What Is a Fake Door Test? Definition + Example
A fake door test advertises a feature or product that does not exist yet — as a button, menu item, pricing tier, or ad — then measures how many people try to walk through it. Those clicks and sign-ups are a demand signal you gather before building anything, greeting each visitor with an honest "coming soon."
Quick Answer: A fake door test (a.k.a. a painted door) places a fake entry point — a button, link, pricing tier, or ad — in front of real users for something you have not built, then counts who clicks or signs up as evidence of demand before showing a "coming soon" page. Popularized by Testing Business Ideas and Alberto Savoia's The Right It, it buys real market data cheaply — provided you stay honest and never charge for vapor.
The premise is simple: an action is worth more than an opinion. Ask people whether they would use a feature and you invite polite, unreliable answers. A fake door makes them do something instead — click, sign up, reach for their wallet — and that behavior is the data. It is one of Alberto Savoia's pretotyping techniques in The Right It, his method for collecting your own market data before you build, and David Bland and Alexander Osterwalder catalog the same move in Testing Business Ideas as a fast, low-cost demand experiment.
How a Fake Door Test Works, Step by Step
A fake door test works by staging a convincing entrance to a product that has no rooms behind it yet, then watching who tries the handle. The mechanics are deliberately cheap — you are buying a demand signal, not shipping a feature.
- Pick one thing to test. Name the single feature, offer, or price you are unsure about. A door can only answer one question at a time.
- Build the door, not the room. Create the visible entry point — a button, a menu item, a pricing tier, an ad — and nothing behind it.
- Show it to real, relevant traffic. Expose it to a slice of existing users or a small paid audience who genuinely fit the use case.
- Measure the attempt to enter. Count clicks, sign-ups, or waitlist emails against how many people saw the door.
- Catch them honestly. The instant someone walks through, show a truthful "coming soon" and offer to keep them posted.
What you are after is a rate — the share of people who saw the door and tried to open it — not a raw tally. A rate lets you compare this idea against a control or against your other bets. Decide what result would make you build before you look, so you judge the number honestly instead of rationalizing whatever appears.
A Concrete Fake Door Example
The most common form is a new menu item placed inside a product that already has users. Imagine a small project-management app weighing whether to build an automated "AI Weekly Report."
Rather than spend a sprint on it, the team adds a "Generate Weekly Report" button to the dashboard. It looks and behaves like a real feature. When a user clicks, a friendly panel appears: "This is in the works — want early access?" with a single email field.
The click is the question; the email is the answer. Over two weeks the team watches what fraction of active users find and click the button, and how many leave an email at the coming-soon panel. A trickle of clicks and near-zero emails is a quiet no. A steady stream of both is a green light — and a warm waitlist to launch into.
Notice what the test captured: not what users said in a survey, but what they did when they believed the feature was one click away. That behavioral evidence is exactly the "your own data" Savoia urges you to trust over borrowed opinions.
Fake Door vs. Painted Door vs. Smoke Test
These three terms overlap so heavily that people use them interchangeably, and that is mostly fine. The only useful distinction is where the door sits and how much you are testing.
The table below maps the family. It is qualitative — a guide to when each label fits, not a ranking.
| Technique | What it is | Where the "door" lives | Best for |
|---|---|---|---|
| Fake door test | Umbrella term for any fake entry point to something unbuilt | Anywhere — button, link, tier, or ad | Appetite for a specific feature or offer |
| Painted door test | Near-synonym; a fake control painted onto a live product | Inside an existing product's UI | An incremental feature for current users |
| Smoke test | Broader pre-launch demand test for a whole product | A standalone ad and landing page | Validating a new product from scratch |
Takeaway: Reach for a painted door to test a feature inside something people already use, and a smoke test to validate a whole new product. "Fake door" is the family name covering both — so if the labels blur in conversation, you have lost nothing, because the method is identical.
For the in-product variant, see our definition of the painted door test; for the standalone ad-and-page version, see what a smoke test is in startup validation.
Running a Fake Door Test Ethically
A fake door briefly misleads real people, so treat it as a small, honest withdrawal from your trust account — never a fraud. The ethics come down to being upfront the moment interest is captured, and never taking money for something that does not exist.
- Land them softly. When someone clicks, tell the truth — the feature is being explored, not broken. A fake error or an endless spinner is a lie; an honest "coming soon" respects them.
- Turn intent into a relationship. Offer to notify them at launch. This captures the demand you came to measure and hands the visitor something real — a place in line — for their small letdown.
- Never charge for vapor. Measure intent with a click or an email, not a credit-card charge for a product that will not arrive. Taking payment for something that does not exist is fraud, not a test.
- Keep the blast radius small. Show the door to a slice of traffic, briefly. You want a signal, not a reputation hit across your whole audience.
- Honor the implied promise. If people knock, build it — or tell the waitlist honestly you decided not to. Ghosting an interested list teaches users to ignore you.
Both The Right It and Testing Business Ideas treat this honesty as non-negotiable, because a demand experiment that poisons customer trust has failed even when the numbers look good. Run cleanly, a fake door is one of the cheapest, sharpest probes you have — best used as one step inside a larger startup idea validation process. A validation workspace like Edmired can help you record the hypothesis and the pass/fail bar before the clicks arrive, so you read the result honestly.
Key Takeaways
- A fake door test measures demand for something unbuilt. You stage a fake entry point — a button, tier, or ad — and count who tries to walk through before writing real code.
- Actions beat opinions. A click or an email is behavioral evidence; a survey answer is a polite guess. Watch what people do, not what they say.
- It is grounded in pretotyping and lean experiments. It is one of Savoia's pretotyping techniques in The Right It and a cataloged demand experiment in Testing Business Ideas.
- Painted door and smoke test are close cousins. A painted door is the in-product variant; a smoke test validates a whole product with an ad and a page. "Fake door" is the family name.
- Judge a rate, and set the bar first. Track the share of viewers who try the door, and decide what result would make you build before you see the data.
- Honesty is the price of admission. Show a truthful "coming soon," capture interest with a waitlist, keep it small, and never charge for a product that does not exist.
Frequently Asked Questions
Is a fake door test ethical?
Yes, when run honestly. The line is transparency: the moment someone acts on the door, tell them the truth and offer to keep them informed. It becomes deception only if you fake errors to hide the test or take payment for a product that does not exist. Keep the test small and brief, and it stays a fair, low-cost probe.
What is the difference between a fake door and a painted door test?
Very little — the terms are often synonyms. When people do distinguish them, "painted door" usually means a fake control placed inside a live product, such as a button that leads to a "coming soon" panel, while "fake door" is the umbrella term that also covers standalone ads and landing pages. Both measure clicks as a signal of real demand.
How many clicks make a fake door test a success?
There is no universal number; success is relative to a baseline you set beforehand. Compare the share of viewers who try the door against a control, against your other bets, or against the threshold you committed to before running it. A rate you can compare against something meaningful beats any impressive-sounding raw count.
Can you run a fake door test without existing users?
Yes. Instead of placing the door inside a product, drive a small amount of paid or social traffic to an ad or landing page for the unbuilt feature and measure sign-ups. At that point the fake door effectively becomes a smoke test — the same idea, staged for an audience you do not have yet.