Questions · Published
What features should the first version of my app have?
TL;DR
Only what's needed to deliver the one outcome your first customers will pay for, and it's usually much smaller than you think. Work back from your promise, then sort everything into Now, Next and Later. Now is the first version, and some of it can be done by hand at first, like setting customers up or sending reports.
The first version of most software is much smaller than its founder expects. Its one job is to keep a single promise to your first customers well enough that they'll pay for it. Everything your product will eventually do can come later.
Work back from your promise
Start with the outcome you've promised, then ask what's the least the software needs to do to deliver it.
Say your promise to temp agencies is "fill a last-minute shift by the end of the day, without ringing round twenty temps". The first version might only need two things: a page where a manager posts a shift, and a text to available temps that they can accept with one reply. That's it.
A simple way to keep it this small is a roadmap with three columns:
- Now: everything needed to deliver the first version. For the temp agencies, that's the shift page and the text, plus whatever you do by hand to make them work.
- Next: what you expect to build once customers are paying, like recurring shifts or a simple weekly report.
- Later: what you'll build only if customers keep asking for it, like a payroll integration or a mobile app.
Now is a commitment. Next and Later are your current best guesses, and what you learn from customers will almost certainly change them. That's fine. They still give you a direction and keep good ideas from getting lost.
Do things by hand first
Some of what's in Now doesn't need building yet. Paul Graham's essay Do things that don't scale makes the case for doing things by hand at the start: setting up each customer yourself, sending reports as emails you write, handling the awkward cases personally.
For the temp agency idea, you could set up each agency's list of temps yourself and write the weekly report as an email. They're still Now items, because the first version doesn't work without them. You'll learn far more about what customers need by doing them than any feature would teach you, and once something has earned its place, it moves into Next to be built properly.
Taking payment can be simple too
You need a way for your first customers to pay, but it doesn't need to be built into the app. A payment link or an invoice is fine for the first few. Plans, seats and usage-based pricing can come once you know what customers actually buy.
Don't sell what you haven't built
Keep your sales conversations to what the first version does, plus what's genuinely coming next. Promising features that are still in Later sets expectations you can't meet, and it pulls your first version back towards being big.
Signs your first version is too big
- It's taking weeks, and nobody has paid yet.
- You're building features no customer has asked for.
- You can't explain what it does in one sentence.
- Much of it covers situations your first customers don't have.
If any of these sound familiar, cut back to the one outcome and get it in front of customers. When you build with an AI tool, give it a clear brief that covers only that first version: who uses it, what each of them needs to do, and the important rules and edge cases.