The screen we pulled

I can still picture the screen. It was, by every standard we had written for ourselves, a good screen.

It appeared right after a user's first match. The language was careful. The framing was trauma-informed. It carried the same restraint we had tried to put into everything. And on a Tuesday in late October 2024, looking at it for what must have been the hundredth time, I saw what was wrong with it, and could not unsee it.

One button was slightly larger than the other. There was a gentle implication, in a line of copy we had revised three times, that the right next move was to send a message - and, underneath that, a quieter implication that not sending one would be a kind of mistake. None of it would have failed an audit. A reviewer would have passed it. It satisfied every internal guideline we had. And it was, unmistakably, a soft version of the exact engagement engineering I had spent eighteen months telling people we existed to refuse.

What frightened me was not the screen. It was what the screen implied about everything underneath it. If a careful screen could drift that far without anyone noticing, then the thing underneath the screens - how we were quietly defining success, which metrics were governing what shipped, the words we used in our own meetings - had to be drifting too. The screen was not the problem. The screen was the symptom I happened to catch.

So we stopped. Not paused - stopped. I want to be honest that this was not a serene, principled moment; it was frightening. We had a product that worked. We had a version that would have shipped and, I am fairly sure, would have done commercially well. Stopping meant telling the cohort of users we had what we were doing and why. It meant refunding subscriptions. It meant losing momentum we could not easily get back, and, over the following months, losing several people who had joined to build the first version and did not want to build the slower second one. The cost was real, and I am not going to pretend it wasn't.

What we did with the stopping was write down the Refusal Register - a short list of things we would never do, each with the reasoning, the growth argument we had heard for doing it anyway, and the deliberately difficult conditions under which we would ever reconsider. The screen we pulled is the reason that document exists. I keep a description of it in my notes, not the design itself, just the account of how close we came and how respectable it looked while we did.

I am writing this as a story rather than as part of the essay it belongs to because I want it to stay specific. It would be easy to tell the company's history as a series of principled stances. The truth is smaller and more useful: it turned on one ordinary screen, on one ordinary Tuesday, that looked fine and wasn't. The whole of what Wouch became is, in some sense, the decision not to ship it - and the much harder, much longer work of rebuilding the team's thinking so the next one would be caught sooner.

Wouch, a relationship-readiness platform (not a dating app).