For one thing, lemmy.ml is categorized as "Games > Games - Other (In United States)" which made me scratch my head to the point of hurting my scalp. The rest are uncategorized (which is better than being miscategorized, imo).
¹ -- reddit.com is included as a point of comparison
² -- lemmy.world didn't exist yet in May 2023
We can see that the larger instances are already performing well in comparison to reddit when it comes to "interaction" statistics. It's a surprise, however that kbin.social trounces everyone else it was compared to--even comparing favorably with lemmy.world in visit numbers. In comparison, lemmy.ml performed quite badly especially in bounce rate and average visit duration. Someone who's better equipped than me in analyzing these figures can perhaps do a better anaylsis, but from what I can see, we're not doing that bad here.
I've also added lemm.ee into the mix just for good measure (and perhaps as a proxy for smaller-ish instances), and it's doing quite good as well.
Something to keep in mind to contextualize the interaction statistics: the density of contributors to lurkers in these numbers will be drastically higher here due to the greater barrier to entry as well as the average user type of the migration. It should be expected that the average user in a niche/early-adopter community will be more active.
Thanks! What do you make of the unusually high bounce rate (and low average visit duration) for lemmy.ml though? That has been a head-scratcher for me. Did it not gain people from the migration as well (which makes for better interaction)?
Could it also be due to load/server issues? When I have trouble with page loads on .world I have a tendency to bounce and have a low average visit time.
I'm actually quite the opposite, I wait until a page loads. Between the two of us, it evens out, lol!
But yeah, I have a pet theory as to why lemmy.ml has a high bounce rate and low average visit duration: people migrating from Reddit dismiss lemmy.ml for whatever reason and check out other instances instead. That's probably also why it didn't gain as many users as other instances (comparatively speaking).
I think lemmy.ml restricted registration and had a PSA telling people to join other instances. They didn't gain that many users because they didn't want to, probably because their infra couldn't handle it (hence also the bouncing).
What do you make of the unusually high bounce rate (and low average visit duration) for lemmy.ml though? That has been a head-scratcher for me.
It shouldn't be a head scratcher. lemmy.ml was running like ass for much of it, with some periods of being completely down. Poor performance will drive up your bounce rate massively.
Okay, but 52% of traffic on reddit is coming from Youtube??? Those channels that just compile reddit posts must really be helping Reddit. Lets make some Lemmy reading channels
As far as I understood that stat, it's traffic coming from "social media sites" that's counted there. What their definition of a "social media site", I've got no idea--especially not since they include Youtube in that designation.
Added lemmy.ca into the table as well, because, damn, what a ride! I suppose given from the explanation I was given regarding lemmy.ml's figures, it wasn't able to cope well with the flood of incoming users, I suppose?
We've actually handled it very well as while it's a major instance we are on the smaller side we also moved to a dedicated server at a good point right before it reached the limits of the VPS it was on.
So from what I can tell the visits to the main lemmy instances are only ~.5% compared to total visits to reddit (1.7 billion vs 9 million). Smaller instances might bump that up to .7% or so. Reddit's 3% drop in visits might be mostly due to people using the site less, while transfers to lemmy account for only a fraction of that. Still cool, though, hopefully we can keep the momentum going over the next couple months.
I really wish someone can compile the figures for all Lemmy instances (and Kbin) so that we can make such a comparison. But yeah, other alternatives also popped up such as Tildes, Squabble, Raddle. I suppose that can also make up for part of the missing 3%. Of course, there's also a sizeable number that have simply stopped using Reddit (be it in protest or other reasons--such as seasonal variability, as in people are actually touching grass!).
I actually forgot to include Tildes, Squabble and Raddle into the table, but then again, I was only trying to compare what we gained here against what Reddit has lost. I hope that this momentum (loss for Reddit, gain for the Threadiverse) keeps going. But more importantly, that we keep the engagement here at a high level in both quantity and quality.