Test whether anyone will use a product at all — at the lowest possible cost — before building it. “Pretend” + “prototype”: make it look real enough to get a true behavioural signal.
▸ Try the interactive toolPretotyping is a set of techniques for testing whether anyone will actually use a product at all, at the lowest possible cost, before committing resources to build it. The word deliberately fuses pretend and prototype: a pretotype pretends the product exists in order to gather a behavioural signal about demand.
Where a prototype tests can we build it and can people use it, a pretotype tests the more fundamental question: will anyone want it in the first place? The techniques — fake doors, a manual “concierge” version where humans fake the product's function, a landing page, a “mechanical Turk” — all share one goal: get a real behavioural signal of demand while building essentially nothing. The trap pretotyping guards against is falling in love with an idea and building it before confirming anyone cares.
Fake door — measure clicks on an unbuilt feature · Concierge — humans manually deliver the product's value · Landing page — measure sign-ups for a non-existent product · Mechanical Turk — fake automation with manual work. All test demand, not function.
The most expensive mistake in product is building something nobody wants. Pretotyping exists to make that mistake nearly impossible by testing demand before the build.
| The shortcut | What it costs | What it gives you instead |
|---|---|---|
| Building before demand is proven | The classic, costly failure: ship it, then learn nobody cared. | Pretotypes test demand at near-zero cost, pre-build. |
| Falling for the idea | Conviction substitutes for evidence of demand. | A behavioural signal replaces founder enthusiasm with data. |
| Confusing interest with demand | “That's cool” isn't “I'll use it.” | Pretotypes measure action, not enthusiasm. |
| Over-investing to test | Building a real version just to gauge interest. | Concierge and fake-door methods fake the product entirely. |
State the belief: will the target users actually use this? Set, in advance, the behavioural signal that would count as real demand — clicks, sign-ups, repeat manual requests.
Match technique to the idea: a fake door for an in-product feature, a landing page for a new product, a concierge test where you manually deliver the value for a high-touch service. Build as little as possible.
Make it look and feel real enough to elicit genuine intent, while doing the actual work manually or not at all. The realism matters — a signal from an obvious test is worthless.
Compare actual behaviour to your pre-set bar. Did people click, sign up, come back, pay? Action is the signal; expressed enthusiasm is not.
Strong demand justifies investing in the build; weak demand saves the investment. The whole point is to make this call before the expensive part.
A team had a conviction-level belief in a new service idea and was ready to build the automated product. Instead they ran a concierge pretotype: they offered the service to a handful of real users but delivered it entirely by hand behind the scenes, faking the automation they hadn't built.
Two things emerged cheaply. First, demand was real but narrower than assumed — only a specific subset came back for more. Second, doing it manually revealed which parts of the imagined product actually mattered to users and which were irrelevant. They built the real thing later, for the right segment, having spent almost nothing to learn what conviction alone would have got wrong.
The deliverable is a behavioural demand signal against a pre-set bar — from whichever pretotype technique fit — plus what the fakery taught about the real product.
| Technique | Fakes… | Best for |
|---|---|---|
| Fake door | An unbuilt feature's entry point | In-product feature demand |
| Concierge | Automation (humans do it) | High-touch services |
| Landing page | The whole product | New-product demand |
| Mechanical Turk | The “engine” (manual behind UI) | Validating a workflow |
Pretotyping answers the question that comes before every other discovery question: will anyone want this at all? And it answers it the only honest way — by measuring behaviour while building essentially nothing.
The distinction from prototyping is worth holding onto: a prototype assumes demand and tests usability or feasibility; a pretotype tests the demand itself, the assumption everything else rests on. Skipping it is how teams end up with beautifully usable, technically sound products that nobody uses. The concierge technique carries a bonus the others don't — by manually delivering the value, you discover not just whether there's demand but what users actually value, which often differs sharply from the product you imagined. Test demand first, fake everything you can, and let real behaviour — not conviction — earn the build.
If you're building the real thing to test demand, you've missed the point. Fake it first.
“Cool!” isn't usage. Measure action — clicks, sign-ups, repeat requests.
If it doesn't feel real, the signal is worthless. Make it convincing.
Decide what counts as real demand before you run it, or you'll rationalise the result.
The fake door (Tool 18) is one core pretotyping technique among several.
Pretotypes are often the cheapest method a lean experiment (Tool 15) chooses for a demand assumption.
Pretotyping is the front-line tool for the value/demand risk in the Four Big Risks (Tool 01).
Pretotype to confirm demand; then prototype (Tool 16) to test usability and feasibility.
Take a product or feature idea you believe in. Pick the cheapest pretotype technique that would test whether anyone actually wants it — fake door, landing page, or a manual concierge version.
Describe how you'd fake the product entirely, and what behavioural signal (not opinion) would convince you demand is real.
If the honest test would cost days while building the real thing would cost weeks, you've found why conviction is the most expensive reason to build — and behaviour the cheapest reason to.