I'm not inherently opposed to a smarter order, but I feel like it (at least in the case of Facebook and Twitter so far) breaks a fundamental part of the use case for me. I check in every so often, and want to scroll back to the last post I saw, so I can know I saw everything. When it is sorted non-chronologically, this is impossible. I see the same thing multiple times because it's sorted near the top, and I miss some stuff completely because it gets sorted to the bottom.
Also, I'm not sure what signals they're using to rank posts. I could imagine using likes and comments, but I don't think that would accurately identify the posts I most want to see. For example, I follow a couple of local restaurants who post specials and similar things on Instagram. I never "like" or comment on those; it seems pointless (I use those actions to communicate with my friends on their posts). If you use that signal, those posts which I always want to see will get pushed to the bottom. I'm not sure what other data they have to use.
> I check in every so often, and want to scroll back to the last post I saw, so I can know I saw everything.
I am on the same boat. I'm not sure how they don't see the contradiction in pushing mobile more and more at the same time ignoring the problem that their default newsfeed is stale. They still have the option to switch between 'Most Recent' and 'Top Stories', but it defaults to 'Top Stories' which wasn't the case a while back -- it used to default to whatever I last chose.
Now I don't check Facebook that often mainly because there is nothing new, and its always the same handful of people I am forced to engage with.
>Also, I'm not sure what signals they're using to rank posts.
In the past, FB has run "experiments" which determine the order you see updates based on the perceived content of the post, in order to manipulate users' behaviors. Without access to the ranking criteria used on a given feed, it would be prudent to assume that similar "experiments" (or worse) could be determining what you see, and when, on any FB service.
Why do all of these social media companies ruin feeds like that? It would be fine if there was a seperate tab for 'best'.. I want to see things as they're posted. No fancy algorithms.
It increases the kind of engagement that results in ad clicks. There's no way to make an advert or sponsored celebrity 'post' travel through time, but you can make it 'better'.
This may not be the kind of experience you want, but after all you aren't the customer here. You may not even be the product - you may just be a side cost of getting ads in front of the lowest common denominator consumer.
Because the majority of social media users subscribe to more people/companies posting more things than the user could ever possibly read. Rather than users being overwhelmed with noise, getting frustrated, and possibly logging in less, social media companies come up with a way to show them enough content to keep them engaged and happy (and visiting the site/using the app) as long as possible.
Here is some speculation on my part, perhaps totally incorrect.
The three social platforms with non-chronological timelines now: Facebook, Twitter, Instagram, may all be seeing diminishing engagement from critical demographics (16-21.) Users log in less, so by "rewarding" them with the stickiest post when the do happen to log in, it appears to increase the retention of those cohorts.
Power users are following/friends with a lot of other users. If 1,000 - 2,000 people are followed there is a lot of noise. This also presents opportunities to completely bury the less sticky content, which also results in increasing ad revenue from publishers who got addicted to the free traffic (with "sticky" ads being rewarded with cheaper clickthroughs.)
The chronology complaint is interesting because reddit & hacker news have a similar sorting algorithm (voting, but with other complexities thrown in.) You log in, look, and log off. A compulsive user may check hourly.
There are definitely problems with dynamic sorting. Noisy producers may become more noisier because less of their stuff is being shown. Other users may not bother posting at all because they don't expect anyone to see what they posted. That audience is going to feel comfortable if another platform offers them a place they feel they are being heard.
Whatever the numbers show right now, the longer term retention cohorts have yet to be seen by anyone.
Despite monster ad revenue I am not convinced that Facebook has made a "slam dunk" on social media and will be the last man standing.
Except the real goal likely isn't usability, it's to show promoted content in with the real good content, ala Facebook, Twitter, Imgur, Reddit, Hacker News, the list goes on really.
Because now Instagram can charge brands to reach their followers.
Facebook pioneered this business model and it has worked very well for them financially. Even if you have millions of followers to your Facebook page, you can only reach a small percentage of them without paying Facebook.
Reddit is Fun has a nice solution (though it's just an interface to reddit's core functionality): at the top of the screen you can select which ordering you want (time order, most points per time period, hotness).
I think Reddit is not quite analogous to the other sites listed. Excepts for posts being deleted or stickied, all content is sorted democratically on a timer. You have the option (by default) to additionally sort by newest or newest post that are also popular with users. I'm sure other Reddit mobile clients have this functionality as well.
I believe the algorithmic ordering is intended to optimize other variables such as "time on site," because if you stay longer they can show you more ads. It seems like the conspiracy stuff would be much harder to actually do, like manipulating the content to influence you to click on more ads.
Maybe "conspiracy" is too strong a word, but I have no doubt that the key motivation behind all these algorithmic changes is to boost ad metrics rather than to better serve the users.
While I agree that there is a conspiracy element to it, research has been done into this area before by Facebook. Whether or not it's in use is a whole different can of worms.
I knew something was up when they removed the time since the item was posted. This change will just make me unfollow the bigger users that I follow and check them manually.
I'm curious what data all of these companies are using to determine that it is better to identify posts by algorithm as "the best" rather than simply showing everything in chronological order.
I thought Twitter originally said several weeks ago that they would introduce the new feed and then let the user choose to use it or not. So when it first started I thought I opted-out. Just the other day on twitter.com I was shown a list of "best posts" that I needed to close, and then another click to say I didn't like seeing that section.
Even if somehow finding which posts/tweets are "the best" is helpful, I don't doubt that some users prefer chronological order.
It's really disappointing, but then I guess it comes down to the issue of using a free service- in the end they get to choose how it's run and not me.
Keep in mind that Instagram is owned by Facebook. So expect the same methods Facebook uses to rank content, and consequently expect to see the same "tricks" that marketers use to exploit EdgeRank. (e.g. emotionally-charged linkbait, content-theft-and-reposting)
I can totally see genuine reasons for that. Feed is a huge part of the product, and you want to control that (which means controlling UX). You don't want to let things go out of hand. Some of us feel like, well, we can control it ourselves by unfriending/unfollowing/unliking things we don't like. That's true. But we shouldn't forget that we can't apply this assumption to the whole active userbase of Facebook/Instagram. Many many of their users won't bother, or aren't even aware of it.
It's kind of odd seeing all frustrated comments about how it's just about ads and etc. Well, duh, they are a business, not a charity.
The only problem with it is if they disable an ability to switch back. But, again, I can see reasons for that. It might be that they ran some experiments and some that people switch to chronological and then they happen to be less engaged over the time because their chronological, raw feed happens to suck big time.
> we can control it ourselves by unfriending/unfollowing/unliking things we don't like... But we can't apply this assumption to the whole active userbase of Facebook/Instagram.
Perhaps that is the UX challenge companies like these should be pursuing? One of the unfortunate side effects of the valley's growth fetish is that features like "make it easier to break connections" get sharply deprioritized.
I naively dream of a world where products teach the masses "it's easy to settle only for quality, and it's okay to stiffarm the other noise that clamors for your attention"
They too have noticed that having a feed of n% stale posts and 100-n% new content is more likely to addict you. A skinner box is a powerful device, and you don't get addicted nearly as easily if you get the reward every time.
I really hope they have an option to switch it back to chronological. I tend to check Instagram at the end of the day and just scroll back until I see things from the previous day, then I know I've seen everything. I don't follow too many people so it only takes a few minutes.
Arguably this is Facebook's most important secret sauce, so it makes sense to try it with IG. I'll be curious if it sticks or if they revert back to a chronological feed.
Also, I'm not sure what signals they're using to rank posts. I could imagine using likes and comments, but I don't think that would accurately identify the posts I most want to see. For example, I follow a couple of local restaurants who post specials and similar things on Instagram. I never "like" or comment on those; it seems pointless (I use those actions to communicate with my friends on their posts). If you use that signal, those posts which I always want to see will get pushed to the bottom. I'm not sure what other data they have to use.