How long should a solo MVP take to build?

7 days. If the smallest version somebody would pay for takes longer than a week for one person, the idea is too big to test, and you should cut it down before you start rather than after.

Where the 7 days comes from

I spent 14 years as a contract developer. What I learned, without noticing, was how long a build actually takes as against how long anyone says it will. That number never left me, and it's why every deadline in my books is 7 days rather than a month.

A month gives a project room to grow features nobody asked for. A week forces the question: what's the one thing it must do?

What fits in a week

The product I spend my Sunday evenings on was built in 7 days, after asking for the money first. Day 1 went on the checkout and the payment account, because paperwork is the part that can't be rushed. The core of the product took 2 days. The machinery that runs it took half a day, because I'd built it once before for a product that died. Day 5 was the check that stops it sending out wrong answers.

Nothing heroic: that's what happens when the stack is dull, the scope is one job, and the slow parts go first.

What doesn't fit, and what that tells you

My first product took 6 weeks, mostly spent learning a framework I'd wanted to try. A tool for tracking which client had signed off which version of a design took 9 days, and nobody had paid by week 3.

When a build runs long, it's usually saying one of 2 things: you're building more than one job, or you're learning on the customer's time. Both are fixable before you start.

The week, in order

1. Day 1: the payment account and the checkout. Not last. First.
2. Days 2 to 4: the one job the product does.
3. Day 5: the check that catches it when it's wrong.
4. Days 6 and 7: the page that explains it, and the first message to the people who paid.

Ship Alone sets out the week in full, with the dull stack that makes it possible.