The client I turned down, and what it taught me

The budget was real, the problem was interesting, and I nearly took it anyway despite everything my gut was doing.

The signals

Three meetings in, I had still not met anyone who would use the thing. Every request came through one person who described the team's needs confidently and got visibly impatient when I asked to speak to them.

Then a sentence that settled it: that the team would use it because they would be told to.

Why that is fatal

I had already built something for a team who had never been in the room. I knew exactly how that ends — it ships, it works, and adoption is one person. The specification is not the problem. The absence is.

You cannot make people use software by instruction. You can make them open it once.

How I said no

Badly, the first time. I hedged and talked about capacity, which was a lie and I think they knew it.

I went back and said the real thing: that I did not think it would get used in the way it was being set up, that I would need access to the team for it to work, and that without that I would be taking money for something likely to fail.

What happened

They were annoyed. Then, about two months later, they came back and gave me the access.

I do not tell that story because it always ends that way — usually it just ends. But saying the honest thing badly is still better than taking work you already know will fail.