10 Fake Door Test Examples for Startup Validation (2026)

A fake door test, also called a painted door test, puts a real-looking entry point for something you have not built in front of real users, then measures how many try to walk through it. The door might be a button, a menu item, a pricing tier, or a landing-page call to action. Behind it is not a finished feature but an honest message: this does not exist yet, here is what we are considering, tell us what you need.

The click is the signal. It tells you whether the people who reached that surface want the thing enough to act, before you spend weeks building. What it does not tell you is whether they will pay, unless you pair it with a pre-sale. On the evidence ladder it sits above a survey, which only captures an opinion, and below a paid pilot, where money actually changes hands.

At Rocket Internet, we built audience before product — email lists, communities, waiting lists, pilot users. A fake-door or waiting-list test is that same instinct made measurable: you put a believable door in front of real people and measure whether they actually try to walk through it, before you spend a month building the room behind it. Keep that instinct in mind through the ten tests below; the door changes shape each time, but the discipline behind it does not.

Below are ten fake door tests across different surfaces and business types: what each measured, how to keep the click honest, and what a pass or a fail looked like. First, the line you cannot cross, and how to instrument the door so the number means something.

The line an honest fake door never crosses

A fake door is a legitimate research method. A bait-and-switch is deception. The entire difference is in what happens after the click.

An honest door lands on a truthful state: we are exploring this, help us prioritize, join the waitlist, tell us what it should do. It never charges a card for a product that does not exist, never confirms an order you cannot fulfill, and never implies the feature is live when it is not. The moment you take money, you have left testing and entered a pre-sale, with real delivery, refund, and consumer-protection obligations attached.

Charging for vaporware, or advertising a product you have no ability or intention to deliver, can be treated as deceptive under consumer-protection rules; in the US, the Federal Trade Commission's advertising guidance covers bait advertising and misrepresentation. This is not legal advice, so confirm your obligations with a qualified professional for your product, market, and channel. But the ethical rule is simpler than the legal one: the person on the other side of the door should never feel tricked when they learn it was a test.

Respect the platform too. Ad networks and app stores ban deceptive destinations and dead-end "coming soon" pages, and a door that annoys users or breaks policy costs you trust and reach worth far more than the data.

What the click actually proves

Read the click as a measure of intent, not value. It shows that people who reached the surface, understood the offer, and felt the problem were willing to spend a few seconds signaling interest. That is real, and most internal feature debates never earn even this much evidence.

It stops short of three things worth naming. It does not prove that the crowd who clicked will pay; only a real commitment shows that. It does not prove they will keep using the thing once it exists; only a working version and repeat use show retention. And it does not prove the market is large; one warm segment can click enthusiastically and still be too small to matter.

So treat the fake door as the lightest rung on the same ladder your other evidence sits on. Use the wider customer-validation framework to weigh a click against an interview, a signed pilot, or a renewal, so an impressive click-through rate does not quietly outrank stronger proof you already hold.

Instrument the door before you open it

A click count with no plan around it is a Rorschach test. Decide these things in writing before the door goes live.

• The action that counts. Name the single event that means intent: the button click, the completed waitlist form, or the submitted answer to "what should this do." A click into an abandoned form is weaker than a completed one, so ask for the deepest action you reasonably can.
• The comparison. A raw number means nothing alone. Hold the door against a baseline you trust: an adjacent real feature's click rate, your normal email click-through, or the same door shown to a different segment.
• The sample. Small samples lie. Five clicks out of forty views can flip to one out of forty the following week. Run the door long enough that the rate would survive being cut in half, and distrust any verdict drawn from a dozen sessions.
• The traffic source. Segment every result by where the visitor came from, because intent is not the same temperature everywhere.

The last thing to fix is the pass line, the result that will change your mind, and you can set that once you have seen the examples below. If you have no audience to send through the door, you do not need to build one first. A small, relevant group sourced the way the no-audience validation process describes is enough to open a door and read a real signal, as long as you hold that audience constant across variations.

1. A new-feature button inside an existing app

The cheapest door to open is one inside a product people already use. A B2B project tool added a Generate weekly status report button to its toolbar for a slice of active accounts, and measured what share of weekly-active users clicked it against the click rate of a comparable, genuinely working action.

The click opened a panel explaining the feature was still being scoped, with one field, "what should a weekly status report include for your team?", and a notify-me toggle. No fake report was generated.

A pass was a click rate at or above the neighboring feature, plus specific, repeated answers pointing at the same job. A fail was clicks that bounced back out of the panel with nothing typed. The trap is novelty: a new button gets clicked because it is new, so watch the second week and compare against a feature users already know, not against zero.

2. A pricing tier for a plan you have not built

Add a tier to a working pricing page and you learn who reaches for the plan above the one they are on. An analytics SaaS placed a third Teams column beside its live Solo and Pro plans. Where the real plans said Subscribe, the Teams card said Coming soon — request early access and opened a short form, never a checkout. It measured how many visitors chose the unbuilt tier over the buyable ones, and which plan those people were viewing from; the form asked for team size and the one capability that would make them switch.

It passed when higher-value visitors, not free-plan tire-kickers, asked for it and named consistent needs, and failed when the interest came only from people who would not pay for the existing plans either. The hard rule sits at the button: never wire the Teams card to a checkout that charges, or one that silently becomes a waitlist. A price beside a working Buy button implies you can sell it today; a clearly labeled request-access state does not.

3. A landing-page call to action that routes to a waitlist

On a pre-launch landing page, the primary call to action is the door. A consumer productivity app ran a hero button reading Start free that led not to a signup but to an honest waitlist: access is opening in waves, tell us your use case. The measure was click-through on the hero button, then completion of the waitlist, split by whether the visitor arrived from an ad, from search, or from a shared link.

Pass and fail diverged on that second step. A strong click-through with high completion and specific use cases from the target audience was real interest. A strong click-through from a link that went briefly viral, then almost no completion, was curiosity; the crowd was wrong for the offer, even though the headline number looked great.

This door tests only the entry click: the headline and the audience, together. To test the whole offer, including price, scope, the follow-up, and whether qualified people finish a real next step, run a full landing-page MVP rather than a single painted button. And never let one traffic spike write your conclusion; hold the audience steady before you trust the rate.

4. A paid ad to an honest 'coming soon' page

When you have no audience, you can rent a little attention. A hardware accessory that did not yet exist ran a small ad budget to a page describing the product with one action: Notify me when it launches. No Buy now, no charge. The read was ad click-through, the share of clickers who left an email, and the cost per interested email, each split by creative and audience so one lucky ad could not carry the verdict.

It passed when the cost per genuinely interested email was low enough that the eventual unit economics could work, and that held as the audience widened. It failed when a single creative produced all the signal and everything else fell flat. Two cautions: ad platforms restrict deceptive and dead-end destinations, so read the policy before you spend, and a paid click is the coldest kind of intent, thick with targeting noise, so set the bar higher than for a click from someone already inside your product.

5. A nav menu item for a future module

A vertical SaaS for salons wanted to know whether customers would adopt in-app Payments before committing a quarter to building it, so it added Payments to the main navigation for logged-in users. The signal was the share of sessions that opened the item and, more usefully, which customers those were. The link led to a short honest page: payments is on the way, here is what we are weighing, how do you take payment today? Retained, paying salons clicking through and describing their card-machine-and-spreadsheet workaround was the pass; clicks from trial accounts already on their way out meant little.

The specific risk of a nav-level door is that it competes with real features and can send a paying user down a dead end mid-task. Keep the test short, watch support tickets for confusion, and take the item down the moment you have your answer. A door you leave hanging for a month stops being a test and becomes a broken promise.

6. A 'request an integration' page

Not every fake door asks build-or-not; some ask build-what-first. A B2B tool with an integrations directory listed connectors it had not built, each marked as planned and never as available, beside a Request this integration button and a catch-all field. It measured not a yes or no but a distribution: which integrations real accounts asked for, and how sharply demand concentrated. A pass was concentration, one or two integrations pulling most of the requests from target-fit accounts, giving an obvious place to start; a fail was a long scatter of one-off asks, or a list dominated by people who were not customers.

The pitfall lives in the labeling. A partner's logo under anything that reads as "available" can imply a relationship you do not have and mislead both users and that company. Mark planned integrations unmistakably, and let the votes, not your roadmap hopes, choose the order.

7. A 'book a demo' for a product still in concept

A Book a demo button is a door with a person behind it. An early B2B team ran one while the product was barely a prototype and counted only qualified bookings, companies that fit the profile, rather than raw calendar fills. Here the honest handling extends past the click into the call: you show a concept or clickable prototype and state plainly what is real, what is mocked, and what does not exist. A qualified buyer who books, then describes the problem in detail and asks about timeline and price, is a strong signal; bookings you cannot staff, or attendees expecting finished software who leave disappointed, are the failure.

The distinctive hazard is that this door creates an immediate human obligation. Unlike a silent click, someone has to show up and be honest. Do not book more demos than you can hold, and never run a hidden-manual workflow as though it were the automated product without saying so.

8. An email announcement to one segment

If you already have a list, an email is a door you can open to one chosen room. A tool emailed a single segment, power users of a related feature, with "New: [capability]. Try it," and measured click-through against its own normal campaign rate. The link led to an honest "we are building this, shape it" page, and the email never claimed the feature was live. Click-through well above the usual campaign rate, plus replies describing the need, was the pass; opens without clicks, or clicks shadowed by unsubscribes, was the fail, and the unsubscribes matter because they are the price of the test.

The real caution with an email door is that you are spending list trust and deliverability. Send to a small, relevant segment, not the whole list, and keep every claim true, because a single "announcement" of something that does not exist teaches your best subscribers to discount the next one.

9. An app-store-style pre-launch page

Mobile stores offer an honest version of this door built in: a pre-launch listing that tells users, in the platform's own words, that the app is not out yet and lets them register interest. A consumer app used one and tracked pre-registrations by source. Because the action is so cheap, one tap, no account, no payment, the number inflates easily, so the pass bar sits not at the pre-registration count but at conversion into real installs on launch day. Thousands of pre-registrations that produce a trickle of installs is a classic cheap-action mirage; a smaller list that shows up and installs is worth far more.

Use the platforms' official pre-order and pre-registration features rather than mocking up a fake store listing, which misleads users and breaks policy. And weight the whole signal modestly: a pre-registration is interest at the lowest possible cost, which is exactly why it is so easy to overread.

10. An in-product upsell for an expansion tier

The warmest door of all opens in front of a paying customer at the moment they hit a limit. When accounts maxed out their dashboards, one product showed Unlock advanced analytics with a see-plans action, though the tier did not exist. It measured how many customers clicked at the point of need, and whether the same high-value accounts kept doing it. The click led to a roadmap state, "this is coming, tell us what advanced analytics must include for you," not a charge and not a fake upgrade screen. A pass was a real share of the best, most-engaged customers clicking and describing consistent needs, which reads as genuine expansion demand; clicks only from low-usage or churning accounts read as noise.

The line not to cross is manufacturing the signal by hiding an existing capability behind a fake "upgrade." Punishing current customers to generate a number is both dishonest and self-defeating; the door has to lead to something genuinely new.

Read the result, then build, pre-sell, or drop

One door, one week, one audience is a single data point, so treat it like one. You wrote the action that counts, the comparison, and the pass line before opening it; now hold yourself to them.

• Build when strong intent shows up from the right segment, across more than one traffic source, with qualitative depth behind it, meaning people did not just click but described the same real need. Even then, build the smallest version that delivers the outcome, not the whole imagined product.
• Pre-sell first when intent is strong but you have never tested money. A click is not a card. Turn the interest into a commitment, a deposit, a paid pilot, or a founding-customer offer, by learning to pre-sell the idea before building, which is the test that actually measures willingness to pay.
• Step up to a fuller MVP when the door proved intent to enter but you still need to prove the thing behind it delivers value. A fake door never tests the outcome, only the wish to reach it, so pick a heavier format that puts a real result in front of the customer from the MVP examples library.
• Drop or revise when the signal is weak, comes from a single source, or comes only from the wrong segment. A door that barely beat a dead link, or that only cold ad traffic touched, is telling you to change the offer, the audience, or the problem, not to start building.

Whatever the number, do not let one test carry the decision. Re-run the door, change one variable at a time, watch the second-week decay once novelty wears off, and read it against a real baseline. In a launch system like the 100 Tasks framework, a fake door is a LAUNCH-stage validation move: cheap evidence bought before committed build, logged with a written hypothesis and a pre-committed decision rule so the result cannot be rationalized after the fact.

Your next step is small. Pick one surface you already have honest traffic to, a button in your app, a tier on your pricing page, or one email segment, then write the single action that counts and the number that would change your mind. Build the honest post-click state first, and only then open the door, to a defined audience, for a set window. Let the clicks argue with your assumptions instead of decorating them.

Inside the 100 Tasks framework, this is Task 25 in the LAUNCH stage, the landing-page demand test, and the SEO Brief Builder, Launch Amplifier, and Ad Copy Machine Marketing Skills exist specifically to drive real, qualified traffic to that test instead of a guess. Set one up inside 100 Tasks AI.