cross-posted from: https://lemmy.ca/post/12971023
Hi folks, out of pure curiosity, I was poking some graphs.
It’s been about half a year since the big API protest, so I was curious to see what Lemmy’s crtitical mass looks like, what the staying power is, etc. Screenshots taken from https://the-federation.info/platform/73 on 2024-01-09. I’m posting screenshots because they’re a snapshot in time, and because that stats server is very slow.
Because I’m posting on lemmy.ca, I’ll post quite a few related to this instance, but it’s probably more widely applicable and you can get graphs from your instance too. I’ll also post some lemmy.world and lemmy.ml graphs, since they make interesting points of comparison – biggest server, and original server.
First, lemmy-wide total users count, where this is a rolling one month window. If a user was online within the month, they count here.
First observation – there’s some jagged edges in the graph due to things popping in and out of the federation. So it’s probably more useful to look at single servers. Lemmy.world came online pretty much coincidentally with the API protest and had open registration, so it makes a good data point. You can see the surge of users, then the plateau of the people who stuck around:
Lemmy.ml below has a similar curve, plus some sort of data artefact.
As does lemmy.ca, below:
I suspect the data artifact is related to the transition from 0.18 to 0.19 and something changed in the way active users was counted in between. Lemmy.world is still running 0.18.5.
Notes: The difference between the peak and the plateau is higher on lemmy.world and lemmy.ml – I suspect this is because they were more popular places to sign up during the protest. Whereas lemmy.ca has retained more users, as a percentage. Still, the total number of active users on each server is quite low.
In the same order (total, lemmy.world, lemmy.ml, lemmy.ca), total posts. The slope of this line represents post rate. Steeper line is better. Flat line means dead instance.
And comments. I wish there was a comments to posts ratio, which would be some indication of engagement levels. But you can sort of work it out.
Anyway, looks like post rate has decreased slightly since the initial bump, but are still looking good. But the comment rate hasn’t flattened as much. So the users that were retained seem to be more engaged than the users from the initial bump. I think this is a good thing for the health of lemmy. Likewise, the growth in supported apps, improvements to the software (Scaled sort in 0.19 is night-and-day better than anything prior!), and others will allow lemmy to not only survive, but be ready for whatever influx happens next.
I want to send a special shout out to all the admins, particularly on my home instance of lemmy.ca, and the coders who keep improving things. Thanks for giving us all a home!
You must log in or register to comment.
I suspect the data artifact is related to the transition from 0.18 to 0.19 and something changed in the way active users was counted in between.
My understanding is they started counting votes as activity.
The difference between the peak and the plateau is higher on lemmy.world and lemmy.ml – I suspect this is because they were more popular places to sign up during the protest.
Makes sense that people who care enough to look into smaller instances would be more committed.
Threw this analysis together because the-federation.info doesn’t provide derivatives:
⚠️ Note that the y-axes are not to the same scale! ⚠️
It looks to me that activity is diminishing slowly. There were peaks and dips (I assume by spammers?) but every server saw a huge spike when Reddit decided to kill third party apps, followed by a slow collapse afterwards. The ratio of Reddit-joiners to slow-leavers seems to be very similar.
Activity is huge compared to what it was before the Reddit event (especially in terms of comments) but it seems that Lemmy is not maintaining its popularity.
Even beehaw, which is a more opinionated, community focused server, seems to suffer from the same trend. Lemmynsfw seems to be keeping pace at the moment, though I do wonder if that’s because of repost bots.
⚠️ Note that the y-axes are not to the same scale! ⚠️
What about Lemmy’s competitor, kbin (and I suppose mbin)?
⚠️ Note that the y-axes are not to the same scale! ⚠️
I have my questions about the data (large amounts of posts disappearing is pretty weird) but if this data is correct, I don’t think the Threadiverse is viable for the long term without another influx of new users. These platforms are content-driven, and only a tiny amount of users post any interesting content compared to the upvoters and commenters, and the more interaction drops, the more people will want to leave.
Excellent work. Based on this, I think the communities on kbin and beehaw are in trouble. But only if you consider posts to their communities to be the key metric. If their users are still participating in the larger fediverse, then maybe it is fine.
Kbin has had a lot of stability issues. And beehaw defederated from some major parts of the network specifically because they wanted to avoid the influx of users from lemmy.world and others. So why would I, as a lemmy.ca user, post to a kbin or beehaw community and limit the potential discussion.
I sort of wish the-federation.info would produce derivatives as you did – far easier to interpret than slope changes visually. Probably could use a 28 day moving window average or something to smooth it so it isn’t as noisy, but that would disguise interesting events.
I find the plateau quite puzzling (lemmy.world, but the total looks very similar):
There was quite a steep increase, and then it suddenly stopped.
I would rather expect it to slow down, than to stop that abruptly.
We’re looking at a fairly large group of people making a decision to create an account on Lemmy. There are plenty of reasons to expect it to be fuzzy. Even if they all responded to one particular event in time, some would have done so immediately, others the next day, few more even later.
It’s because it is a one month rolling window.
Lemmy has better user retention than Diablo IV confirmed
and the diablo 4 community is dead lmao
There are dozens of us! DOZENS!
Looking at the rest of the data (especially the sustained linear increase in posts across the whole network), I’m increasingly skeptical that the drop in “active users” is really all that meaningful. Speaking for myself, when the big migration happened I created three accounts on different instances, but I’ve found myself only consistently using one of them. If a significant percentage of the rest of you did similar, that means there could’ve been what looks like a huge drop in the number of “active users” even though the number of actual people using the platform remained the same!
This is true in my case too.
Yes, i made four, because when i joined Lemmy, everyone seemed to urge new users to spread across the fediverse. So, i did. But over time, i did away with two accounts and am contemplating ditching another one.
I think posts is being inflated with bots copying reddit, my subscribed feed has noticeably slowed and even trying to find more communities to get more posts hasn’t been a huge help.
Thanks for this. Unlike on Reddit I feel much better posting here knowing I’m not helping some company make more money.
Gives me the old internet vibes I’ve come to crave
I wasn’t conscious of it until I had stopped, but on Reddit I was censoring myself to avoid my comments getting deleted.
You were probably being racist. There’s only 2 things I can’t stand: intolerance and the Dutch.
Here there’s a different kind of self-censorship. Anything you do (including your upvotes) gets propagated out using the ActivityPub protocol to all instances that are subscribed to that community. So in theory, admins on different instances can tell what you’re upvoting. A bad acting admin could stalk you here in a way that a mod never could on reddit - because mods couldn’t look at your upvote history.
The good news is that they cannot delete or modify your content on other instances (only their own), so they’ll never pull a spez and edit someone else’s comment globally. And, bad acting admins will simply get defederated, so we should be self-policing (in theory).
Do we have an expectation of privacy for our upvotes, or is that generally supposed to be public information? I like to think my comments and upvotes match up pretty well.
I’m not sure how that expectation lines up. On reddit they were private (except to yourself and the admins). Because there’s no warning anywhere on lemmy that they are public (or at least semi-public), I’m going to presume that most reddit refugees believe it works like reddit.
Fair. I’ve heard kbin allows viewing, so there are federated sites which can see them without needing to be an admin or run an instance.
Thanks for sharing this, this is really interesting.
My hope is that when Reddit announces their IPO, more people will start talking about wishing for alternatives. I hope this motivates a few people who checked it out and left and lots of new people to take a first look, and when they do I hope they find an already active community that produces enough content to retain more people and generate more content.
Those results might be slightly skewed by alternate accounts. When I first joined during the Reddit Exodus I created this account on lemmy.world, but the instance suffered a LOT of downtime for the first month or so, so I created a few other accounts on lemmy.ml and sh.itjust.works so I could still browse while lemmy.world was down.
After the instance stabilized I pretty much stopped using the other accounts, so I, personally, am 2 of the people who “left” by leaving the other accounts inactive.Same, it might have been this
Same. It wasn’t clear how to choose an instance, so I ended up creating accounts in three different places and posting a couple times before settling on this account. I haven’t used the other accounts in months, so they’re part of that surge.
When the reddit API protests occurred, lemmy wasn’t really ready for the influx either. Historically, when a social network dies, it’s some combination of a protest and there being a pre-existing landing place that is ready to receive the influx. In the case of digg dying, that was reddit ready and waiting.
But lemmy had so many rough edges and was almost entirely unknown at the time of the reddit protest – bugs, missing features, no apps… For most reddit users, even with the 3rd party shutdown, moving to lemmy at the time was objectively worse.
You’re right though – the next time something happens, lemmy is now established, the apps exist, many of the bugs and missing features have been dealt with, etc.
Totally remember the lack of apps. Initially, I just had to use Lemmy through a mobile browser. Lots of devs were working hard to publish their apps, and after a few months we had lots of options. That was just amazing how quickly it happened.
BTW shout out to Bean, my favorite Lemmy client. It’s not perfect, so in some cases I still use Voyager to fill in the gaps, so bonus points for Voyager too.
Another important detail is that Digg v4 pissed off most of the userbase, so the impact was pretty much immediate. Reddit APIcalypse pissed off only power users instead; the impact will only come off later (sadly likely past IPO).
Well, there was also the DeCSS key censorship debacle…
09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0
Could you explain what this was? I couldn’t find anything about it while searching.
https://en.wikipedia.org/wiki/AACS_encryption_key_controversy
Good overview. But it was a huge on-site protest, where a large percentage of all comments and posts contained the above string in some form.
Awesome work! Thanks for sharing!
The line is going up. My sources tell me this is good.
That graph gets worse the longer I look at it.
My sympathies to anyone who has to use reddit because their niche community either doesn’t have enough activity or doesn’t exist at all.
I’m more of a casual user who’s just here for the news and memes, so fortunately I don’t have that problem.
sadly even things like memes were better on Reddit, maybe I could do something to change that
In that case, I recommend !tenforward@lemmy.world ;)
Subbed! Thanks!
Okay, more serious answer. You look like you’re on kbin, so I don’t know if this applies – nevertheless.
On Lemmy 0.19, the Scaled sort algorithm is such a good improvement over (Hot/All/Top/…) that existed prior to 0.19. It’s basically a Hot sort, but it’s weighted by community size. So if you’re subscribed to a small community, that gets one post a week, it’s still likely to end up in your feed. I’ve noticed a huge improvement when switching to it as my default sort – suddenly that weird music community I subbed to, but never noticed any of the posts – is in my feed. Etc.
Lemmy.world is still on 0.18, but when they upgrade (I have no information on that process) I suspect that people should be switching to it as their default sort for a better experience if they’re into niche topics.
Hey, cool! I didn’t know about Scaled sorting. Thanks again!
So you’re saying I should try Hot sort now? Hmm. Let’s see.
My sympathies to anyone who has to use reddit because their niche community either doesn’t have enough activity or doesn’t exist at all.
I’m more of a casual user who’s just here for the news and memes, so fortunately I don’t have that problem.
Great post, thanks!
Meanwhile if you’re curious how reddit has been doing, just look at this chart
chart doesn’t seem to show up on lemmy. Not sure what I did wrong there, but it’s the first time I tried an image in a comment. I guess this way works better.
Sweet post. To me this looks like the makings of a sustainable community and I remain pretty optimistic. Curious what the numbers for Kbin would look like.
Here’s the kbin graphs directly on the site that generates the charts. I’m unfamiliar with kbin as a backend, so I don’t know how to interpret any of their numbers. But you can play: https://the-federation.info/platform/184
I don’t even use lemmy and that was interesting to read.
