A Community Becomes a Different Community Without Anyone Deciding

The forum I moderated was, in year one, a place where people helped each other with a specific technical problem. By year five it was a place where people talked about the industry around that problem, with occasional technical questions that were mildly unwelcome. Nobody decided this. The rules never changed. Roughly forty percent of the year-five active members had been there in year one.

The mechanism, as far as I can reconstruct it, is a small feedback loop running continuously. Every post that gets a warm response makes similar posts more likely. Every post that gets ignored makes similar posts less likely. Neither event is significant on its own, and over five years and a few hundred thousand posts, the composite is a completely different publication.

What made it invisible is that at no point was the change large enough to notice. Compare any two consecutive months and they are the same place. Compare year one and year five and they share almost nothing. This is a general property of gradual processes and it is why the people best placed to notice — the long-standing members — are the worst placed, because they experienced every increment as normal.

There is a version of this that is fine. Communities are allowed to become different things, and a place that had solved its original problem should arguably move on. But two things go wrong reliably.

The first is that the stated purpose stops matching the actual one, and new members arrive on the basis of the stated one. Our description still said we were a place to get help with the problem, and people would arrive with the problem and post about it and get a lukewarm reception, and leave thinking they had done something wrong. They had not; the description was four years out of date and I was the person who had not updated it.

The second is that the people the place no longer serves leave quietly, and quiet departures do not register. There is no event. I only found out how many technical people we had lost because I ran a query on posting frequency by cohort, and the answer was that a specific group had thinned out over eighteen months with no complaints and no goodbyes. Complaints come from people who are staying. The strongest signal is silent.

What I would do differently is one small thing: read the description of the place out loud, to yourself, once a year, and ask whether it is true. It takes two minutes and I did not do it once in six years. And if it is not true, either change the place back — which is hard and usually fails — or change the description, which is easy and at least stops the mismatch from being a trap for newcomers.