Skip to content

Part II: Building the first version

8. What your first version is for

Build something small enough to try and useful enough to learn from.

By the end of this part, you'll have a version of your product at a live URL that you can give to a friend. They should be able to do something useful with it. You'll be able to watch what happens and change the parts that need work.

Bring the problem you explored in Part I. You have someone in mind, an outcome they care about, and an idea for helping them achieve it. Building gives you another way to investigate that idea. Some questions become much easier to answer when someone can actually try the thing.

Build without guilt

With AI helping you write software, you can try an approach before you've argued every detail into place. Build a rough interaction to see how it feels. Try a different layout. Find out whether the idea makes sense once you have to put it on a screen.

Build a rough interaction to see how it feels.

Give yourself permission to throw that work away. A version that helps you discover a simpler approach has done something useful, even if none of its code survives. You don't owe a feature a place in the product because you spent an evening on it.

The cost of an experiment still matters. While you're building with made-up content, you can change a screen and throw it away without affecting anyone. Once someone is using the product to save their work, an experiment could change or delete something they need. Use made-up data while you explore, and keep a way to undo your changes.

Once your friend has a version they can use, you can try new ideas in a separate development version. Keep their version working while you decide which changes are worth sharing.

Decide what this version should teach you

“MVP” stands for minimum viable product. For this handbook, think of it as a small version someone can use to get a worthwhile result. The useful question is what that version needs to do for the person you've chosen.

We'll continue with the hypothetical designer from Part I. Suppose you want to find out whether putting a draft beside a clear request for a decision helps a client respond. A useful first version could let the designer create that request, share it, and read the client's answer.

You could learn whether the client understands what is being asked. You could see whether the response gives the designer enough information to continue.

Write down what you want to learn before you build. It gives you a reason to stop adding things and try what you have.

Keep the first promise small

Your friend should be able to complete the main task. If a button says it saves their response, the response needs to be saved. If something fails, they need enough information to know that it failed and what they can do next.

There can still be rough edges. You might create their account yourself, or send the invitation manually. Decide which work you'll handle behind the scenes and make sure it doesn't prevent them from trying the part you need to learn about.

A friend's reaction will also have limits. Someone who has never needed client feedback can help you spot a confusing button, but their enthusiasm won't tell you whether designers want the product. Choose a friend with relevant experience if you can, and keep talking to the people from Part I.

Payments and analytics can wait for the forthcoming Part III. For this first trial, observing what your friend does and asking about the experience can give you plenty to work on.

In the notes you've been keeping, write a short description of what your friend will be able to do. Then add the question you hope trying it will answer. We'll use both to choose the first useful journey.

Before you move on0 / 3