The 7-day kit for launching your business

Start without the AI. Write three lists of ten lines, by hand if you can: what you do better than most people around you; the industries you know from the inside, even from a summer job; the things that genuinely annoyed you at work over the last three years.

The third list is the valuable one. A business idea is almost always an annoyance somebody else shares — and already pays to make go away.

Only then do you cross the three lists with the AI.

You are a product strategist helping a solo founder pick a first sellable offer.

Here are three lists about me.
SKILLS: [paste your list]
INDUSTRIES I KNOW FROM THE INSIDE: [paste your list]
WHAT ANNOYS ME ABOUT THOSE INDUSTRIES: [paste your list]

Give me 8 offer ideas at the intersection of these three lists.
For each one, one line each:
- exactly who it is for (a job title, a company size);
- the concrete problem, phrased the way that person would say it themselves;
- what those people do about it today, and what that costs them;
- who already sells something comparable.

Rule out the ideas where nobody spends money today, and tell me why you ruled them out.

Keep three ideas. One rule to choose between them: is somebody already spending money on this problem — on a tool, a contractor, an employee’s hours? If the answer is no, that’s not an untapped market, that’s an absent one.

This is the day everyone skips, and the only one that really decides what follows. You are not selling anything today. You are listening.

Find ten people living with the problem: LinkedIn, industry groups, trade forums, your own contacts. Write them something short that offers nothing.

I'm preparing 5 discovery conversations with [describe the person precisely:
job title, company size, city].
I want to test this problem: [describe the problem in one sentence].

1. Write me a first-contact message, 4 lines maximum, that sells nothing and asks
   for 15 minutes.
2. Write me 6 interview questions about what the person HAS ALREADY DONE,
   never about what they would do.
3. For each question, tell me which answer should make me drop the idea.

Banned: "would you use", "would you pay", "what do you think of".
Rewrite everything in the past tense.

During the call you are after four things: the last time the problem came up, what they did about it then, what it cost them in money or hours, and what they use today to cope.

An offer that fits on one page sells. An offer that needs three pages gets negotiated.

1. Who it’s for — a job, a size, a situation. Not “small businesses”.
2. Which problem — the sentence from day 2, in their words.
3. What they get — an outcome, not a feature list.
4. How it works — three steps, no more.
5. How much — a stated price, even a provisional one.
6. Why you — one sentence, true and checkable.

For the price, don’t start from the time it takes you: start from what the problem costs your client. If you save two days a month for somebody who bills £400 a day, you are not selling three hours of your time.

One page. Genuinely one: the promise at the top, who it’s for, how it works in three steps, one piece of proof, one button. One button — two different buttons is one more decision asked of somebody who just met you.

Which tool depends on what you sell. Selling a service means you need a site: Framer is faster and looks better. If the product is the software, Lovable or Replit give you a real application in a day, even if you rewrite it later.

The rule is three numbers: two colours, one typeface, three assets.

• One main colour, one dark text colour. That’s it.
• One typeface, in two weights. Pairings show, and they go wrong.
• A logo written as text. A symbol comes later, once you know who you are.
• Three reusable assets: a banner, a quote background, a cover.
• The same profile picture everywhere. That’s what people recognise, not the logo.

This is where most launches stop. The product exists, the page is live, the identity is clean — and nothing happens, because nobody knows it’s there. The five previous days are only worth what this one makes of them.

Thirty isn’t a modest number out of caution: it’s the point at which silence starts to mean something. Below it you learn nothing. Above it you can no longer personalise.

Every name has to pass three questions: does this person have the problem from day 2? Do they have the authority to say yes? Can I write, in one line, why I’m contacting them? If the third question has no answer, the name goes.

That’s the work Lihn does: you describe what you offer and to whom, it looks for the companies that match, and it tells you why each one was kept. You keep control of the list — you’re the one who removes names. It isn’t the tool that makes your offer good; it’s the one that stops you building a business without ever going to find customers.

The message is four lines. One line showing you looked at who’s on the other side. One line on the problem. One line on what you do. One question at the end — a single one, easy to answer yes or no.

What brings replies isn’t writing talent: it’s the follow-up. A first message, a follow-up at three days, a last one at seven. Most replies land on the second. That is exactly the part nobody keeps up by hand, and the part a tool keeps up perfectly.

Three rules not to negotiate, before you press send:

• Write from a professional address on your own domain. A personal inbox has neither the reputation nor the right to send commercial mail.
• In B2B, you can contact somebody at their work address if your offer relates to their job. That’s the boundary: if you have to strain to explain the connection, you’re outside it.
• Every message says who you are and lets people opt out in one click. That isn’t a formality, it’s what keeps you out of spam folders.