Place a button, link, or landing page for a feature that doesn't exist yet, and measure who tries to use it. The cheapest honest test of demand — behaviour, not opinion.
▸ Try the interactive toolFake Door Testing is a demand-validation method in which a team places an entry point for a feature that doesn't yet exist — a button, a menu item, a landing page, an in-product card — and measures how many users try to use it. Clicking the fake door leads to a short, honest message explaining the feature isn't available yet, often with a way to register interest.
Its power is that it measures behaviour, not opinion. Asking “would you use this?” gets a flattering, unreliable yes; counting how many people actually click to use it gets a real demand signal at almost no cost. It tests the value/demand risk — the one most likely to kill a feature — before any of it is built. The ethical key is the honest follow-through: the door must lead somewhere truthful, never a deception that wastes or misleads the user.
Place a real-looking entry point (button/link/page) for an unbuilt feature → measure click-through (the demand signal) → on click, show an honest “coming soon / register interest” message. Behaviour, measured cheaply, before building.
The deadliest feature risk is that nobody wants it — and the cheapest way to test demand is to watch whether people reach for it, not to ask whether they would.
| The shortcut | What it costs | What it gives you instead |
|---|---|---|
| Asking about demand | “Would you use this?” gets a polite, inflated yes. | Click-through measures real behaviour, not stated intent. |
| Building to test demand | Shipping a feature to see if it's wanted is the costliest test. | A fake door tests demand before any feature is built. |
| Survey-based demand signals | Stated-preference surveys over-predict adoption. | Behavioural click data is a far more honest signal. |
| Late demand discovery | Finding out nobody wanted it after launch. | The fake door surfaces weak demand in days, pre-build. |
State exactly what level of click-through would count as real demand, in advance (a testable hypothesis, Tool 13). Without a pre-set bar, any result will be rationalised.
Add a genuine-looking door — button, card, landing page — where a real user would encounter the feature. It must look real enough to attract real intent, not a survey in disguise.
On click, show a truthful message: the feature is in development / coming soon, with an option to register interest. Never deceive — the honesty is both ethical and a second signal (sign-ups).
Count how many users who saw the door clicked it, relative to the pre-set success criterion. This is the demand signal; the sign-up rate is a bonus.
Strong demand justifies building; weak demand saves you from building something unwanted. Either outcome is a win — you learned cheaply.
A team was confident a requested feature would be popular — users had asked for it in interviews, after all. Before committing weeks of engineering, they placed a fake door: a realistic entry point in the product, leading to an honest “coming soon” message.
Click-through was far below the bar they'd set. The same users who'd said they wanted the feature mostly didn't reach for it when it appeared in their path. The stated demand from interviews had been real as conversation and hollow as behaviour. A few days of fake-door testing saved weeks of building something that would have shipped to indifference.
The deliverable is the click-through rate against a pre-set bar (plus interest sign-ups) — a behavioural read on demand, not a survey result.
| Asking demand | Fake-door testing demand |
|---|---|
| “Would you use this?” | Counts who actually clicks |
| Inflated, polite yes | Honest behavioural signal |
| Requires building to confirm | Tests before building |
| Stated intent | Revealed intent |
The single deadliest feature risk is that nobody wants it, and the fake door retires that risk for almost nothing — by measuring what people do when the feature appears, not what they say in an interview.
The ethical design is non-negotiable and also practically smart: the door must lead to an honest message, never a deception. Misleading users to harvest a click damages trust and taints the signal; a truthful “coming soon, want to be notified?” both respects the user and yields a second data point (how many register interest). Used this way, the fake door is one of the highest-leverage tools in discovery — it converts the most expensive question (will anyone want this?) into a days-long behavioural test, and a “no” result is just as valuable as a “yes,” because it's a feature you now won't waste a quarter building.
A fake door must lead to an honest message. Tricking users is unethical and poisons the signal.
Without deciding what counts as success first, you'll rationalise any click rate. Set it in advance.
If it's obviously a test, you measure curiosity, not demand. Make it genuine.
Weak demand discovered cheaply is a win — it saved you a wasted build.
Fake doors are a core pretotyping technique (Tool 19) — both test demand before building.
A fake door is often the cheapest method a lean experiment (Tool 15) selects for a demand assumption.
It's a primary tool for retiring the value/demand risk in the Four Big Risks (Tool 01).
Where survey demand (Tool 10) over-predicts, the fake door measures real behaviour.
Pick a feature you suspect users want. Describe where you'd place a realistic entry point for it, and what honest message the click would lead to.
Set the bar in advance: what click-through rate would convince you the demand is real?
If you find yourself wanting to set the bar after seeing the results, you've spotted exactly the rationalisation a pre-set criterion exists to prevent.