The Platform Changed the Deal Four Times and Each Time People Were Surprised

Over the years I have watched the same event happen on four different platforms, and each time the reaction was a mix of anger and genuine surprise that I found more interesting than the change itself. The event is: a platform alters the arrangement it had with the people who make things on it, in a way that reduces what they get.

Reach is throttled. An API that a hundred tools depended on is closed or priced. Revenue share moves. A format that was promoted stops being promoted. In every case there was an announcement, in every case it was framed as an improvement, and in every case a group of people who had structured their working lives around the previous arrangement discovered that they had no standing to object.

The surprise is the part worth examining, because the terms were never secret. What creates it, I think, is that a stable arrangement starts to feel like a property. If a platform has sent you a certain amount of traffic every month for four years, that traffic stops feeling like something you are being given and starts feeling like something you have. It is not, and there is usually a line in the terms of service that says so explicitly.

The useful frame, which I arrived at slowly and after being caught out once: everything a platform provides is a term, and terms have a revision schedule you are not on. The question for anything you build is not whether the terms will change but what happens to you when they do.

That question has a concrete answer and it is worth writing down. For each thing you depend on, what is the fallback and how long would it take to activate. For a community, the answer is usually about contact: if the platform vanished tomorrow, could you reach the people? If the only route to them is inside the platform, you do not have a community, you have an audience the platform is lending you.

The single most useful thing I did in six years of moderation was, for that reason, unglamorous: I maintained a mailing list. About eleven percent of active members ever joined it, which felt like a failure at the time. When we had to move — for reasons that were nothing to do with the platform, as it happened — that eleven percent was the seed that made the move survivable, and the eighty-nine percent were mostly lost.

The other thing I would say is that a platform changing terms is not a betrayal, and framing it that way makes people worse at preparing for it. It is a company making a decision about its own product, which is what companies do, and the expectation that it will not is the actual error. The people I have seen weather these changes best were not the ones who trusted the platform least; they were the ones who had, quietly and in advance, answered the question about what they would do.