Build · Lesson 1 of 2

Start from a real problem

Talk to real people about what they have actually done before you build.

5 min read · Free — no account needed to read

Many projects fail not because they were built badly, but because nobody needed them. The fastest way to avoid that is to talk to the people you want to help before building anything.

Ask about their past, not their opinions: "When did this last happen? What did you do? What did it cost you?" Past behaviour tells you far more than "would you use this?"

AI can summarise interviews and suggest patterns. Sitting across from a person and noticing the moment their voice changes is still your job — and that is where the real problem usually shows up.

See it happen

You are convinced that a shared calendar would fix the chaos in your community group.

The read that arrives on its own

Everyone complains about scheduling, so build the calendar.

The same situation, with this idea

You ask four members what the most annoying part of their week in the group is. Three describe the same thing: they never find out when something has moved. Nobody mentions scheduling. The group already has a calendar and nobody opens it.

What changed

The first version generalised from a complaint you had already interpreted. Asking what people actually did that week, rather than what they wanted, put a different problem in front of you — and a considerably cheaper one to solve.

The takeaway

Talk to real people about what they have actually done before you build.

Easy to take the wrong way

Asking people what they would like instead of what they did. "Would you use this?" reliably produces a yes that means nothing. "What did you do the last time this happened?" produces facts you can build on.

Try this

Ask one person: "What is the most annoying part of your week?" and just listen.