The most expensive sentence in startups is “let’s build it and see.” Building is the slowest, priciest way to learn whether anyone wants your product, and the market can usually tell you for a fraction of the cost, before a line of product code exists. Here are five tests, ordered by cost, and the signals worth trusting.
1. The problem interview (costs: conversations)
Before testing your solution, test the problem. Talk to ten people in your target market, and ask about their current behavior, not your idea: how do they handle this today, what have they tried, what does it cost them. The trap: asking “would you use an app that…”, people say yes to hypotheticals to be nice. Past behavior is evidence; polite enthusiasm is not.
2. The landing-page test (costs: a page and small ads)
Describe the product as if it exists, one clear page with the promise, the price, and a call to action, then send a little targeted traffic. You’re measuring whether the promise pulls, before the product exists. Signal that matters: visitors who leave an email, or better, click a priced “get started” button. A conversion rate in the low single digits from cold traffic is genuinely interesting; a hundred likes on the announcement post is not.
3. The concierge test (costs: your time)
Deliver the service manually to a handful of customers, you are the algorithm, the spreadsheet is the backend, WhatsApp is the interface. A meal-plan app becomes you writing meal plans for five paying customers. What it buys you: you learn the real workflow, the edge cases, and whether the outcome is worth paying for, knowledge that later makes the actual build dramatically more accurate.
4. The wizard-of-oz test (costs: a thin front)
Like concierge, but the customer sees what looks like a product while humans run the backstage. It tests the product experience rather than the service outcome, do users come back to the interface, or did they just like being served by a human?
5. The presale (costs: courage)
The strongest signal in existence: money before the product. A discounted founding-member offer, a refundable deposit, a signed letter of intent for B2B. Ten presales beat a thousand survey responses, because paying is the only answer that can’t be polite.
Reading the signals honestly
- Strong: payments, deposits, emails from cold traffic, people using a manual version repeatedly.
- Weak: likes, encouraging comments from friends, “I’d definitely use this,” survey results.
- The pivot rule: if two tests in a row produce weak signals, change the promise or the audience before changing the amount you’re spending.
Then build, small
Validation doesn’t replace the MVP, it aims it. Once a test produces strong signals, you know exactly which core action the product must nail, which is what keeps scope, cost, and timeline sane, we covered those numbers in how much an MVP really costs, and how to choose who builds it in our agency vetting guide.
Have signals already and want the build scoped? The first workshop is free, bring your validation data and leave with a plan.