The number-one reason startups fail is building a product nobody wants. That's why the first step is idea validation. Talk to 100 potential customers and find out if they'd actually pay. "Build the product, then sell it" is the old, risky approach.
MVP (Minimum Viable Product) means testing the market with minimal features. Even a landing page plus a waitlist can be an MVP. The first 10 customers are found by hand — not through marketing, but through direct conversation. Get feedback, improve the product, then scale. Paul Graham: "Do things that don't scale" — in the beginning, deep relationships beat automation.
Our courses on this topic