Most first products do not fail because they were built badly. They fail because they were built for nobody in particular, and nobody noticed for six months. A 30-day plan forces the question early: does someone care enough about this to use it, and then pay for it?
This is a plan for a solo founder who cannot code. It does not promise customers; no honest plan can. It promises a clear answer in a month. It assumes you build with AI tools or an AI team, so the building is a few days, not a few months.
The short answer. Week 1: talk to people and write one sentence of proof. Week 2: build the smallest version. Week 3: launch to ten people, personally. Week 4: fix the biggest drop-off and ask for the first payment.
Before day 1: choose a narrow target
Pick a specific person with a specific problem you can reach. “Small business owners” is not specific. “Solo yoga teachers in my city who take bookings by text” is. You need to be able to name ten of them, or know where they gather.
Write the proof sentence you will judge yourself by: “Ten teachers use a booking page for two weeks and at least five say they would pay.” We explain why this sentence matters in how to launch a real product without code.
Week 1 (days 1 to 7): listen, then commit to one problem
Goal: know the problem in your customer’s own words.
- Days 1 to 3: reach ten people who match your target. Message them personally. Ask for fifteen minutes about how they handle the problem today, not whether they like your idea.
- Days 4 to 5: in each conversation ask: what did you do the last time this happened? What did you try? What does it cost you? Do not pitch. Write down their words exactly.
- Days 6 to 7: find the repeated pain. Pick one problem. Write the one-line offer and a list of what you will not build.
Done when: you can say “X people told me Y”, with real names in a private note, and you have a one-line promise.
A caution: friends and family are polite. Their enthusiasm is not evidence. People who have the problem and are already paying for a workaround are.
Week 2 (days 8 to 14): build the smallest version
Goal: a working product that does the one thing, on its own address.
- Days 8 to 9: describe the one flow to your AI tool or the PM: who it is for, the one thing they do, and what counts as success. Ask for the edge cases.
- Days 10 to 12: the build. Look at the preview on your phone every day. Cut anything that is not the one flow.
- Days 13 to 14: put it live on a real address with a real database, test it as a stranger, and add analytics that count visits, the main action and returns.
Done when: a person you have never helped can complete the main task on their phone.
Small products are genuinely small. The example products on equipe.im, a booking page for a tutor among them, were each built by an equipe team in roughly 35 to 45 minutes. They are examples with made-up businesses, and a real customer product takes longer because it has real requirements, but the first version should still fit in days.
Week 3 (days 15 to 21): launch to ten, personally
Goal: real people use it, and you watch.
- Day 15: message the ten people from week 1. Do not post publicly yet. Send the link and ask them to try one specific task.
- Days 16 to 19: watch the numbers daily: visits, the main action, who returned. Ask every user one question: what confused you?
- Days 20 to 21: list what you learned. Choose the single biggest drop-off.
Done when: at least a few people have completed the main action without your help, and you know where the others gave up.
If nobody finishes the main task, that is not a failure. It is the most valuable thing you could learn in week 3. Change one thing and ask again. The loop is described in build, measure, learn on autopilot.
Week 4 (days 22 to 30): improve one thing, then ask for money
Goal: a clear answer on whether people will pay.
- Days 22 to 24: fix the biggest drop-off. One change, shipped behind your approval.
- Days 25 to 27: ask your active users directly: “If this cost $X a month, would you pay for it?” Name a price. Better still, offer a pre-order or a simple way to pay.
- Days 28 to 30: count the answers honestly. Decide: continue, change the target or the offer, or stop.
Done when: you have a number you trust: how many people used it, how many came back, how many would pay or did.
Charging is a feature. Taking a first payment, even a small one, teaches you more than a hundred compliments. Use a payment provider you trust, keep payment keys in a secret store and never in the code or a chat, and have the team ask your approval before anything that touches money. More on that in is it safe?.
What it might cost
If you run this with equipe, the starting credit ($5, and $10 more with a card, no charge) and up to 2 hours of team time are included in the free start. A live product on its own can be kept up for $19 a month with Hosting only, and the Team plan is $19 a month or $190 a year for 20 hours of team time, plus credits for AI work. Every request shows its expected cost first, and you set a monthly cap. Details are on the pricing page and in our cost breakdown. Domains are billed at cost.
Common traps
- Building before listening. The fastest way to waste a week.
- Counting compliments as demand. Count commitments instead.
- Adding features after a quiet week 3. If people did not use the first flow, a second flow will not help; fix the first.
- Launching big. A public launch with nothing learned burns your best chance of a first impression.
- Hiding the price. If you do not ask, you do not know.
What to do with the result
After day 30, you have one of three outcomes, and each is a good result:
- People pay or clearly would. Keep the weekly loop going and widen the audience slowly.
- People like it but would not pay. Change the offer or the target. The product may be a feature, not a business.
- Nobody uses it. You found out cheaply. Pick a new problem from your week 1 notes.
When you are ready to build, you can tell the PM your idea and your proof sentence and start free. If you built something earlier in Claude, see Turn a Claude artifact into a live app. To see who on the team does what, read what an AI startup team actually does.