Most products don't fail because they're built badly. They fail because the wrong thing got built well. Validation is how you avoid that — cheaply, before you commit months of engineering.
Start with the problem, not the feature
Write down the specific job your user is trying to get done. If you can't state it in one sentence without naming your product, you're not ready to build yet.
Talk to ten real users
Five to ten focused interviews will tell you more than weeks of internal debate. Ask about what they do today, what's painful, and what they've tried — not whether they "like" your idea.
Run the smallest possible experiment
- Landing page test — describe the value, measure who signs up.
- Concierge MVP — deliver the outcome manually before automating it.
- Prototype test — put a clickable prototype in front of users and watch where they get stuck.
Decide with a metric, not a feeling
Define success before you run the test: "20% of visitors join the waitlist," or "7 of 10 users complete the core flow unaided." If you hit it, build. If not, change the idea — not the goal.
Then build the smallest thing that proves value
Ship the narrowest version that delivers the core outcome, instrument it from day one, and let real usage guide what comes next.
Validation isn't a phase you skip when you're confident — it's the cheapest insurance you'll ever buy. If you want a second set of eyes on your idea, tell us about it.