Hacker News Reader

Comment by mjr00 | Hacker News Reader

The methodology is "(likes + shares + comments)/followers", and the two platforms that allegedly provide the highest engagement are the ones with an order of magnitude fewer followers.

A more likely explanation is that fewer followers means higher engagement, which totally makes sense; for a larger social network account with 100k+ followers, many of those will be bots and people who follow anyone and everyone without really caring. A smaller account will only have followers who care.

I don't think there's anything specific about Mastodon here other than it being a smaller social network so you naturally have far fewer followers.

In either case, this is a single account talking about a single post, and shouldn't be used to generalize different levels of engagement across social networks.

View on HN

> I don't think there's anything specific about Mastodon here

These numbers tell me that:

1. Mastodon users are engaged, with an active mind.

2. The grifters/leeches/scammers and their bot legions haven't invaded yet.

> The grifters/leeches/scammers and their bot legions haven't invaded yet.

Unfortunately, they have already started. In early May several instances limited mastodon.social due to a large amount of cryptocurrency spam.

https://mastodon.social/@Gargron/110310670042459716

I haven’t followed the case closely so can’t provide much insight, but a web search might return more information.

I'm not saying anyone was wrong to limit mastodon.social, but for comparison: since Twitter was bought, I have been periodically checking the main feed Twitter shows an non-logged-in visitor, maybe a couple times a week. The "large amount of cryptocurrency spam" my instance saw from mastodon.social was maybe ~15% of what that couple-times-a-week visit showed on Twitter (to say nothing about my old account's poor DMs). Sure, anecdotal as heck, but I'd unscientifically class Mastodon's current spam problem as approximately equivalent to the level of spam one gets as an Instagram user being tagged into posts for sunglasses or diet pills etc.

Actually, it’s even simpler than that. Mathematically, if you are in a network with 100 people, the platform can show you a higher proportion of posts before you hit that Dunbar number or limit of how many posts are going to read. If the same network has 1 million people, and you have thousands of connections there, then obviously, each person’s posts are seen less and there is more filtering happening. The real question is, how much advertising, memes, cat videos in general garbage the network pushes into your feed.

Incidentally this is also the mundane reason behind the Women in Tech controversy. Women in tech reported more asshole behavior by men because the probability of encountering a man was higher, than in other industries, and not because tech has more assholes (if anything I have found tech people to be nicer and meeker in general than others, like comics or lawyers). If the ratio was more 50/50 then this phenomenon would go away. (Also, tech workers get a lot of money and prestige which is why it makes the news more than, say women in plumbing or carpentry.)

I don't think an experiment with a single post on a single topic with accounts on vastly different scales can tell you either of those things.

There are a few holes glaring holes in this methodology but just one is that she has 8.7k followers in Mastodon since joining in late 2022, while she has 241k followers on twitter since 2009. It's not surprising that the users that followed her in the last 5 months are more engaged than the ones that followed her in the previous 14 years. Are the people that just aren't interested in her content or that just stopped using twitter after a 14 year span necessarily grifters/leeches/spammers?

I don't know about 1 but simply due to scale 2 is probably true, however this experiment isn't saying much.

Anecdotally, I've noticed a much higher engagement % on Mastodon compared to Twitter with my own posts as well.

On Twitter, with over 8K followers, I get very good engagement in general (imo). But on Mastodon, with only 1.7K followers, I'll get engagement of almost 50% that of Twitter whenever I duplicate posts. If I had 8K followers on Maston and if those numbers continued the same trend I'd probably drop Twitter altogether.

Off-topic I also find the folks that comment on my posts on Mastodon to be friendly by default. It's a very refreshing vibe.

For a much smaller audience, I'm seeing a very active user base for the topics at hand. It also helps that I tag my posts in Mastodon (like people do on Instagram) to help increase reach a little bit.

Depending on your social circle, it may be the case that a lot of twitter followers have essentially gone dead. I still have a Twitter account (I don’t want anyone taking the username; the account is over 16 years old, and, who knows, Twitter may sort itself out one day) but no longer tweet; if I look at that account, most of the people I followed also appear to be no longer tweeting much (though, who knows, maybe the ‘Following’ tab is just broken).

Go on Youtube or Twitter and you'll see that small creators always have a much higher engagement ratio than larger ones. It's quite natural that a small community is more involved than a large one.

> The grifters/leeches/scammers and their bot legions haven't invaded yet.

They are, unfortunately, on their way; just had my first crypto spam on Mastodon the other day.

Another fairly reasonable hypothesis is the difference in account age. If you've been on Twitter for a decade, a lot of the accounts following you are just dead: people who no longer use the platform, who forgot their passwords, etc. The followers you have on Mastodon are more likely to be engaged because the vast majority joined in the past 12 months or so.

I think this phenomenon happens to all early social services. A large social network might have the normal 90/9/1 rule apply, but in the early adoption space your userbase skews toward the 9/1 portion. It will be interesting to watch if it regresses to the mean as it grows.

I don't buy the bot explanation, even if 10% of followers are bots it wouldn't change the calculation.

The more likely explanation is feed manipulation by as demonstrated most prominently by FB where "following" is barely more than a suggestion to the algorithm.

This is just a single user with a single post so you can't draw any useful conclusions,

but... The main difference with Mastodon here is there is no fancy algorithm. It's just a chronological timeline.

So as a user, you quickly unfollow people who waste your time with uninteresting posts. Whereas on algorithm-driven services (literally every other social network) the algorithm will suppress posts of people you don't engage with, but leave you showing as following them (even though you don't see their posts)

So with mastodon no matter how big the network gets, you'll always have fewer followers (as bored people unfollow you) but they will be more engaged (as your posts are never hidden)

Also note that this is observation by one person from one post. I was hoping it was a larger scale study of sorts that covers users who did not have as large a follower base to begin with.

This one is easily explained by the fact that Mastodon is young and people who migrated are the most active social media users.

Reply on news.ycombinator.com