On online communities

I wrote this thread a short while back and figured I should edit it for posterity here. It’s from two decades ago, but I promise, it’s relevant for this age where we’re again trying to grasp how online communities work.

In the summer of 2003, I was hired to run technical operations for an early virtual world for teens called Habbo Hotel. The thing had been existence for a couple of years, with the originally-Finnish service localized into English (for the UK), German (for the Swiss) and Japanese. UK and Swiss? I’ll come back to that later.

We’d been funded by first-tier European venture capital for this, while the business model was to sell premium items inside the virtual world. Very few people understood at the time what the hell that meant, and even fewer believed it could ever work. Free-to-Play wouldn’t be invented as a term until several years later. And our audience being teens, we couldn’t rely on credit cards as a payment mechanism, so it was all more than a bit weird. But that’s a story for another time.

My first job was to make sure we could relaunch it localized for Spain, too. With a small team in Madrid supported by us in Helsinki, the software was localized and a marketing campaign planned.

Meanwhile, I was focused on finding a hosting service we could reliably run it with. This was before cloud services were widely available, and anyway, the software had features which would not be ideal for cloud hosting for years to come (namely, a very large, very busy MySQL database).

We did find a new hosting provider in Germany which we ended up leaning on heavily for the next decade. My team, 4 people at the time, got to work re-installing everything and deploying the software for its final campaign implementation, and the team on the community side got busy with marketing. That launch in Spain turned out to be quite successful.

It was followed within a month by a launch of the service for Italy, a few months later in Sweden, Netherlands, then Germany (separate from the one for Swiss German that already existed), Canada, Norway and France. Later on, we’d follow with another wave of sites, but these ones kept us busy for my first year.

Every one of these was the same software, but a complete re-localization, new community, and a local team to market and support that community. New media partners, new ad campaigns, new local payment services, often partnerships with the local telcos (remember, this was 2 decades ago), new safety team, and so on. And here’s where that country, not language focus comes in.

Habbo’s market was (and still might be, I haven’t kept track — it does still exist) teens, and we had very early on identified that launches worked much better for us when they were hyper-localized to not only countries, or to cities, but to towns and schools.

This was an audience that in theory is curious about new things, but also highly concerned about their peers’ opinions. Few would want to be seen doing things others were not doing — so it’d be better to have the majority of a single school to play, rather than a few people in many schools.

Facebook would use the same strategy when it started spreading from Harvard to Stanford, Columbia and Yale. Remember how in the beginning you couldn’t open a (The)Facebook account without having a specific school’s .edu email account? Yeah, using exclusionary access to concentrate demand and buzz into a specific community had been, was, and still is a tried and true launch strategy.

Anyway, in the case of Habbo, these sites were entirely disconnected from each other. Even if you had been inclined to do so, you couldn’t connect from a Spanish account to a German member — you’d have to sign up again in the other service.

This worked for Habbo, because we wanted to keep the communities focused on their local languages and cultures. It also made the job of moderation and safety much, much easier for us, because moderators would only need to understand the local youth culture.

Moderation is a very culture-specific affair, after all. And we indeed had moderation. A lot of it. At no time would there not be at least one moderator logged in and able to react in real time to events — we even closed the majority of the service for night if traffic didn’t justify having moderators awake.

Of course we also had a lot of software doing auto-moderation, flagging suspicious activity, suspending users for bad behavior, and so on, but it was always supervised by a person. Safety was a section of its own in our business reporting, and my team was also responsible for the tools used in the work.

The reason I decided to write this down was that we’re again in a bit of an experimental time in online communities. I feel that some of the lessons from the earlier experiments may have faded, and it would be wise to keep in mind that while technologies have changed and the world evolved, us people are still just people.

We like to be surrounded by people we understand, and feel that they understand us. We’re not really equipped to deal with global reach, and even those among us who can boast of having followerships in the tens of millions, really truly connect with hundreds (or fewer!)

Communities are small, and most of them are local. Twitter may have been a global service, but most of us weren’t looking at it globally. Most of the alternatives we now are testing out may also be global, but we’re better at experiencing them through a macro lens with local focus. Mastodon, in particular, is not a community, it’s software used by thousands of communities, many of which weirdly carrying the name of the software, revealing their tech enthusiast origins.

We have identities, and we fundamentally feel better when we can say that identity is grounded to something we understand. So by all means, reach out to the world, but also, invest in that local community around you. And do think about more than just the software.

Thanks for reading.