The list arrived as a spreadsheet: eight feature requests, roughly ranked by who had asked for them loudest. Our discovery pass turned it into two we recommended building this quarter, one to revisit later, and five we argued against outright.
Why "we could build all eight" was the wrong frame
Every one of the eight was technically feasible, most within a few weeks each. That was never the constraint. The constraint was the support and maintenance load a stable, working product would carry afterwards, and nobody had priced that into the request.
The question that did the sorting
For each item: who, specifically, asked for this, and how many other people have the same problem. Two items on the list came from a single vocal user each. Three others turned out to describe the same underlying problem in different words, which meant the honest count was one feature, not three.
What we recommended and why
- Built now: a bulk-edit tool, requested independently by four different users doing the same repetitive task by hand. High frequency, high pain, low ambiguity about what "done" looks like.
- Built now: a permissions refinement that was blocking a specific customer segment from adopting the product at all — not a nice-to-have, a sales blocker.
- Revisit later: an integration with a third-party tool that only one client currently uses, worth building once a second client asks, not before.
- Declined: the remaining five, each solving a real but narrow problem better addressed by a workaround we could describe in the meeting, in under a minute, for each one.
The conversation itself
This is the part that actually matters and the part most write-ups skip. We did not say no and move on. For each declined item we explained the trade-off in plain terms: this would take three weeks, serve one user well, and add a component the whole team now has to keep working for everyone else, forever. Framed as an ongoing cost rather than a one-off decision, four of the five requesters agreed unprompted that it was not worth it.
Every feature you build is also a feature you now have to keep alive. The eight-item list was really a request to take on eight ongoing obligations, and only two of them paid for themselves.
What made this possible
A relationship where the client trusts that "no" comes with a reason, not a shrug. That trust is built slowly, over engagements where we did say yes and delivered well, and it is the actual asset that let this particular conversation go well rather than becoming a negotiation.