The difference between a closed, algorithmically-curated silo like Twitter and the fediverse.

Same post.

Twitter: ~43K “followers”, 8 boosts, 58 likes
Fediverse: ~8.43K followers, 57 boosts, 106 likes

Your thousands of “followers” on Twitter mean nothing because the algorithm (i.e., Twitter, Inc.) decides who gets to hear you.

@aral one feeds you constant micro-doses of endorphin by constantly showing you how well you "did", the other knows it is unhealthy
@aral Twitter's algorithm makes some people with more or less exposure. Fediverse show everything, sort by date

Indeed, @aral, also the fediverse is less saturated so people are generally more interactive, perhaps due to the bystander effect?

@aral Exactly my experience, back when I used both.

@aral Or it could be because on the fediverse there are more people who are critical of big tech companies than on twitter. Aka sampling bias.


For this post, perhaps. But I've noticed higher engagement here, too, even with a lot fewer followers.

@attilakinali @aral This appears more likely to cause such a result than display and sorting algorithms, though the latter almost certainly plays a part.

In any case, only accurate samples can help reveal the causes of any such differences.

@attilakinali You realise these are still my followers we’re talking about, right?

@aral Are you assuming that all your followers are the same? That none of them has a different personality? A different opinion?

@attilakinali Not at all. But I am implying that following me is a sampling bias that’s stronger and more relevant than that of the generic sampling bias difference between Twitter and the fediverse’s general audience.

@aral Look man. You write something that elicits a response from a specific type of people. You look at the reaction you get on different platforms from that one single post? Sample of one statistics? Not accounting for any biases at all? Really?
That's the kind of pseudo scientific behavior I detest. It's the kind of thing that deceives people. Or rather, it's how people deceive themselves. You think you understood something. But what you really did was feed a GIGO system. And lied to yourself.

@attilakinali @aral I think it's more complicated. For example, my RMS-style Facebook/Meta meme got a lot of boosts/stars at a scale disproportionate to my account age and follower count, mostly due to having second-degree connection with a power user. So it will also depend on the influence of each follower on the respective platform. Note that this does not rule out algorithm manipulation, but too many variables makes determining the exact cause difficult.

@aral I agree, but I do wonder how much of this is because of the differences in audience. I would expect people on the fediverse are more likely to agree with the message.

@mike You realise that you’d have to be a masochist to follow me on Twitter or anywhere else if you didn’t agree with me, right?

@aral I would guess that the enthusiasm level might be a bit different, at least.

@aral @mike Theories:
I follow people I don't agree with, mostly because I don't agree with everyone about something. Maybe I love most of what you do but just put up with your "big tech" opinions. Twitter would know that and filter those out.

I limit how many people I follow on the Fedi so I don't have an overwhelming amount to read. If I were on Twitter, I could follow more people because the algorithm sorts for me. So I'm more likely to interact on the Fedi.

So with all sincerity then, why are you still using Twitter? Don't you think that your presence there somehow validates the platform?

I don't mean that as some random bitchy (twitter style) comeback. Really. isn't it time to "vote with our feet"?
Or if not now, when?

@keith I’m forwarding from here. I still have people there that see my tweets who are not here (members of European Parliament, etc.) When I feel it’s no longer useful, I’ll close it down. Does that make sense?

That is to say.. I do not support in anyway the idea that mastodon is a twitter "replacement".

@aral Lovely example, and ❤️ the sentiment too 👍🏻😊
@aral and it has adsolutely nothing to do with the anti-bigtech mood difference here and there 😅

@albi Well, let’s just say if you’re pro-Big Tech and you follow me, you’re likely a masochist ;)

@aral @albi :bunthink: are you assuming ignoring that your followers also have followers, who are more likely to boost your posts here than on twitter?

@xarvos @albi Nope. It wasn’t a PhD thesis, just an observation.

@aral I have made were similar experience.
#Twitter is very much in favour of big-players among users.
So for small user like me #Fediverse is by far the better choice.

@aral the Fediverse is also great because you can create dedicated communities while retaining the ability to connect with others on the outside of that community.

It's the modern equivalent of old BBS systems with Fidonet.

@aral can I use these screenshots for a presentation of the fediverse?

Sono uscito da twitter proprio perché annoiato dall'algoritmo

@aral plus probably more meaningful comments than just proper bullshit. also what's already known is the thing about hate speech. the problem is that twitter puts hate speech comments at the very top, because they generate more ad revenue for twitter. this can be clearly seen when using non-algorithmic twitter frontends like nitter.

@Goutte That's the core of the argument . @aral has fewer followers than .

Naively one could assume that the base rate of reactions on the Twitter account would be five times that of the Fedi account.

@aral Well, there is also a bias with the content of the message and the audience of each network/platform.

@aral Same text, but not same post obviously, since those are 2 different accounts on 2 different portals)

@aral Wait, you were (still are?) using Apple iPhone to post those.

@poputchik Wanna buy me a new phone? (Kidding, please don’t. No, I’m not going to throw away my phone overnight but I am looking for alternatives.)

@aral @poputchik The Pixel 6a will receive 5 years of support, which is relatively long for an Android device.

Kind of pathetic that the bar is that low; if google just released one phone every 2 years they'd be able to easily support the model for over a decade.
@Aral Balkan this post is 22 days old. Can we now get an update on the numbers?
Sign in to participate in the conversation
Aral’s Mastodon

This is my personal Mastodon.