Eugen Rochko came up with the idea for and built Mastodon some six years ago during another one of Twitter’s dips. A developer who had already been interested in and was working with open source software, he got the idea for Mastodon from a federated version of a forum he’d built in high school.
That project was called Zeon Federated, and it’s no longer active. While developing that, he also built and sold a platform to manage escrow for artists around commissions.
Mastodon’s success has somewhat taken its creator by surprise. Rochko didn’t jump into this project as a power user of social media nor is he prone to sharing much about himself. When we spoke, he dialed into our video chat from an undisclosed location. He’s never even used Instagram. If growth hackers look at building audience or revenue as an end in itself, Rochko seems to be the opposite when it comes to development.
This week we spoke with Rochko about the early days of Mastodon, its recent surge in users and how advertising may factor in its future.
You’ve probably seen significant growth in the last six weeks or so. Has the rate of growth maintained pace, increased or tailed off since the first days of the handover to Elon Musk? How many users and servers do you have now?
If you look at it on the graph, we had a huge spike around the news of Elon Musk buying Twitter. And there was another spike when Musk fired most of the employees at Twitter. It’s trailed off now, but the rate is way higher than it was before October. We now have 2.5 million monthly active users across Mastodon, across 8,600 servers.
We don’t chart the growth rate, but right now, app downloads on iOS and Android are about 4,000 each per day. The highest spike we saw was when Musk fired employees — we had 149,000 downloads on Android and 235,000 on iOS. Over the last 90 days, the iOS app has had 1.8 million downloads. Android provides different figures, but in October, the installed audience for the Android app was 53,000 devices. Now, it is 907,000 devices.
I can’t give you much on whether mobile is more popular than desktop: I don’t track it. We haven’t built dashboards for that.
“Moderation work is not automation-friendly. The simple cases are so simple that even if it’s a person doing it, it just takes a couple of seconds to do it. And when it’s complicated, then no automation can help. It requires a human to read into the context of the situation and to make the call.” Eugen Rochko
You say “we,” but how many people do you have at Mastodon?
I’m the only full-time employee, and the rest — five people — are contractors at the moment. I’m looking to expand the full-time team and have been working on some job listings. It’s kind of a slow process; I wish I could do it a lot faster. But it’s a new frontier for a company that has been a one-person venture for six years. It has been fine so far, but now we need more people.
Is Patreon the only vehicle you’ve been using to fund it so far?
Patreon is the main one. We built a custom sponsorship platform as well for when a business wants to sponsor us to save on Patreon fees. We also got a public grant this year from the European Commission to finance some of the work on features. But mainly, it’s Patreon.
So the bulk of it is coming from around 8,500 backers on there…
Yeah, that brings in $31,000 per month. That number has risen dramatically over the past month — it was only $7,000 last month. That’s the only reason we can even think about getting new employees.
This is the kind of scary part of running a nonprofit based on donations. I’m responsible for myself if the donations dry up, but if you hire other people and the donations stop, suddenly you’re responsible for other people’s livelihoods. That’s been the stopper for getting other people as employees before now.
I think now there is some buffer, so we want to get a few more people involved.
Do you think you’ll keep the whole operation as a nonprofit that’s crowdfunded? Or do you think you’ll look for other sources of funding?
It’s going to stay a nonprofit. But I can tell you that what we’re looking into is a kind of split model like Mozilla’s, where the nonprofit will continue working on the core product, which will remain open source, nonprofit and so on, and we might start a for-profit business for software as a service, first to provide hosting for Mastodons for those who desire that.
Imagine if TechCrunch wanted to have their own Mastodon server but didn’t want to have their own IT department do it. You could go to us, have a contract and pay us a monthly hosting fee. Then you would have your own server on your own domain, but we would do the technical work.
This is in its early stages, by the way; this is in the planning stages, so not too many details to offer right now. But that is quite the classic open source business model. Nothing groundbreaking there. But it should be a sustainable and fair business.
You said “first.” Is the idea that you would not just host it but offer moderation and things like that, and provide services for those federated servers?
No. We would do just the hosting and the server would be completely under your control. Obviously, we would allow you to take your data and move it to your own hosting provider in the future or migrate from another hosting provider and so on.
It’s not really a new business model. There are hosting providers for Mastodon in the ecosystem right now, like masto.host. But they are overwhelmed with demand right now, which is why we’re looking into this. I think that we could offer something to help in this space, as well.
Is advertising something you’d want to provision across the platform or leave it to individual servers to decide? How would that work?
Well, you have to consider the fediverse nature of the network. Anyone could develop another platform using the same ActivityPub protocol but with completely different software around it, with different expectations and different features. If they wanted to build ads into it, they could, in theory.
The question is only as a user: Would you go to a service that has ads? And to make those ads effective, the service would have to keep track of your interests and location. Or would you just go to another one that doesn’t have that?
We, Mastodon, are not interested in the ad business or implementing ads into our code. But as I said, it’s free and open source, so anyone could modify it. They’d do it at their own risk with different business models.
Do you think there is more to the business model for Mastodon beyond hosting instances? What about the people who run those instances? Do you think that at some point, there’s going to be a need to come up with ways for them to generate revenue, to incentivize them? If so, what do you think that would look like?
I see a method in the sort of framework for building interoperable social media networks. You could think of an individual server as a separate social network, like Tumblr or Instagram. With interoperability built in straightaway, I think that it makes sense that they would be able to explore different business models or maybe build different features.
The fairest model that could probably appear in the ecosystem would be the paid accounts model. This is something in the past that App.net tried to do, but, I feel, did not succeed. It wasn’t clear if that was because of the paid account part or because they didn’t really build a good flagship product.
I don’t think they had a critical mass of people there, talking, to get anything off the ground properly.
I remember the idea was to build an open API ecosystem. The idea of building everything around the API and relying on third-party apps inspired me to take the approach that I did with Mastodon — making the API the main sort of way to interact with the server, and then building the front end around the same API that any other app could use.
But yeah, after that, the app that they had by default was not really very good. And they just always said, “Oh, somebody’s going to build something better.” But nobody did.
That was probably the main reason they didn’t succeed, and not necessarily because of the paid accounts thing, but who knows? That would definitely be something to explore. There are some servers right now that do this sort of thing in a roundabout way by, for example, requiring a Patreon subscription before you can sign up. That exists in the ecosystem.
There are also servers that run as co-ops, where you become part of the co-op before you join, and then you have part ownership of the server and you get to vote on decisions on moderation and so on. But I don’t think that’s very scalable, because you can’t imagine a call with millions of users all voting on something. But I’m not an expert on co-ops.
There’s room for a variety of different business models. Ultimately, the beauty of Mastodon is that you don’t need to have a server that grows to millions of people. Even if you have a small one just for yourself, your friends and family, you can participate in this global network. Running a server of that size is basically pocket money. So it really does not require a monetization strategy.
I noticed that the biggest server you run yourselves, Mastodon.social, is closed now for new registrations. Is that a matter of costs or operational power, and do you think you would ever open that up to more users, maybe as paid accounts? I can see a bit of cachet developing now on some servers over others. Someone might try to capitalize on that over time.
I’m not entirely sure. I can’t say anything right now about what we might do there. The main reason registrations are closed right now is just that it is a big burden on DevOps to scale up beyond the number of users.
I don’t want to say, “Oh, the software is not good enough to scale.” That’s not really the reason. It’s just a question of not having a dedicated DevOps employee right now.
I can’t run all of these organizational things and everything else. It’s just easier to close registrations and ensure that the people who are already on there have good quality of service instead of allowing more people to sign up and have everything slow down. The decentralized nature and the fact that there are plenty of other servers to choose from means that it’s kind of a victimless decision to make.
That said, I think my server is good. If we do a good job of moderating it, we’ll have a good balance of sensible moderation and not being too aggressive with our policies. Some servers on the political spectrum go a bit, in my opinion, too far with requiring content warnings.
There are a variety of different approaches on the fediverse. I think that in my service, it’s an approach that is closer to the mainstream.
Are you doing all the moderating yourself at this point? Or do you have people pitching in and helping?
Luckily, there’s a team of moderators now; I believe five people. They’re paid, but they are not full-time. I found them by posting a request saying we’re looking for more moderators. A bunch of people emailed me, and I went through the process of telling them what the job entails, what the remuneration is, and how we’re going to manage the communication and payments, and so on.
Is moderating all manual at this point? Are you trying to bring in any automation?
We don’t have any automation at the moment. I think that is one of the selling points of Mastodon — you get this personal approach with moderation because the size allows for that. The relationship between you and who hosts the server is a lot closer than between you and Twitter, Inc.
I think it’s also just the fact that moderation work is not automation-friendly. The simple cases are so simple that even if it’s a person doing it, it just takes a couple of seconds to do it. And when it’s complicated, then no automation can help. It requires a human to read into the context of the situation and to make the call.
That means it’s not proactive, but reactive, report-based moderation, right?
Yes. You submit reports, moderators look through the reports queue and then act upon it.
Do you think there’s scope for proactive moderation?
Some Mastodon servers do that — moderating every post — but I find that a little uncomfortable. At the end of the day, that increases the workload by a lot. I feel like it’s kind of a whack-a-mole situation, where you’ll never be able to achieve complete cleanliness of content.
At the same time, it reminds me too much of those upload filter laws that the European Parliament tried to pass, which the tech community resisted. If there is something problematic, it will get reported and it’s absolutely fair to just act on those reports.
What is the average size of a server on Mastodon now?
I don’t have that number. Mastodon Social has 881,000 users and 210,000 active users as of today, but I have really stopped promoting raw user numbers. A while ago, it used to be the main metric that Mastodon exposed both on the front page of a server and open statistics, and stuff like that.
I now think it’s a misleading metric. There’s a reason why platforms like Facebook and Twitter don’t really mention it, either. They always talk about either daily active users, monthly active users or monetizable daily users. The reason is that number just always grows, but it includes everything that’s not relevant.
The number of active users is always going to be a smaller percentage of the total that just goes up all the time. So we switched to just talking about active user numbers.
Did you think that you would have 881,000 people signing up? Do you think that you might kick some of the dormant users off at some point?
I have had different thoughts about this over the years. In the past, I’ve considered this, because, yes, dormant accounts consume space. But I realized a couple of years ago that it’s a negative PR move. Even if it makes sense, and even if it’s fair to kick off dormant accounts, it’s just bad publicity. So there’s no point in doing it. It’s not worth it.
Can I ask you about Zeon Federated, which I noticed was a company you registered in the U.K., but is now deactivated. What was the story there?
It’s a mildly interesting story. I used to run a forum in 2010, when I was still in high school. That was also the first time that I came across the concept of federation and the protocols involved back then.
I and another developer, who later worked with me on the forum, had the idea of doing a complete revamp of that forum and building a federated version of it. We started to do that a little bit but then we gave up. The name of that software was supposed to be Zeon Federated, so we registered the domain. However, the software back then didn’t work out. It kind of became my personal homepage.
I think a year or so after that, I started another business called Artists and Clients. That wasn’t a very creative name, but it was an escrow marketplace for getting hard commissions. We registered a limited company in the U.K. because it was so easy. It was just £15 or less and you could do it online. That was my first company registration ever.
So that existed for a while. And I do think, yeah, it was Zeon Federated Ltd. or something. I think that only existed until 2014, which is when I found a buyer for the platform, sold it for $2,000, and was very happy. That U.K. company was dissolved, liquidated or whatever happens to it. Honestly, I was a bit too young to really understand the implications of company registrations and stuff like that.
So wait, you founded an escrow marketplace for commissioning artists and then somebody acquired it from you? It’s probably pivoted into an NFT company now.
[Browsing the internet] Yeah, they are! They have NFT mentioned on their homepage! I just opened it.
Earlier, you were talking about possibly building a nonprofit alongside a business that could generate some revenue. Would you use your existing crowdsourced funding to build the latter, or would you go out and look for fundraising?
No, the funds that the Patreon receives are for the nonprofit and are untouchable for that purpose.
So you would need to raise outside money for the services business, you think? Have you approached VCs yet? I cannot believe you have not been contacted by investors.
Yeah, over the years, I’ve definitely received a lot of unprompted cold contacts from various VCs. I previously ignored them, but now we have Felix [Hlatky] — who basically works as the CFO, but he doesn’t have the title officially, yet. Now I forward those to him and he tries talking with them. Sometimes I tune into the calls.
We’ve tried talking to some VCs about this hosting business thing in the past couple of weeks. I have to say, though, they’re interested in somehow getting into the main product, and they’re not so much interested in a sustainable hosting business. So, VCs are not going to help here. We’re not letting them into the main product in any way.
So yeah, we’re probably going to have to go with an angel investor or just crowdfunding the hosting business separately, or just, I don’t know, maybe personal funds would be enough. That’s not entirely clear.