How to cut your MVP down to something you can actually ship
George · 2 min read
The short answer: write down the one job your product does for one kind of person. Keep only what that job needs to work end to end, do the rest by hand or not at all, and set a ship date before you start. An MVP that does one thing well teaches you more than one that does five things halfway.
Almost every founder has a project that was "nearly done" for months. It rarely stalls because the code is hard. It stalls because the list keeps growing: one more setting, one more integration, a dashboard nobody asked for.
Find the one job
Finish this sentence: "[A specific person] uses this to [do one thing] so that [one result]."
"Freelancers use this to send an invoice in under a minute so they get paid faster" is a product. "A platform for freelancers to manage their business" is a roadmap. If your sentence has an "and" in it, you probably have two products.
Cut everything around the job
Go through your feature list and ask of each item: does the job fail without it?
- Accounts: often yes. Social login or a magic link is enough; you do not need profiles, avatars and settings pages.
- Payments: if you are charging, yes. One plan is enough to start.
- Admin dashboard: almost never. You can look at the database.
- Team features, roles, exports, themes, integrations: almost always later.
Everything that is not essential goes into a "later" list. Writing it down matters: it is easier to cut a feature when you know it is saved, not lost.
Do things by hand
Many features can be replaced by you, for now. Send the onboarding email yourself. Generate the report manually for the first ten customers. Approve accounts by hand. Doing it yourself is slow, but it shows you exactly what users need before you build it, and you will build it better later.
Set a date, then cut to fit
Pick a ship date two to six weeks away and tell someone. When the date gets close and the list is still long, cut more instead of moving the date. A fixed date turns "should we add this?" into "what can we drop?", which is the question that gets products shipped.
What not to cut
Scope is not quality. Keep the core flow fast, clear and reliable. Users forgive a missing feature; they do not forgive losing their data or a sign-up that breaks. Small and solid beats big and shaky.
The short version
- Describe the one job in one sentence, without "and".
- Keep only what that job needs to work end to end.
- Do the rest by hand, and write the later list down.
- Fix the date and cut to fit it.
- Cut features, never reliability.
When it is shipped, show it to people. That is what the launch is for.