analogy is too generous. It’s more like putting an engine on your stapler because people buy cars more often than staplers at a higher margin despite having a monopoly in the stapler market.
They also need to be honest with themselves about whether TikTok is a competitor, or if TT has simply won a completely different business that happens to take customers' finite resources (time) from Facebook.
The French Laundry is a very niche product, which Facebook is not. I think a more apt analogy is suggesting McDonald's to include more "healthy" options or a coffee bar because the competitors do it and it's a successful business strategy.
The real danger FB finds themselves in is that the initial value prop (make and maintain your friendships with less friction) is LONG gone.
Instead, they are relying almost entirely on the network effect now. You have to be on Facebook because your grandma is on Facebook. This is not sustainable long-term, for obvious and perhaps morbid reasons.
A 'strategic shift' to present suggested content may be what Facebook wants to do (or would like to tell investors it is doing).
However, the amount of real content being posted appears to have fallen off a cliff. This is obviously very subjective but holds true for both my own experience and friends/family who have commented.
So there isn't just a similar level of content but the algorithm is displaying less - but literally everyone I know no longer posts to Facebook in remotely the same way.
The only time I see regular updates are 'happy birthday' posts.
Everything else has moved to other apps - mostly Instagram stories.
The drop in real content is astounding. It used to be the case that I’d never ever come close to the end of the content from people I actually follow when I checked Instagram, and 5-6 years ago I interacted with the app daily. (Note: I deleted my FB account 7 years ago but kept insta).
Now I’m in my early thirties and I maybe open it two or three times a week — I hit the “you’ve run out of content from your friends” message in less than five minutes (and I follow ~400 people). It’s even worse because half of that five minutes is spent scrolling past the INSANE number of ads and suggested posts (most of which are terrible ads as well).
My friends are still actively posting “stories,” but I can’t imagine that will last forever (stories are also littered with ads, it’s insane).
Facebook has been out of ideas for a while, but it wasn’t so apparent until recently since they faced relatively soft competition, and they either acquired popular social media start ups or if they couldn’t, stole from them wholesale (snapchat). They now have a real competitor in TikTok and on the last earnings call it became clear they have no plan for regaining users / organic engagement, so they’ve decided to pump the short term financials by serving everyone a shit load of ads. Facebook’s long term prospects are bleak imo.
Has anyone kept track of the number of times they shifted the Facebook newsfeed algorithm?
At different point in time, the feed prioritized:
posts from friends, posts from games from your friends, friends' sharing of links to compete with Twitter, videos to compete with Youtube, moments to compete with Snapchat, comments that friends made, posts from friends again as growth slows down
I wonder what percentage of users drift off every time they shifted the algorithm.
The difference of course is that TikTok's suggested content is generally relevant.
My entire suggested feed relates to US politics. I keep getting suggestions for Second Amendment groups, "Stand your Ground" groups, pro life groups, "Isn't she beautiful" (a gun lobby group), "Texans for God" and so on. I'm Australian. I guess this is what I get for doing an occasional checkin at a place named Texas BBQ.
> but the OP is likely enrolled in some sort of test that is giving them 98%.
Couldn't it just be that suggested content is not only placed "between" in-network posts to achieve a certain minimum ratio, but also some small number of suggested-content posts are inserted at a constant drip into your feed (e.g. every N minutes in the feed's timeline) to ensure a certain minimum absolute amount of advertising happens per day per account?
Which would mean that, if all your friends are people who don't post very much on FB any more, there could be 50 dripped in suggested-content posts (to achieve the minimum absolute suggested posts per day) for each one in-network post.
Even without "ensuring a minimum absolute amount of advertising", I presume this would have to be happening on some level, to ensure that anyone who tries to use Facebook "like TikTok" (effectively an RSS reader for suggested content), ends up with a non-empty feed.
It makes sense tho for them to go for suggested content, because the can make you like stuff. That's their thing. It's no random thing they've been hiring psyops veterans of the afghanistan war. They can push things
I'm not sure it's the advertising dollars that are addictive as much as it is "growth". To keep the stock price up you have sustain a PE that reflects revenue growth, and once you are done signing up new users... Well... Something's got to give.
I see the ads as a lot like cutting timber out of a finite forest - it's profitable, but potentially harmful to the ecosystem in a way that can make it unsustainable. As long as you don't cut too much, replant aggressively, then you can make a sustainable business out of it - but it doesn't scale past the point where your harm to the ecosystem causes collapse. Once people exit FB because it's all ads, then the network benefits begin to evaporate, a little bit at first, then all at once...
Then again I don't run a social media company, so what do I know. ;-)
>Then again I don't run a social media company, so what do I know. ;-)
Maybe not a little as you think. Just because the social media companies are "runing", running into the ground is still running. They just have so much momentum that the ship is going to keep moving along.
> To keep the stock price up you have sustain a PE that reflects revenue growth, and once you are done signing up new users... Well... Something's got to give.
Facebook need to just settle back and offer dividends. It's over.
Doesn't help that their recommendations are always "Oh, you bought an air conditioner? Come look at this wide variety of other air conditioners we have!" rather than, say, adapters for connecting air conditioners to different window types or home electricity usage monitoring or something.
This is a common complaint. The reason they do it is because... it works!
It can be counter-intuitive but this is often seen in recommendation systems. In music for example, the song most likely to be played after a song is... that same song again.
However, it is painfully clear that there are very obvious categories of products that it does not work on. 99% of the time, if someone buys a refrigerator, they will not be interested in buying another refrigerator.
It would not be that difficult for Amazon to detect these product types and give other kinds of recommendations for them. Unless their algorithms and the data that drive them are much, much less comprehensive than they want us to believe.
Nope, even with items like fridges it's intentional.
Most people have not bought a fridge in the last month. Most people will not buy one in the next month. Taking your estimated 99%, that leaves 1% who just bought a fridge but it doesn't work, or didn't fit their room, or for whatever other reason they've sent it for a refund... and are about to want a fridge again!
1% might sound low, but if you look at the group of people who haven't just bought a fridge... maybe only 0.1% of them will buy one in the next year, vs. that suddenly huge-seeming 1% of recent fridge buyers. Suddenly, not such a bad idea to advertise to them :)
But if I've bought a fridge from Amazon, and it doesn't work, surely the first thing I'm going to do is try to return it to Amazon?
Most people, after all, don't really have the spare money to buy two fridges at once, so they're going to want their money back from the first one before they buy the second.
Once they've started the return process, then Amazon can start recommending new fridges to them.
Again: this is all data Amazon totally has access to. Whether they are failing to gather it, failing to recognize its value, or simply choosing not to use it when they know it's creating a lot of bad recommendations, is unclear.
I'd guess fridges are annoying enough to be without one that many people who can afford, either through having cash or access to a credit card, would indeed be willing to pay for the second fridge while waiting a few days for the first fridge to ship back and be processed before they get the refund.
Or maybe by the time they've fully made up their mind to return the fridge, it's often after they've first decided there is indeed a better fridge out there, and already chosen where to buy it in the day or two before they start a return. In which case the time to advertise to them would be the window after purchase and before initiating a return.
Maybe... etc.
I don't know for sure, having neither worked at Amazon nor tried to market fridges. But I have seen many examples of this sort of retargeting of people who've just bought something with ads of the same type of thing, where general intuition, even that of experienced marketers yet alone normal shoppers, made it seem like a dumb idea and yet the data actually did prove it was a great use of advertising.
So maybe your gut instinct is better than the ad conversion predictions Amazon's code is making, with all their data. But I'd be surprised.
>Again: this is all data Amazon totally has access to. Whether they are failing to gather it, failing to recognize its value, or simply choosing not to use it when they know it's creating a lot of bad recommendations, is unclear.
I'm not sure as to their reasoning. But maybe this decision is in fact based on their data? They're a pretty data driven company
Or the random notifications I get on my phone from Amazon: "We thought you'd like a new air fryer." I bought one 6 months ago from you, I don't need a new one.
Somehow it's always about some big appliance that lasts years and I just bought. It's never about "you might be running out of toilet paper."
It seems to me that Amazon has the same problem, but with their physical products instead of ads. Cheap junk, fakes, and scams pushed up sales for a while, but then people give up on the site and go to other sources that are perceived to be more reputable.
If the logging analogy holds, then I don't think this is purely an issue of profits above all - it's more an issue of short-termism. If you are Facebook, you could reasonably stay relevant, focus on content quality, and shift to a profit model that looks a lot more like Morton Salt's - steady as she goes. In the long run, sans other goals, that might be the rational thing to do.
On the other hand, if you are looking to preserve stock price in the short-term you need to at least create the illusion of sustainable revenue growth. And there is a lot to be said for this later strategy if you don't really want to run Facebook in maintenance phase. Seems that Mark and Co. might not be all that concerned about "deforesting" Facebook with ads because the end game here is the Metaverse...?...
Exactly. By adding more and more ads, their own widgets, and pushing down smaller sites in the search results to favor brands, Google made it very hard for smaller sites with user-generated content to grow and profit organically. And then fewer people searched for them or expected to find them.
I’m kind of enjoying the fact that newer sites and apps like Discord don’t let Google index them after Google killed off a large part of the independent Internet. We’re missing out on a lot of good content that would be put up in an alternate universe with a less greedy and shortsighted Google.
Users click more on big brands. The average internet user doesn’t want to browse hobby websites, they use the top 10 sites and never leave them. Accept it. Some of them never leave only Reddit or TikTok.
I was interested as well, but couldn't find anything specifically answering the question. According to the link below social media, email and watching videos counts for about three-quarters of most people's internet usage, which could easily mean they spend most time on just three or four sites. The two smallest categories - searching and shopping - could also just mean Google and Amazon for the most part, so it's the "Surfing content (19%)" that's ambiguous... although even then I assume that Google is a big part of that, and otherwise surfing for content is pretty much the opposite of visiting a site regularly, although I guess it probably includes news websites, which probably are a regular visit.
I'd guess that it is possible most people visit a dozen or fewer websites regularly, of which half a dozen or so are in the Top 10, and the rest include things like news/media sites that are most likely to be popular (if only nationally). Seems likely the older the person the narrower range of things they do online, apparently the most common activities for the over-60s are email and checking the weather lol.
Never thought I’d see the day, but I actually use Yandex for image search now because they basically deliver the google image search experience of a decade ago. While Google Images has had no one with Vision in charge and been tweaked, buffed and downgraded to the point of uselessness. And you can tell no one on the team actually uses it.
Google Search could honestly be dethroned if someone really wanted to. The usage at this point is habit, not because it’s still a good search engine.
I've noticed recently Google image search has gotten surprisingly terrible for some things. I use Bing image search when the Google image search lets me down. Haven't tried Yandex.
Yandex in particular is good with reverse image search is the main thing that won me over. It uses the same system Google used before they replaced it with machine learning and made it useless.
Google Image search these days seems to be "Oh that image you sent me is an Apple, here are pictures of other Apples" rather than showing me other sizes of the same image, and other images that are very similar looking.
I still weekly make the mistake to try to find something that I know to exist. Sometimes I know there are hundreds of pages on the topic. I type query after query, again and again, nothing shows up. It's like in the 90's, I tell stories about.. I forget his name and the place, when it happened but it was something like this and that. pfft!
edit: I just attempted to look up what kind of taxes on has to pay if one receives donations. I cant carve out a query that gives me anything other than tax deductible donations. The "receiving" part is completely ignored.
I agree. More generally, any business model where your users and your customers are disjoint sets will cause a misalignment of interests that generates an ever-increasing negative feedback loop. It seems like it can take a long time to play out though.
For those annyoyed with ads on Google search, Kagi [1] is worth trying. It's not free, but does not show ads and uses the Google index behind the scenes. I've been using it on desktop for few months and have no complaints.
Does it give better results than Google? I find Google's results to be pretty horrible, and worry that if Kagi is using Google's index, then I wouldn't get better results with them.
It does give better results. The filters can also be quite powerful.
The problem is paying a steep fee to proxy Google's index... It benefits you and Kagi in the short term, but strengthens the Google/Bing/Yandex index oligopoly [0] long term. Bad search results are the symptom - scarcity and opaqueness of viable indices is the root cause
> but strengthens the Google/Bing/Yandex index oligopoly
The way we look at it as that it gives Google a non-ad based source of revenue, which is a good thing for the future of the web. If one day that revenue could displace all ad-based revenue, by Google selling data to providers like Kagi, we would reverse the current state of the web.
--that coupled with inability to realize that blasting ads at people doesnt make your product look better, it makes your product a source of frustration, and a daily/hourly annoyance, your product is associated with bad, and is avoided.
the answer to the problem includes the concept of less is more.
It actively makes the product worse. E.g. from yesterday, if I search for "high tc record," I'm not looking for the set of bedsheets with the greatest number of threads.
In fairness, reading your search term, I had no idea what you were talking about so it's not entirely surprising the organic content producers don't match those terms well. Looks like the SEO folks do though.
Pretty sure at least 80% (probably 95%+) of low-quality SEO content is generated by computers, or at least by very low-skill outsourced labor in a machine-assisted way.
It then makes sense that the mature Internet can probably match any search query in a manner that leads to a site that sells something vaguely matching the query, through sheer quantity of available content.
In other words - there is enough SEO content generated and indexed that most queries will be able to be monetized by a search engine like Google.
Tc is the high temperature for superconductivity.
I noticed this about google a long time ago - they started "dumbing" down their search results. Don't blame them because that's where the money is, but it is annoying when you're searching for something and the results are full of celebrities because someone famous happened to say something like your search term.
Then again, Tc means different things in different fields. E.g. critical temperatures for all kinds of critical phase transitions. Note that the google search you posted above gives results for various phase transitions many of which are not superconducting phase transitions. I also was lost when I saw the original query.
Perhaps there are multiple groups searching with the same terms, but expecting completely different results. Google takes its best guess at which group you're in. To get better at that they have to gather more info about you personally. I'd prefer to change my search query than give lots of extra personal data to Google.
But you don't have a 20-year dossier of almost every web site he's visited; a list of 90% of his credit card purchases; an AI scrutinizing every photo he's ever been in that's been in, or in the background of, on the internet; a list of his previous searches; and a list of the places he's brought his cell phone.
Google says it spends about 0.3Wh per search, of which roughly one third is for scraping and two thirds for the search. I think you may be overestimating how much processing they could do with that budget (assuming they want to do that significant user-based tweaking, which is IMO not a given).
I don't know what you are trying to search for. A record for total compensation? Something related to Tesla coils? Turing Complete? Thread Count is a totally valid interpretation.
I have no idea what you are expecting from "high tc record"
I'm getting links about the highest temperature on earth and it is listing it in Fahrenheit so if the "tc" means "temperature celsius" it is failing for me.
All the other links are about high temperature superconductors.
Tumblr - banned 18+ content in 2018(?) in an attempt to make their site more advertiser-friendly and lost most of their userbase within a couple of months. Site was thriving before that.
We are talking about platforms where the ads themselves caused users to leave. This is not what happened with Tumblr and it is not a good faith view to say so.
I only see you attempting to constrain the topic that far, rather than just the ways that advertising based models in general can pervert a company's incentives by turning away their core user base. Sometimes it's just changing the UX until your userbase no longer wants to show up. Sometimes it's literally just banning a huge chunk of your userbase for some reason.
LinkedIn went through the same algorithm led spiral of decline.
Initially, they seemed to stop pushing out good on topic content. Instead, they rewarded and promoted Facebook style posts and made up stories probably due to metrics and A/B tests.
The good authentic posters then either stopped posting or had to feed the beast by pushing out the same crappy content in the house style which the algorithm rewards.
Almost everyone on there now is writing purely for the algorithm and not their industry peers whether they realise that or not. They are just chasing the doamine hit of likes.
The product is worse for it and I’m sure it ultimately kills the golden goose for social platforms despite a short term uplift. They become much less engaging and sticky when the social element is lost.
When LinkedIn arrived, it seemed like a good idea -- sort of a way for people who aren't great at networking to see how to expand their network. Then, like any dating site, you needed to buy credits to participate, and only "players" are willing to do that. Then, they erased any additional value brought by their thought leaders by promoting third-rate content with catchy headlines like you mentioned. Now, why would I trust them with my resume? And if I have no interest in them, likely anything else on there is either out of date or some sort of lie.
Friendster, Myspace, Orkut, Facebook.... Similar things for social. Use the network to find more people to hang out with, feel less alone. But then, again, that is not actually a business and people aren't paying for it. So, you either show generic ads (not a bad idea) or you get people to pay to get their content in front of you. But then people stop caring about what they see there and move on.
I agree with other posters -- I used to keep Facebook around because my family was there. Now I don't even see their posts when I'm on it. So, there's no value anymore.
I treat LinkedIn as I treated Facebook: a way to keep in touch with people you know or once knew, as a way to keep track of a list of these people, and as an easily findable public handle. For Facebook, it was of course less oriented towards money making.
Now LinkedIn still does the same job. As far as I’m concerned, while “thought leaders” and other spam is more prominent, dealing with egos is the price to pay for making it easier to find a job, and to be found as a candidate. It’s not materially different to IRL corporate world.
I'd also add that the optimizing the timeline for engagement (including negative engagement & arguments) wound up making me quit the platform entirely. There were no positive interactions, just deeply unpleasant arguments with people spewing vitriol.
Like ... why would I stay on a platform that was so deeply unpleasant to interact with all the time?
suppose you're in the meeting room and you're the product guy and you'll make this case. Indeed let us not optimize for the presently advantageous local maxima lest it come back to swallow and end us. How is this communicated? And what does it take to convincingly persuade people, or investors of this -- to even bear through a temporary downward period with the understanding that fundamentals must be respected?
>And what does it take to convincingly persuade people, or investors of this
As long as their top priority is next quarter's result, you can't. You may get small wins here and there, but nothing fundamental will change unless they have a long term vision. When I say "long term", I'm talking 10 - 20 years, not 1 year.
They didn’t build the iPhone or even like Google Maps and Google Images by walking into the meeting room with data. They instead said “what if maps were as easy as we’ve made search”
Of course it becomes a problem when no one even to the CEO level has lost vision. Never put the numbers guy in charge.
From my comprehension of Mad Men? You hire a PR firm to convince those investors and decision people, to sell them an idea instead of KPIs.
It can go both ways, of course: Mad Men is full of examples of stupid business ideas that went spectacularly wrong. And while it wasn’t the focus of Mad Men, the real world is full of examples of business that went the metrics way and were successful.
But IMHO, if I was the CEO of a social media company, I sure wouldn’t focus on A/B testing and engagement metrics.
It's not just the ad revenue or suggested content. FB's most sticky offerings are groups and pages. Well, FB actively hides posts in these to their own followers. You end up having to pay just sponsor messages to your own group.
I quit FB a few years ago when I realized the suggested posts were just way more interesting than anything my friends posted. Idk what that says about their algorithm. Or my friends.
Almost like building a business around “engagement” is a model that leads to a death spiral. Social media sites are basically giant MLM schemes that sell the idea that become real companies before the scheme runs out.
I was going to disagree with the GP for the same reason you pointed out. I liked all the page and group posts because they were funny and not someone I haven't spoken to eight years planning their wedding in public. There aren't as many organic posts from my friends to share anymore.
