Skip to the page

Plainly Put

One sentence about building things, taken apart

HomeFounders and teams

The cheap yes and the expensive one that follows

A yes costs nothing to say and a great deal to deliver. A no is uncomfortable for a minute and then costs nothing at all.

No is easier to do. Yes is easier to say.

Jason Fried

The line is eight words long and describes a timing problem. Saying yes feels good at the moment of saying it: the customer is pleased, the colleague is grateful, the meeting ends warmly. The cost arrives later, as work. Saying no is awkward at the moment and then nearly free, because nothing has to be built, maintained or explained afterward.

Jason Fried and 37signals have made a habit of the harder answer. Their products have stayed deliberately small, and their public writing has repeatedly argued that most feature requests should be declined by default, with the popular ones allowed to prove themselves by coming back again and again. For a company selling simple tools to small businesses, that stance has worked. Each feature they declined is one they never had to support, document, test on every browser or explain in the help pages.

The asymmetry shows up well beyond product decisions. A founder who says yes to every coffee meeting, conference panel and advisory request has no afternoons left. A manager who agrees to every deadline commits the team to work it cannot finish. In each case the yes was easy and the consequences were distributed across weeks.

Saying no well is its own skill. A flat refusal can cost goodwill that a young company needs. The 37signals habit has been to explain the reason, often in public, so that customers understand the product is small on purpose rather than by neglect. A customer who hears why a feature will not be built is more likely to stay than one who hears nothing.

The line needs one correction for very young companies. A startup with ten customers cannot afford the posture of a profitable firm with thousands. Early on, saying yes is often how a company learns what it is: a founder who builds a custom report for the first serious customer may discover the product’s real use. When Airbnb’s founders flew to New York in 2009 to photograph their hosts’ apartments themselves, they were saying yes to work no software company would plan for, and it taught them why bookings were stalling.

What changes with time is the cost of the yes. When a feature is used by one customer and maintained by the founder over a weekend, agreeing is cheap and informative. When the same feature must work for thousands of accounts, be supported by a team and survive every future release, it is expensive and mostly permanent.

A practical rule follows. Before agreeing, estimate what the yes will cost to do, not to say, over the next year. If the answer is small and the lesson is large, say yes. Otherwise the uncomfortable minute of saying no is usually the cheaper choice.