The Roadmap Item That Paid for Itself Twice

You can fit anything in a suitcase if you're willing to sit on it. What you can't do is get to any of it easily once it's in there.

I've had the same conversation with sales more than three times this year, and it's always some version of the same shape: a big customer wants a specific thing on the roadmap, and the size of the check makes it very hard to say the word no. So I don't say it. I say yes, we push the feature up, engineering shifts two sprints of work around it, and six months later I'm looking at what got delayed to make room and asking myself if the trade was actually worth it. Usually the honest answer is no.

I'd been doing a worse version of Peter Drucker's idea about deciding what not to do for years without knowing it had a name — I just called it triage, or firefighting, or "whoever's loudest gets the resource." What his framing changed for me wasn't the discipline, I already had that on paper. It was the order of operations. I'd been deciding what to do first and treating "what not to do" as the leftover. He flips it. You decide what you're not doing on purpose, before revenue is in the room making the case for itself.

The part I don't love admitting: I still say yes more than that framing would want me to. A customer that size, asking directly, with a renewal coming up — I feel the pull every time, and I don't always resist it. I've built things that shipped fine, worked fine, and then sat there mostly unused by everyone except the one account that asked for them, while something with broader pull sat in the backlog another quarter. I don't think I've fully fixed this. I think I've gotten slightly better at noticing it happening in real time, which is not the same thing.

If you run product and a big account asks for something specific, where's your actual line — and how do you know when you've crossed it?