The smallest thing that could work
July 22, 2026 · 4 min read
A colleague of mine used to open every planning conversation with the same question, and it drove me slightly mad for about a year before I understood it. She would let everyone finish laying out the plan, wait a beat, and then ask: what is the smallest thing that could work?
Not the smallest thing we could ship. Not an MVP. The smallest thing that would actually solve the problem for a real person, even if it embarrassed us.
Why it lands harder than it reads
Most plans are not too ambitious. They are too wide. Somebody names a real problem, and then five people attach their own adjacent concerns to it, and by the time it reaches a document it has become a platform. The question does not attack ambition. It attacks width.
The honest answer is usually uncomfortable, and it is usually a form. A form and an email. A spreadsheet somebody updates by hand on Fridays. I have watched a team spend two quarters on a scheduling service whose first honest answer was a shared calendar, and the shared calendar would have told us within a fortnight that the real problem was nobody trusted the data.
The version I actually use
I have added one clause. After the smallest thing, I ask what we would have to learn for the small thing to be wrong. If nobody can name it, we do not have a hypothesis — we have a preference, and preferences do not need two quarters.
Small is not a budget. It is a way of finding out sooner.
The thing nobody warns you about is that this gets harder as you get more senior, not easier. A small answer from a junior engineer reads as pragmatism. The same answer from someone with a title reads as a lack of vision, and you will feel the pull to say something larger. Say the small thing anyway.