The Real First Steps of Starting a Business
I have an idea. What do I actually do first?
Write your idea as one sentence naming a specific person and a specific painful moment. Then have ten conversations with that person about the last time it happened. Then ask one of them for money — a deposit, a pilot fee, a pre-order. Everything else (name, logo, entity, website, product) comes after somebody has tried to pay you.
Most people start a business backwards. They begin with the artefacts — the name, the logo, the company registration, the website — because artefacts feel like progress and conversations feel like risk. Six months later they have a beautiful shell and no evidence.
The order that works is unglamorous and has not changed in three hundred years: find the person, confirm the pain, take the money, then build the thing. A market trader in 1740 did it by standing in a square and watching what people picked up. You have better tools and the same job.
The four moves, in order
- 01Name the person. Not 'small businesses'. A dental practice owner with two chairs and no office manager. Specific enough that you could list ten of them by name.
- 02Find the painful moment. Not the category of pain — the moment. 'The Friday afternoon where she reconciles card payments by hand and always finds a gap.' If you cannot describe the moment, you do not understand the problem.
- 03Ask for money before you build. A €200 deposit, a paid pilot, a pre-order, a signed letter of intent with a number on it. This is the only validation that cannot be faked by politeness.
- 04Build the smallest thing that resolves that Friday afternoon. One person, one job, one price.
What the first ten conversations should sound like
- ·You talk for a third of the time, maximum.
- ·Every question is about the past: 'walk me through the last time', 'what did you do instead', 'what did that cost you'.
- ·You never describe your solution until the last five minutes, and then only as a question.
- ·You end with: 'who else should I speak to?' — a warm intro is a stronger signal than any compliment.
Ten of these takes about two weeks around a job. It will feel slow. It is the fastest thing you will ever do, because it removes the twelve months you were about to spend building the wrong thing.
This is industry-agnostic
| Business | First step that isn't building | The paid signal |
|---|---|---|
| Software tool | Ten calls with one job title | A paid pilot or annual pre-pay |
| Agency / consulting | Three scoped conversations | One paid diagnostic engagement |
| Food / retail | A market stall or pop-up weekend | Cash through the till on day one |
| Physical product | Twenty conversations plus a rendering | Pre-orders or a crowdfunding floor |
| Creator business | Publish for six weeks before selling | First 20 paid subscribers |
The first milestone is not a product. It is a stranger trying to give you money for something that does not exist yet.
Do this next
- Write the idea in one sentence naming a person and a moment.20 minutes
- List ten real people who live that moment. Names, not segments.45 minutes
- Book five conversations for this week.1 hour
- Draft the paid ask — deposit, pilot fee, or pre-order — with a number.30 minutes
Avoid
- —Registering a company before you have a paying signal.
- —Buying a domain, logo or brand kit in week one.
- —Surveys. They measure politeness, not demand.
- —Building a prototype 'just to show people' — it converts the conversation into a demo and kills the information.
“Turn my idea into a two-week first-steps plan: the exact person, the painful moment, ten names to contact, five interview questions, and the paid ask I should make.”
Ask Aury thisLast edited 2026-07-28