You've got an idea — how to pressure-test it before you write code

The frame An idea is a guess about a problem. Before you spend three months building, you want the guess to survive contact with the people who supposedly have the problem. That takes about two weeks and no code. The mistake is asking people whether they like your idea. They will say yes to be kind, and you will hear a market that isn't there. Talk about their life instead — what they did last time this came up, what it cost them, what they tried. Facts about the past predict behaviour; opinions about the future don't. The two-week test · List the people who have this problem now. Not "everyone" — the ten you could actually message this week. · Talk to at least five. Ask what they do about it today, how often, and what the workaround costs them in time or money. Let them talk; you're there to learn, not to pitch. · Watch for the pull: someone asking when they can use it, introducing you to a colleague, or trying to pay before it exists. That is the only strong signal — politeness is noise. · Write the one-liner: [who] struggles to [do what] because [why now]. If you can't fill the blanks from what you heard, keep talking. Reading list · The Mom Test — the short book on how to ask so people can't flatter you into a bad decision. · Y Combinator Startup School — a free course; the early lessons are on exactly this. · "Schlep Blindness" by Paul Graham — on why the best ideas look like too much work, so everyone skips them. Go / no-go Green: several people have the problem badly, already do something painful about it, and lean in when you describe a fix. Amber: real problem, weak urgency — narrow to the segment that feels it most. Red: everyone is "interested" and nobody has changed anything — the idea isn't disproven, it's just unproven. Go find a hair-on-fire user, or a different idea.