The order is the whole method: choose before you spend, spend before you sell, sell before you judge. Each step produces the input the next one needs, so the $20 buys evidence rather than enthusiasm. Skip ahead and you end up deciding from memory instead of notes.
Step 1: Pick one idea you can start
Order matters here: the pick comes before any spending, because a $20 budget buys a test of one idea, not a survey of four. Choose something deliverable within seven days using tools you already own, with a buyer you can name out loud today. Mowing works if you own a mower and can point at three neighbours with long grass; tutoring works if you can name two families; a market stall works if a pitch is free this Saturday. Write it as one sentence including who pays and how much. What derails this step is choosing the idea you are most excited about rather than the one with a reachable buyer, which sends the next $20 into a product nobody asked for.
Step 2: Cap the test at twenty dollars
With one idea fixed, the $20 exists to buy evidence, not infrastructure. Before spending, name the single question the money answers: will strangers stop at this stall, will anyone pay $18 for this, will ten parents reply. Then buy only what a paying customer would receive or see. That might be one sample you make yourself, $10 of printed flyers for a Saturday pitch, or a $12 ad targeted at five postcodes. Keep every receipt and note the date, since step 4 asks what the test actually cost. Where people slip is buying the fun parts first: a logo, a domain, branded packaging. Those feel like progress and generate zero information about demand, so by step 5 you are deciding on a feeling instead of a result.
Step 3: Get one paying stranger
Evidence beats enthusiasm, and this step produces it. You have something physical or bookable now thanks to step 2, so go ask ten specific people directly rather than posting and waiting. In person or by message, one line: I am doing this on Saturday for this price, do you want it. A real customer is someone who hands over money or receives a delivery you can name. Free counts only if they state what they would have paid, and that number goes in your notes as the first price test. Likes, group-chat encouragement, or a relative's polite yes are the classic substitution, and it feels so much like traction that people skip ahead to celebrating. One stranger's yes outweighs fifty warm maybes.
Step 4: Write five lines of notes
Five lines, written within a day, while the details are still exact. The constraint is the point: a fixed shape forces facts in and feelings out. Line one, what you sold and at what price. Line two, who paid and in what quantity. Line three, who said no and their exact reason in their own words, which later becomes your ad copy. Line four, total spent, including the $20 and any hours you would resent repeating. Line five, the one thing that surprised you. People commonly turn this into a long diary of how the week felt, or skip it when the result was a flop, which means step 5 gets settled by the loudest memory instead of the numbers. Two rejections saying too expensive is a different finding from two saying not for me.
Step 5: Decide before the week ends
Three outcomes, and the notes from step 4 pick between them. Double down if a stranger paid full price and you can repeat the delivery in under an hour, since that combination means the money scales without eating your week. Pivot if buyers kept asking for an adjacent thing: a different size, a weekend slot, a bundle. That is demand pointing sideways, and it is the most common useful result. Drop it if you had to chase every yes, or if the total came in at a loss price cannot fix. Dropping is a win because $20 bought a real answer, and the only genuine failure is arriving here with no notes to read. Watch for the sunk-cost reflex, where money already spent pushes you to continue on an idea the evidence has closed.
A phone reminder for this experiment vanishes the moment you dismiss it, leaving no trace of whether you ever asked anyone. A paper strip taped where you work visibly gets shorter as each line is cut off, and it stays there until the last one is gone. Decide once, on paper, and the $20 buys an answer instead of an intention.