The cost of a fast no

Every team I have worked on has eventually discovered that saying no is free. You decline the feature, the roadmap stays legible, the release ships, and nobody has to maintain a thing that three customers asked for and one of them meant. The discipline gets celebrated. It goes on a slide.

What the slide leaves out is that a fast no and a good no cost very different amounts, and only one of them is actually cheap.

The two refusals

A fast no closes the ticket. It is a sentence about scope, usually correct, usually delivered before anyone has understood what the person was trying to do. It protects the roadmap and it teaches the requester one thing: that asking has a low expected value. They stop asking. You keep shipping. From the inside this looks exactly like focus.

A good no costs an hour. You find out what the person was doing when they hit the wall. You almost always discover that the feature they requested is a workaround they designed themselves, in a hurry, without access to the codebase — and that the thing underneath it is either much smaller than the request or much larger.

A feature request is a bug report about a workaround, written by someone who cannot see the code.

Half the time the hour ends with the same no. But it ends with a no that names the real problem, and the requester leaves with something they did not have: a reason.

What you actually lose

The cost of a fast no is not the feature you declined. It is the report you will not get next time.

Teams that decline quickly are not flooded with bad requests — they are starved of information. The people closest to the work stop telling you what is broken, because telling you has never once changed anything. You end up with a roadmap built entirely out of the opinions of people who are paid to have opinions, which is the worst possible sampling of a product.

The tell is a quiet backlog. If nobody is asking you for anything unreasonable, it is not because the product is finished. It is because the channel is closed.

Making it affordable

The hour is only worth spending if it produces something durable, so write it down. Not in the ticket, which will be closed and never read again — somewhere the next person can find it. Three lines is enough: what they were doing, what stopped them, what we decided and why.

Six months later that file is the most valuable document in the project. It is the only record of what the product refuses to do on purpose, and the difference between a principled product and a haunted one is entirely whether that record exists.

The refusals accumulate into a shape. Read fifty of them and you will see what your product actually believes, which is usually a slight surprise, and always more coherent than anything on the slide.