Most founders do not start with a clear product plan. They start with a problem that keeps showing up. The job is not to pick the perfect feature. The job is to find the smallest thing you can build that teaches you something.
1. Start with the problem, not the idea
Write down the problem in one sentence: "People who ___ have trouble doing ___." If you cannot describe it in terms of a current struggle, you are describing a solution, not a problem. Solutions are expensive guesses. Problems are cheaper to validate.
2. Find five people living it
You do not need a hundred users. You need five people who have the problem right now. Talk to them about what they are doing today, what they have already tried, and what they have given up on. You are looking for language you can reuse and a pain strong enough to pay for.
3. Map the smallest possible first step
Ask yourself: "What is the smallest thing I could build that would let a real person make progress?" Not your whole vision. One job. One workflow. One screen. The smaller the first step, the faster you learn.
4. Decide what must be true for this to work
Before you build, list the assumptions. What do you need to believe about users, behavior, and willingness for this to be worth building? Pick the riskiest one and test it first. Interviews, prototypes, and fake-door tests are all cheaper than shipping code.
5. Build, measure, then decide
Build the small thing, put it in front of users, and watch what happens. Did they understand it? Did they finish? Did they come back? The goal is not to prove you were right. The goal is to learn whether to keep going, change course, or stop.
You don't have to do this alone
This is the kind of work product people are trained for. If you are stuck between an idea and a first feature, you can reserve hours from the pool and talk through it with someone who has done it before.