Test your idea before you build it

Before I build anything, there is a step called sketching: fast loops on paper, in Figma (a program for designing screens) and in throwaway prototypes. A prototype is an imitation of your product, with just enough in it to try it out. I would rather throw away the wrong idea cheaply than ship it expensively.

You can start testing yourself too, without being able to program. Say you want an app that lets people in a neighbourhood order their bread from the baker in advance.

Write down what you want to know

Every idea rests on assumptions. Here there are at least three: people sometimes find an empty shelf, they would order in advance because of that, and the baker wants to handle those orders. If the first one is wrong, the other two fall away as well, so start there.

Also write down in advance what you would count as a good result. For example: out of ten neighbours, at least four bring up the empty shelf themselves. It is easy to explain away any result afterwards, and a number on paper makes that harder.

Talk to people

A conversation only costs time. Ask about what people have done, because they can answer that truthfully. “Would you use an app like this?” usually gets a polite yes, and that yes costs them nothing. Ask this instead: “When did you last go to the baker? Did they still have what you wanted? What did you do then?”

Say nothing about your idea for as long as you can, because once people know it they start helping you instead of answering. Talk to people you don’t know well if you can, and write down literally what they say. Pay most attention to what they already do to solve the problem. Someone who phones the baker every Friday to have a loaf set aside really does have that problem.

Don’t forget the other side. Ask the baker how orders come in today and what goes wrong. The book The Mom Test by Rob Fitzpatrick is entirely about conversations of this kind.

A clickable prototype

A clickable prototype is a series of screens linked together. You press a button and the next screen appears, but nothing works behind it. You make it in a program like Figma, and paper screens you swap by hand also work at the start.

Then give someone a task: “Order two loaves for Saturday.” Say nothing else and watch. Where does that person hesitate? What do they press first? Which word don’t they understand? Changing a screen costs next to nothing at this stage, and you can carry on testing with someone else the same day.

A prototype shows whether people understand how something works. Whether they would also use it is something you are better off finding out another way.

A landing page

A landing page is one web page that says what the product does and for whom, with one button to sign up. You can make one with a site builder, without programming. Say honestly that the product doesn’t exist yet, and that people will hear from you when it does.

Then people have to get there: a flyer at the baker’s, a post in the neighbourhood Facebook group, perhaps a small ad. Count how many people visited the page and how many signed up. Leaving an email address asks more of someone than nodding politely in a conversation, so that figure tells you more. Someone who pays up front gives an even stronger signal, as long as you say clearly that they get their money back if the product never comes.

An email address is personal data. Say on the page what you use it for, use it only for that and check what the privacy rules (GDPR) ask of you. If in doubt, put it to a lawyer.

Doing it by hand for a while

You can also offer the service with no software at all. The neighbours send in their order through an online form or WhatsApp, and you take the list to the baker every evening. That mainly costs you time.

This shows you whether people order a second time. You also learn the exceptions you would otherwise only discover after the build: who cancels, who doesn’t pick up their loaf, what happens when the baker closes for a day. Those exceptions are exactly what someone building the app wants to know, as you can read in You have an idea for an app. What do you do first?

What to do with the outcome

A few conversations and a page give you a direction, and you don’t get more certainty than that yet. If nobody recognises the problem, you change the idea or drop it, and that will have cost you little. If people recognise the problem but get stuck in your prototype, the design has to change. If they sign up and order a second time, you can choose what goes in a first version.

Take your notes to whoever is going to build it. What people said word for word and where they got stuck can go straight into a brief.