How I validated an idea before writing code

I almost built a full analytics dashboard because three friends said it sounded cool. Instead I made a Typeform with five questions, posted it in two Slack communities, and offered a 15-minute call to anyone who completed it. Twelve people responded in four days. Only four had the problem urgently enough to pay. That was enough signal. I scoped a tiny v1: one report, CSV export, email delivery on a schedule. Shipped in ten days instead of three months. If I had coded first, I would have polished charts nobody wanted. The form took an hour and saved a quarter.