

A Great Place to Upskill
Learn the difference between MVP experiments and fake door tests, when to use each, plus examples, metrics, ethics, and PM experimentation best practices.

.png)
Imagine you want to open a new restaurant, but you are not sure whether people in your neighbourhood actually want Thai food. You have two ways to find out cheaply. The first is to cook one small Thai dish, sell it at a local market stall for a weekend, and watch whether people come back for more. You spent a little effort, but real customers ate real food and told you something true. The second way is even cheaper: you put up a poster that says "Thai restaurant opening here soon, sign up for the launch," and you count how many people scan the QR code. You cooked nothing at all, yet you still learned whether the interest is there.
Both approaches are experiments, and both are far smarter than signing a lease and building a full kitchen on a hunch. But they are not the same thing, and confusing them is one of the most common mistakes product teams make. The first is a minimum viable product experiment. The second is a fake door test. This guide explains exactly what each one is, how they relate, when to use which, and the ethical line a fake door test must never cross.
An MVP experiment builds the smallest real, usable version of a product or feature to learn whether it delivers value once people actually use it. A fake door test builds no working functionality at all; it places a realistic-looking entry point for a feature that does not exist yet and measures how many people try to use it, purely to gauge demand before anything is built. The MVP tests "will this work and will people use it?" The fake door tests the narrower, earlier question "do people even want this?"
The key relationship to understand is that these are not opposites. A fake door test is actually one specific, extremely lightweight type of MVP experiment, sitting at the demand-validation end of the spectrum. Both belong to the same family of lean product experiments popularized by the Lean Startup methodology, and both exist to replace expensive guesses with cheap evidence. They simply answer different questions at different stages, and the deeper foundations of the MVP concept are covered in ProductSpace's complete MVP guide.
A minimum viable product experiment is a test in which you build the smallest genuinely functional version of a product or feature, put it in front of real users, and measure whether it delivers value and gets used before committing to a full build. The emphasis is on "functional." An MVP experiment produces something a person can actually use to accomplish a real task, even if it is rough, manual, or missing polish.
MVP experiments come in several flavours, and the differences matter:
Concierge MVP: you deliver the service manually, by hand, for a small set of early customers, learning exactly what they need before automating anything. The user knows a human is helping them.
Wizard of Oz MVP: the product looks fully automated on the surface, but behind the scenes humans are manually doing the work the software will eventually do. The user does not know it is manual.
Single-feature MVP: you build one core feature, done well, and nothing else, to test whether that single slice of value is enough to attract and retain users.
Piecemeal MVP: you stitch together existing tools and services to simulate the product without building it from scratch.
The common thread across all of these is that the user receives real value from a working (if minimal or manually-powered) product. The experiment answers questions like: does this actually solve the problem? Will people use it repeatedly? Does the value hold up in real use, not just in a survey? These are the deeper questions covered in ProductSpace's guide to product analytics fundamentals, since an MVP experiment is only as good as your ability to measure what users actually do with it.
Real-world example: The founder of an online shoe store, before building any inventory system, manually photographed shoes at local stores and fulfilled each order by hand when someone bought. Customers received real shoes; the founder learned people would genuinely buy footwear online, all without building warehouse infrastructure. That is a classic Wizard of Oz style MVP experiment.
A fake door test, also called a painted door test or a 404 test, places a realistic-looking entry point, such as a button, menu item, banner, or pricing tier, for a feature that does not actually exist, then measures how many users click or engage with it to gauge genuine demand. When a user clicks, they reach an honest "coming soon" message, often with an option to join a waitlist or leave feedback.
The name comes from architecture, where a door or window is painted onto a wall to create a sense of balance and symmetry, looking real but opening onto nothing. A painted door test works the same way: the feature looks real, but there is nothing behind it yet.
The crucial distinction is that a fake door test builds no working functionality whatsoever. The entire experiment is the front of the feature and the click it generates. That click is the signal. The logic is simple: where there is smoke (clicks), there is probably fire (demand). Because clicking takes a small but real effort, it reveals genuine intent in the moment far more reliably than a survey, where people say one thing and do another.
Fake door tests typically live in one of a few forms:
A button or menu item inside an existing product, like an "Export to PDF" option that does not exist yet
A banner or promotion teasing an upcoming feature with a call to action
A pricing tier or add-on that is not yet built, to test willingness to pay
A landing page for a whole product that does not exist, to validate a brand-new idea
Real-world example: A social media scheduling tool famously created a landing page for a product that did not yet exist, promoted it, and collected thousands of email sign-ups before writing a single line of code. Those clicks and sign-ups validated real demand and became the foundation for the actual product that followed.
Dimension | MVP Experiment | Fake Door Test |
|---|---|---|
What gets built | The smallest real, usable version of the product | Nothing functional, just the entry point |
Question it answers | Will this work, deliver value, and get used? | Do people even want this? |
User outcome | The user receives real value | The user hits a "coming soon" message |
Signal measured | Usage, retention, satisfaction, real behaviour over time | Clicks, sign-ups, engagement in the moment |
Stage | After demand is suspected, to validate value and usability | Very early, to validate demand before building |
Effort and cost | Low, but still requires building something real | Minimal, often just design or a banner |
Main risk | Building something, even minimal, that flops | Damaging user trust if handled dishonestly |
The simplest way to hold them apart: a fake door test asks "should we build this at all?" while an MVP experiment asks "now that we are building it, does the smallest real version actually work?" A fake door usually comes first. If the demand signal is strong, an MVP experiment is often the logical next step to validate that the built thing genuinely delivers.
You have a crowded backlog of unvalidated feature ideas and need to prioritize based on real interest, not opinion
The feature would take significant engineering effort, and you want directional demand data before committing
You already have an existing product with real users who can encounter the fake door naturally
You want to test willingness to pay for a new tier or add-on before building it
You need a fast, cheap answer to "is there interest?" in days, not weeks
You already have evidence of demand and now need to know whether the actual product delivers value
The real risk is not whether people want it, but whether you can build something they will keep using
You need to learn about usability, retention, and real behaviour, not just initial interest
A manual or scrappy version can genuinely serve early users while you learn
You are ready to move from "do they want it?" to "does it actually work for them?"
A common and powerful sequence is to run a fake door test first to confirm demand, then build an MVP experiment for the ideas that earned a strong click signal. The fake door narrows the field cheaply; the MVP validates the survivors properly.
Fake door tests carry a real risk that MVP experiments do not: they can damage user trust if handled carelessly. Because you are, by design, showing people something that is not there, honesty in the reveal is not optional.
A few guardrails separate a responsible fake door test from one that feels like a scam:
Be honest at the reveal. The "coming soon" message should thank the user, explain you are gauging interest to build the right things, and never pretend the feature was real and simply broke.
Avoid over-promising language. Framing like "Are you interested in..." or "Would you like..." is safer than a confident label that implies the feature already works.
Keep tests small and targeted. Showing a fake door to your entire user base repeatedly erodes trust. Target a relevant segment and limit exposure.
Offer a real next step. A waitlist sign-up, a short survey, or a path back into the product turns a moment of mild disappointment into genuine involvement.
Never take money for something that does not exist. Testing willingness to pay is fine up to the click; actually charging for a phantom feature crosses a clear line.
Key Insight: The disappointment of hitting a "coming soon" page is small and forgivable when you are transparent. The damage comes from deception, making users feel tricked rather than consulted. Handled honestly, many users happily join a waitlist, because being asked what to build feels like being valued, not fooled.
Treating them as opposites: a fake door test is a type of MVP experiment, not its rival. The real skill is knowing which question you are answering.
Skipping the fake door and over-building: running a full MVP experiment for a feature nobody clicked on wastes the exact effort the fake door was meant to save.
Reading clicks as proof of value: a strong fake door click signal proves interest, not that the built feature will satisfy or retain users. That is what the MVP experiment is for.
Running a fake door with no clear hypothesis: without a defined success threshold set in advance, you will not know whether the click rate you got is good, bad, or noise.
Ignoring who clicked: a raw click count hides intent. Users comparing options or clicking out of curiosity are not the same as users with genuine need, so pair clicks with follow-up signals.
Damaging trust with a dishonest reveal: the fastest way to poison a fake door test is to make users feel deceived rather than consulted.
Forgetting the "so what": an experiment that does not change a build decision was theatre, not learning. Define upfront what a positive, negative, or inconclusive result will make you do.
An MVP experiment builds the smallest real, usable version of a product to test whether it delivers value and gets used
A fake door test builds nothing functional; it measures demand for a feature by counting clicks on a realistic-but-empty entry point
A fake door test is actually one lightweight type of MVP experiment, sitting at the demand-validation end of the spectrum
The fake door answers "do people want this?" while the MVP experiment answers "does the real thing actually work for them?"
A strong pattern is to run a fake door test first, then build an MVP experiment for the ideas that earn a clear demand signal
Fake door tests demand honesty at the reveal, since their main risk is damaging user trust, not wasted build effort
Just like testing a Thai restaurant with either a weekend market stall or a simple "opening soon" poster, MVP experiments and fake door tests are both far smarter than betting everything on a hunch. The poster (the fake door) tells you cheaply whether anyone cares. The market stall (the MVP experiment) tells you whether what you actually make holds up once real people use it. Neither replaces the other, and the best product teams use them in sequence: validate the demand first, then validate the value. Do that consistently, and you stop shipping features nobody wanted and start building the ones people were already reaching for.
If you want structured, hands-on practice designing and running real product experiments like these with mentor feedback, ProductSpace's AI Product Management Fellowship covers validation, experimentation, and discovery across a 14-week, project-based curriculum. And for the full foundations of the MVP concept, ProductSpace's complete MVP guide is the natural next read.
How well do you know What Is an MVP Experiment vs. a Fake Door Test?? Let's find out with this quick quiz! (just 10 questions)
Question 1 of 10
An MVP experiment is a minimal but functional version of a product or feature built to test whether customers get enough value to justify further investment. It delivers real value, even if rough or partly manual, so the team can observe actual behavior. Common forms include concierge MVPs, Wizard of Oz MVPs, and single-feature MVPs, all designed to maximize learning while minimizing wasted effort.
A fake door test measures customer interest by presenting an option or feature that looks available but has not been built yet. When a user clicks, they usually see a "coming soon" message or a waitlist. It captures real intent through behavior rather than opinion, making it one of the cheapest and fastest ways to validate demand before any engineering work begins.
An MVP experiment delivers real value and tests whether customers actually benefit from a working solution. A fake door test delivers no working feature and tests whether demand exists in the first place. In short, MVPs validate value while fake door tests validate intent. Many teams run a fake door test first to confirm demand, then build an MVP to confirm value.
PMs should use fake door tests when they need to validate demand quickly and cheaply, before committing engineering resources. They are ideal for comparing several feature ideas, sizing an opportunity, or settling internal debates with real intent data. If the main question is "does anyone even want this?" a fake door test is usually the fastest and lowest-cost way to find out.
Fake door tests are ethical when run responsibly. The key is honesty after the click, such as a clear "coming soon" message, and never taking payment or making firm promises you cannot keep. Handled carelessly, they can erode trust and leave users feeling deceived. Handled with transparency and respect, they are simply a behavioral way of asking users what is worth building.
For fake door tests, the key metrics are click-through rate, signup rate, and waitlist registrations, which signal intent. For MVP experiments, activation rate, retention, and willingness to pay reveal whether real value is delivered. Conversion rate and customer interviews support both. The right metric depends on your question: use click metrics for demand and value metrics like retention for delivery.
Yes. Many validation methods require little or no code. Fake door tests need only a button or landing page. Concierge and manual service MVPs deliver value by hand. Video MVPs and landing page MVPs test demand with no product at all. These approaches let founders confirm whether an idea has real demand before investing in engineering, saving significant time and money.
Alternatives include fake door tests, landing page experiments, concierge MVPs, Wizard of Oz MVPs, manual service MVPs, and video MVPs. Each tests a different risk at a different cost. Customer interviews and smoke tests also help validate assumptions early. The goal across all of them is the same: learn whether an idea is worth building before committing heavy engineering resources.
Learn what change management is for product rollouts, why most failures are people problems not tech problems, the ADKAR model, and mistakes.
Learn what a fake door test is, the real ethical tension behind it, why a click only measures interest, when to use one, and when to avoid it.
Learn what a product health dashboard is, why most fail without anyone noticing, the HEART framework, a gut-check test, and mistakes to avoid.