Validate a business idea · Built at Google, taught at Stanford

How to validate a business idea before you build it

Most ideas fail because nobody wanted them — not because they were built badly. Here's how to get real customer evidence in days, using pretotyping, before you spend a dollar building.

4,000+ experiments run · $30M+ saved across 50+ teams

Banked

$68M+

$38M generated $30M saved

Fund the best, save on the rest.

Definition · Answer first

What does it mean to validate a business idea?

To validate a business idea is to gather real evidence that people will actually pay for it — before you build it. The fastest way is pretotyping: a small, real-world experiment that puts your idea in front of real customers and measures what they do, not what they say. If the data shows genuine demand, build. If it doesn't, you've saved the cost of building the wrong thing — the way AGL, PEXA and Tabcorp have saved $30M+ by killing weak ideas early.

80%+

of new products fail in the market — most for lack of real demand

Nielsen / Harvard Business School

$30M+

saved across 50+ enterprise teams that validate before building

Exponentially client results

4,000+

validation experiments run to date

Exponentially

AI idea validators vs real validation

AI can guess your odds. Real validation proves them.

A wave of AI ‘idea validator’ tools will score your idea in 120 seconds. Useful for a gut-check — but a model has never put your idea in front of your customers. These answer different questions. See how pretotyping produces the evidence →

DimensionReal validation (pretotyping)AI idea validators
What it measuresReal customer behaviour — clicks, sign-ups, money on the tableA language model's guess about the idea you typed in
Who's tested onYour actual target market, from real trafficNobody — no real customers are ever involved
Time to a signalDays — run a fake door or concierge test this weekSeconds — but it's an opinion, not evidence
What you getEvidence a board will actually fund againstA score or summary, with no proof of demand
Best used forDeciding whether to commit real money to an ideaA quick directional gut-check while brainstorming
The process

How to validate a business idea in 4 steps

Set your pass/fail threshold up front — before you see the result — so the test stays honest. Deep dive: idea validation, step by step →

01

State a testable hypothesis

Turn “I think people want X” into a falsifiable claim naming the exact audience and action. Vague ideas can't be tested; sharp ones can.

02

Set your XYZ pass/fail threshold

Use Savoia's format: “at least X% of Y will do Z.” Decide what result would kill the idea before you run — so weak data can't be rationalised later.

03

Run the cheapest pretotype

Don't build the product — fake it. Pick the method that produces a real demand signal fastest for the least build, and put it in front of real customers.

04

Decide on the evidence

Compare the result to your threshold and act: fund the idea, iterate the test, or kill it and free the budget. Real signal, not opinions.

FAQ

How to validate a business idea, answered

How do I validate a business idea?

Put the idea in front of real customers as if it already existed and measure what they do. Write it as a testable hypothesis, set a pass/fail threshold in advance (at least X% of Y will do Z), run the cheapest pretotype — such as a fake door test — on your target market, then decide to fund, pivot, or kill the idea based on the evidence.

What's the difference between validating and prototyping an idea?

Validation asks ‘should we build this at all?’ — you test demand before building. Prototyping asks ‘how should we build it?’ — you refine a solution you've already committed to. Pretotyping comes first: it validates the idea with a real-world experiment before any prototype exists.

How long does it take to validate a business idea?

With pretotyping, days — not months. A fake door or concierge test can produce real demand data within a week, because you're faking the product and measuring behaviour rather than building anything.

How do I validate an idea without spending money?

Run the smallest real experiment that produces genuine signal — a landing page with a real sign-up (fake door), or manually delivering the outcome to a few customers (concierge). You measure real actions, not opinions, at near-zero cost.

Can AI validate my business idea?

AI idea validators can estimate an idea's odds quickly by pattern-matching, which is useful for a first gut-check. But they can't test your market. Real validation requires a live experiment with real customers — that's the only evidence strong enough to fund a build decision on.

Bring us one idea. We'll tell you if it's worth building.

Twenty minutes with Leslie Barry. We look at one idea on your roadmap, the evidence you already have, and design the test that would validate it. No pitch.