Every roadmap is a list of no's wearing a yes's clothing. Saying yes to one thing is always, implicitly, saying no to everything else you could have built with that time — the difference is whether you do it on purpose.
Decline the request, not the person
The instinct to soften a "no" with vague language ("maybe next quarter") is worse than a direct answer. It keeps a request alive in someone's head indefinitely, which costs more goodwill than a clear no followed by a real reason.
State the tradeoff explicitly: here's what we'd have to not build to do this instead. Most reasonable people accept a clear tradeoff faster than a soft maybe.
Keep a public "not now" list
A visible backlog of declined-for-now ideas does two things: it shows requesters they were heard, and it gives you a paper trail when the same idea resurfaces in three months with a different champion behind it.
nurvex