An MVP app in weeks, not months

Ask an agency how long an MVP takes and the answer is three to six months. Ask a founder who shipped one last year and the answer is a few weeks of evenings, most of them waiting on the stores. Both are telling the truth about different things. The agency is building a product. You should be building an MVP, which is a different size. Here is the timeline for the right size.

1. First, the right size

MVP stands for minimum viable product: the smallest version of the app that real people can use for the real job. Minimum is the operative word, and it has a number: about five screens. The screen that does the job, one detail, one add-or-edit form, one list of people or things, and a profile. One job, done end to end, for one kind of user.

The trap is adding “just one more” screen at the design stage, because each one is cheap to draw and expensive to wire, test and support. Decide the five and write them down before anything else.

2. Days: the design

Two to three days. Day one, the job in one sentence and the five screens as boxes and arrows. Day two, the real screens, starting from a working starter rather than a blank canvas, in your own words and with your own name on it. Day three, on your own phone, tapped through by two people who have the problem, and the fixes from that.

If this takes a week, you have more than five screens.

3. Days: the build

With the design as structure (screens, arrows, content), a coding agent produces the working app in a day or two: a real project, the screens wired, the phone’s own controls. You run it on your own phone the same day. The day after goes on the edges: a list that loads real data, a form that saves, the empty states.

Where it goes wrong: prompting the agent from a description instead of a design (it guesses, you redo), and adding a backend before the first user exists. Start with sample data. Connect a backend when someone other than you is entering data.

4. Weeks: the stores

This is where the calendar fills, and none of it is your fault:

• Google Play: $25 once, identity checks, and for a new personal account a closed test with at least 12 testers for 14 days before you can apply for public release. Then a review of a few days.
• Apple: $99 a year, identity checks, and a review that is usually under a day when it passes, but first apps often go through two or three rounds. Most rejections are for unfinished things: placeholder text, a dead link, no privacy page, a sign-in with no demo account.

Start both accounts on day one of the design, because the identity checks take days of their own and the 14-day test can only begin once you have an app to install. Your two Tuesday testers plus ten friends are the twelve.

5. Where the months come from

Months come from building the product instead of the MVP: accounts and payments before the first user, a design agency, a custom backend, feature requests from people who have not used it yet. Each is a good idea for version two and a month added to version one.

So the honest answer to “how long” is: days to design, days to build, two to four weeks to get into the stores, and a first user on your phone in week one. Open the App Builder, pick the starter closest to the job, and get the five screens onto your phone today. The store clock starts when you have something to install.