The validation method
How to validate a startup idea before you build it.
AI has made building cheaper than ever — which means it has made building the wrong thing cheaper than ever. Here is the exact method we use to find out whether an idea is real before a line of code gets written, and the engine we built to run it.
Why it matters
The dominant reason startups die is building the wrong thing.
Roughly 42% of failed startups cite no market need as the primary cause — not bad engineering, bad luck, or thin capital. As building gets cheaper, it becomes easier to build the wrong thing faster, and at greater scale.
The math has changed but the principle hasn't: a week of validation now saves about five weeks of building the wrong thing. The point of the first build isn't to launch — it's to find out.
The method
Four questions, answered before you build.
Is the problem real?
Not 'would people use this if I built it?' but 'is this problem painful enough that they're already trying to solve it?' If the answer is no, nothing else matters.
Who has it most acutely?
The specific segment whose pain is sharpest and whose willingness to pay is highest. Most failed products have a real problem but aim at the wrong target.
What would they pay?
Not vague intent — real signal. A deposit, a pre-order, a credit card, a signed pilot. Soft signals are noise; hard signals are the truth.
What's the smallest test?
Not the smallest version of the final product — the smallest experiment that returns a binary result. A landing page, a concierge service, a single sales call with a paper prototype.
The method, in practice
We built an engine to run this method: Smokescreen.
The principle of validation is free; the execution is what teams skip. Assembling a landing page, copy, tracking, a waitlist, traffic, and analysis used to burn two weeks. xlabs Smokescreen crushes that cost — what took three to four weeks of founder time now takes under thirty minutes.
Landing page + hard-signal waitlist
From a one-line idea to a live, hosted page in minutes — with a waitlist that captures willingness to pay, urgency, and what they use today, not just an email.
Dashboard with a pre-set bar
Signups, conversion, traffic sources, and themed feedback in one view — against a success threshold defined before the test goes live. No moving goalposts.
Explore + Monitor
Once there's signal, deep strategic reports tell you what to build and who to sell to, and continuous monitoring tracks the market so a single validation never goes stale.
Reading the result
Hard signal, then a real decision.
A waitlist that asks for an email collects emails. A waitlist that asks how much someone would pay, how soon they need it, and what they use today collects evidence about whether the idea is real. They cost the same to build and produce very different qualities of insight.
Then comes the hardest discipline: the decision. Either the signal clears the bar you set — go build. Or it doesn't — pivot or kill. Most teams gather weak signal and rationalise it. A clean no is as valuable as a clean yes; both are answers you can act on.
Validation FAQ