We gave users exactly what they asked for. They stopped coming back.

Last year, our most-requested feature at Murror was a chat function. Users wanted to talk to the AI like a friend — open-ended, anytime, about anything.

It made perfect sense. The data backed it up. Surveys, support tickets, App Store reviews — everyone was asking for the same thing.

So we built it.

Within two weeks, our core metrics started dropping. Session length went down. Return rate declined. The journaling habit we'd carefully nurtured was breaking apart.

What happened?

The chat feature gave users exactly what they wanted in the moment: instant comfort. But it replaced what they actually needed: the slow, sometimes uncomfortable process of sitting with their thoughts and writing them out.

Journaling is hard. It forces you to organize messy feelings into words. That friction is the product. The chat feature removed the friction — and removed the transformation along with it.

Users felt better in the short term but grew less over time. They were consuming emotional comfort instead of building emotional resilience.

We ended up redesigning the feature completely. Instead of open-ended chat, we created guided reflection prompts that appear after a journal entry — helping users go deeper into what they already wrote, not escape from it.

The lesson that changed how we build: user requests tell you what people want to feel. Your job is to figure out what will actually get them there. Sometimes those are the same thing. Often they're not.

Now before we build any requested feature, we ask one question: does this make the user more capable, or more dependent?

Has anyone else shipped a beloved feature only to watch it quietly erode the thing that made your product work?