The order matters because each step removes a reason to spend money you have not earned back yet. You name the person before the product, prove demand with manual work before automation, and let five real reactions rather than your own attachment decide the final call. Building an offer or a brand before step 2 usually produces something polished that nobody asked for.
Step 1: Name one person, one problem
Begin with one sentence that names a specific person and the specific problem you would help them with, written the way they would say it, not the way a business plan would. "Freelance illustrators who lose track of which client approved which revision" beats "creative professionals seeking workflow optimization" because it points at a real moment of frustration you can respond to. Write it, then read it aloud and ask whether you could find ten of these people this week. The usual failure here is vagueness dressed up as ambition: a sentence so broad that every later step becomes guesswork. If you cannot name the person, you cannot describe their problem, and you cannot tell later whether anyone actually wanted what you made.
Step 2: Define the manual deliverable
Take the sentence from step 1 and ask what you could hand that person this week using nothing but your own time and tools you already have. A useful result is something they could act on or show someone else: a short list of three vendors, a rewritten paragraph, a one-page plan, a 20-minute call with notes. Write the deliverable as a noun, not a promise, so it reads "a 5-item sourcing list" rather than "help with sourcing." Cut the scope until it feels almost too small, because the point is to learn whether the outcome matters before you invest in producing it at scale. What people skip here is the word "manually" — they jump straight to a tool, an app, or a template, which hides whether the value is in the result or in the delivery mechanism.
Step 3: Build the offer with free tools
With the deliverable named in step 2, build the smallest container that can hold an offer: a one-page doc, a two-field form, a plain post, or a free scheduling link. It needs four things and nothing else — who it is for, what they get, what you need from them, and how to say yes. Use wording close to your step 1 sentence so the offer and the audience match. Design, logos, and a domain can wait, since a free form with a clear question outperforms a beautiful landing page nobody visits. The common slip is polishing the wrapper before checking the contents: three hours on font choices, ten minutes on whether the offer is comprehensible to a stranger. Read it as if you had never heard of the idea, then rewrite anything that requires explanation.
Step 4: Take it to five people
Five is enough to start seeing patterns and small enough to do today. Choose people who already have the problem — a forum thread where it is being discussed, a Slack group, a Discord server, five individuals you can message directly — and say plainly what you are doing: "I am testing whether this is useful, here is what I would make you, would you want it?" Admitting it is an experiment is not a weakness, because it lowers the stakes for them and makes honest answers more likely. Do not inflate numbers, fake testimonials, or hint at a team you do not have, since false signals now will send you down the wrong path at step 6. What most people get wrong is broadcasting to a general audience instead of asking five specific people, which produces polite silence rather than usable feedback.
Step 5: Record questions and confusion
After the five conversations, write down four columns: what they asked, what they ignored, what they misunderstood, and what they wanted changed. Questions reveal missing information, silence on a section tells you it does not matter, misunderstanding usually means your step 3 wording failed rather than that they are slow, and requested changes are the closest thing to a roadmap you will get for free. Capture their words, not your summary of them — "can it cover weekends?" is data, "interested in flexibility" is not. Do this within an hour of each conversation, while the specifics are still intact. Skipping the write-up is what turns a week of real feedback into a vague feeling that it went well, which is exactly the state to avoid before deciding anything.
Step 6: Decide on evidence, not attachment
Now compare the log from step 5 against a simple bar you set in advance: for instance, at least two of five people ask to actually receive the deliverable, or agree to pay something small for it. Two yeses means improve, with the same person, same problem, and a clearer offer. Repeated interest in a different problem means pivot, and you go back to step 1 with a new sentence. Zero traction across five conversations means stop, and that costs you a few days instead of an inventory order. Set the bar before you look at the notes, because attachment grows with every hour you invest and will quietly move the goalposts. The failure mode is treating a polite "this is interesting" as a signal, then spending money to satisfy a demand that was never there.
A phone reminder about this experiment disappears the moment you dismiss it and leaves no trace of what you decided. A paper strip taped above your desk gets visibly shorter as you cut off each step, and it stays there until the last one is gone. The order is the whole point: person, deliverable, offer, five people, notes, decision.