The takeaway: Default to "not now" for every feature request - yours included. Listen, log it, and wait. The requests that keep coming back earn a serious look; the rest quietly disappear. When you do say no, be honest and direct, and remind people what your product is for.
Why "yes" is more expensive than it looks
A feature request feels small in the moment. The founders of 37signals (the company behind Basecamp) argue the opposite: in REWORK, they write that "you rarely regret saying no, but you often wind up regretting saying yes." In a 2024 episode of their REWORK podcast, Jason Fried puts the trade simply: yes is cheap now and expensive later. Every feature you add is design, code, testing, documentation, and support - forever. And as David Heinemeier Hansson notes in the same episode, commitments accumulate: say yes easily for a few months and you have silently spent your next year of decision-making.
Their suggestion is a useful test for any request: would you say yes if you had to do the work tomorrow? If the honest answer is no, the request may not deserve a yes just because it sits far in the future.
Start with no
37signals' older book, Getting Real, names the practice directly: start with no. Their version of the workflow:
- Listen, but do not act. Every new request - from users or from your own team - meets the same initial response: "not now." This is a filter, not a dismissal.
- Let requests prove themselves. Getting Real compares it to a feature having to stand on the porch for days before being let in. If a request keeps coming back from different people, that repetition is your signal to look deeper.
- Protect the product's shape. The book's line is that the secret to building half a product instead of a half-ass product is saying no. Each yes is adopting a child: you are stuck with it, and taking a released feature away from customers is far worse than never building it.
The same chapter carries a Steve Jobs anecdote via Derek Sivers: pressed with a thousand feature ideas for the iTunes Music Store, Jobs answered that innovation is not saying yes to everything - it is saying no to all but the most crucial features.
How to say no to a user without losing them
A no lands better when it is plain and respectful. A few patterns that tend to work:
- Answer directly. "We are not planning to build this right now" beats silence or a vague "we'll consider it." False hope turns into a worse complaint later.
- Explain what the product is for, not why their idea is bad. Getting Real suggests reminding people why they chose you in the first place: because the product stays focused and does not try to please everyone.
- Offer the closest path today. If a workaround or an existing feature covers part of the need, name it.
- Log the request where you can see patterns. A simple tally is enough at first. Our post on how to decide which feature to build next covers what to do when a request does earn a real look.
When a no should become a maybe
Starting with no is not ending with no. The filter works because it is cheap: truly important requests resurface, through different users, in support emails, in your own usage. When you hear the same need for the third or fourth time, that is when it deserves the deeper evaluation - user problem, smallest useful version, time budget - rather than a reflex build. Talking to the users behind a recurring request also helps; our guide on how to talk to your users and hear the truth can sharpen those conversations.
If you are still pre-launch or early, a focused product is also easier to describe in one sentence - and early adopters notice that. You can launch your startup on Maidensail and test whether a sharp, narrow promise lands with real users.
The takeaway
Make "not now" your default, let repeated requests earn their way in, and deliver your no's honestly with a reminder of what the product is for. A small product that does a few things well is not a compromise - it is the strategy.
Your launch deserves a focused story
A product with a clear scope is easier to explain and easier to adopt. When you are ready, launch your startup on Maidensail and put that one-sentence promise in front of founders and early adopters.
Sources: Getting Real - Start With No (37signals); REWORK podcast - Say No by Default (37signals, 2024). Quotes attributed as in the sources.