Look Before You Leap
The cheapest test that still answers the question
You don't need to build the whole thing to find out if anyone wants it. You need the smallest, cheapest test that still gives a real yes or no.
The instinct is to build the whole thing in secret, then reveal it fully formed. That is the most expensive possible way to discover you were wrong.
A minimum viable test is not the cheapest thing you can make — it's the fastest full loop that answers your riskiest question. Climb only as high up the ladder as the question requires, and no higher.
- Landing page / fake door: describe the offer, add a 'buy' or 'reserve' button, and count who actually clicks.
- Video: if the value is hard to explain, show it working in a short demo and measure the sign-ups.
- Concierge: deliver the service by hand for a few real customers, with no systems yet.
- Wizard-of-Oz: the customer thinks it's automated; behind the curtain, it's you doing it manually.
- Single-feature build: build only the one thing that answers the question — nothing else.
The very day they announced the idea — before any device existed — the founders let people preorder from a description and a few renderings, collecting real names, addresses, and verified credit cards, but charging nothing until it could ship.
→ Orders poured in. One month later investors put in $2 million, and a year later the first real Fitbit shipped. Real preorders proved the demand before the product was built.
The honest condition: this only works if you can actually deliver later. Taking money for something you can't build is fraud, not a test — always keep an honest, no-charge exit if plans fall through.
You think there's demand for a meal-prep service. Don't build a kitchen. Put up a one-page site with three plans and a 'reserve for next week' button. Ten paid reservations means real demand; zero clicks just saved your savings.
Apathy is the real killer, not criticism. Harsh feedback means people care; silence — no clicks, no orders, polite shrugs — means you have no business yet. Fear a quiet test more than an angry one.
A demo test proves demand, not that you can build it. Dropbox's short video drove its waitlist from about 5,000 to 75,000 overnight — but Drew Houston already had a working prototype, so his test measured wanting, not feasibility. Know which question your test actually answers.
Pick the smallest, most embarrassing test that can still give a real yes or no, run it fast, and let a quiet result kill a bad idea for a few hundred — instead of your life savings.
Choose the lowest rung of the ladder that answers your killer assumption, and set it live this week — a page, a post, or one hand-served customer. Give it a number it must hit.
Look Before You Leap