Pretotyping: How to Test the It Before You Build It

Pretotyping is testing whether people actually want your product before you build it, by simulating its core experience with the smallest possible investment of time and money. Coined by Alberto Savoia in The Right It, it answers one question that decides most startups: are you building The Right It, or the Wrong It? Validate demand first, engineer later.

Quick Answer: Pretotyping tests demand for an idea in hours or days, not months, by faking the product's core experience just enough to see if real people will commit real skin in the game: money, time, effort, or data. The goal is to make sure you are building The Right It before you build It right.

Most founders skip this step. They fall in love with an idea, spend a year building it beautifully, launch, and hear silence. The problem was almost never the code or the design. The problem was that nobody wanted the thing in the first place, and no amount of engineering polish fixes a product the market doesn't want. Pretotyping is the discipline that catches that failure in an afternoon instead of after a funding round.

What is pretotyping, and how is it different from prototyping?

Pretotyping tests whether you should build something; prototyping tests whether you can build it and how well. The words look almost identical, which is exactly why founders conflate them and waste months answering the wrong question first.

A pretotype is a "pretend prototype" — a mash-up of pretend and prototype. Where a prototype is a real, if rough, working version of your product, a pretotype fakes the experience convincingly enough to observe how real people respond, without you building the underlying product at all. Savoia's own definition is precise: pretotyping tests the initial appeal and actual usage of a potential new product by simulating its core experience with the smallest possible investment of time and money.

The distinction is the whole point of the book. Its subtitle — make sure you are building The Right It before you build It right — separates two questions founders collapse into one:

• The Right It is a demand question: does the market actually want this?
• It right is an execution question: can you build it well, at quality, at scale?

Prototyping, design sprints, and clean architecture all serve "It right." They assume the idea deserves to exist and ask how well you can pull it off. Pretotyping refuses that assumption and interrogates it first.

The two words answer opposite questions, so it helps to see them side by side before you commit a single sprint to either.

A prototype assumes the idea is worth building and asks how well you can build it. A pretotype withholds that assumption and asks whether you should build it at all. Pretotyping comes first for a hard economic reason: it is far cheaper to discover the Wrong It in a day than after a year of flawless engineering.

The Law of Market Failure: why competent execution isn't enough

Most new products fail in the market even when they are competently built and skillfully marketed. Savoia calls this the Law of Market Failure, and it is the uncomfortable premise the entire method is built to survive.

Failure is the default, not the exception. This reframes what founders think they're up against. If you assume most ideas succeed and yours is the rare dud, you'll treat validation as a formality. If you assume most ideas fail — including yours until proven otherwise — you'll treat validation as the main event. The Law of Market Failure argues for the second posture.

Execution is not the usual cause of death. The seductive story is that failed startups were sloppy: bad code, weak marketing, a mediocre team. Savoia's argument is that plenty of failed products were built well and marketed hard. They died because they were the Wrong It — a competent answer to a question the market wasn't asking. You cannot out-engineer, out-design, or out-market a fundamental lack of demand.

This is why the sequence matters so much. Pouring your best work into "It right" before you've established "The Right It" doesn't reduce your risk; it just makes the eventual failure more expensive and more painful. For the wider decision framework this sits inside, see the complete guide to startup idea validation, which places pretotyping alongside the other checks a founder runs before committing.

Escape Thoughtland: turn beliefs into an XYZ hypothesis

Get out of your own head and convert vague optimism into a specific, falsifiable claim you can actually test. Savoia calls the place founders get stuck "Thoughtland" — the comfortable zone where ideas live as beliefs, opinions, and predictions that nobody has checked against reality.

Thoughtland feels like progress but produces none. It's where you brainstorm, poll friends, run a survey, and collect enthusiasm. The trouble is that everybody's baby is beautiful in Thoughtland. Ask people whether they'd use your idea and they'll be polite, imaginative, and completely unreliable, because a hypothetical "yes" costs them nothing. Opinions are cheap, abundant, and a poor predictor of what anyone will actually do.

The exit from Thoughtland is a hypothesis, not more discussion. You start with a belief — some assumption about how the market will engage with your idea. That belief has to be rewritten as a claim precise enough that data can prove it wrong. Savoia's format for this is the XYZ Hypothesis:

At least X% of Y will Z.

• X is a percentage — a testable threshold you commit to in advance.
• Y is your target market — a specific, identifiable group.
• Z is the concrete action that counts as engagement — buy, pre-order, sign up, click, book a call.

A vague belief like "people will love a meal-kit for one-person households" becomes "at least 10% of single-occupant apartment residents who see our offer will place a paid pre-order." Now it's falsifiable. The number forces honesty; you can't quietly move the goalposts after the fact.

Then you scale it down with hypozooming. A market-wide XYZ claim is still too big to test today, so you zoom in until you have a small, local, immediate version — an xyz hypothesis — that a representative slice of your market can prove or disprove this week. You don't test "20% of all commuters nationwide." You test the commuters walking past one station exit on one Tuesday. If the local slice is representative, the local result carries a real signal. For the mechanics of choosing X and writing a number you can defend, see the deep dive on writing a testable XYZ hypothesis.

Collect YODA: your own data with skin in the game

The only data worth betting on is data you collected yourself, from real people, making real decisions with something at stake. Savoia's term for this is YODA — Your Own DAta — and he contrasts it sharply with other people's data.

Other people's data can't answer a genuinely new question. Market reports, analyst forecasts, and competitor benchmarks describe a world that already exists. Your novel idea, by definition, doesn't exist yet, so nobody has measured demand for it. Borrowed data feels rigorous and reads well in a deck, but it's answering a question adjacent to yours, not yours. YODA is first-hand, fresh, local, and specific to your exact idea — which is the only kind that predicts your outcome.

Skin in the game is what separates signal from noise. The strength of a data point comes from how much the person had to risk to produce it. A "like," a nod, a survey checkbox — these cost nothing, so they mean almost nothing. Real evidence shows up when someone parts with something they value: their money, their time, their contact details, their reputation, their effort. That's skin in the game, and it's the dividing line between an opinion and a fact.

Not all commitment is equal, so it helps to rank the signals you can collect by how much the person actually risks to give them.

Verbal enthusiasm sits at the bottom because it costs the person nothing. As you move down the ladder, each signal asks the prospect to give up something they'd rather keep, which is precisely why it predicts real future behavior better. When you design a test, aim as low on this table as you can — a paid pre-order tells you far more than a thousand waitlist emails. The trade-offs between these tiers are worth studying in the guide to the strength of different validation evidence.

The pretotyping techniques: which pretotype fits which question

Pick the pretotype that fakes the riskiest part of your idea for the least money. Savoia catalogs a handful of named techniques, and the skill is matching the technique to the specific belief you most need to test.

Each of the named pretotypes fakes a different layer of the product, which is what lets you test demand without building the real thing.

The catalog isn't a checklist to run in full; it's a menu to pick from. Start with whichever pretotype puts a real commitment decision in front of a real stranger fastest and cheapest.

The Mechanical Turk: fake the automation with hidden humans

The Mechanical Turk pretotype makes people believe a machine is doing the work while humans quietly do it behind a curtain. Savoia's classic illustration is a speech-to-text experiment run long before the technology existed: a person spoke into a microphone, words appeared on screen as if transcribed by a computer, and a hidden typist in another room was actually producing the text. It let the team study real usage — and surface real drawbacks, like a sore throat from talking all day — before committing to build anything.

Use it when the hard, expensive part is the technology, but the demand question comes first. You don't need working AI to learn whether people want the AI's output. Deliver the output by hand, watch how people actually use it, and only build the automation once demand is proven.

The Pinocchio: carry a lifeless model as if it were real

The Pinocchio pretotype is a non-working model you pretend is alive, to test whether you'd genuinely use the real thing. The archetypal story is a handheld-organizer inventor who carved a block of wood the size of the device, carried it in his pocket, and "checked" it during the day — pretending to look up appointments — to find out whether he'd actually reach for such a gadget before a single circuit was built.

Use it for personal-behavior questions. When the real risk is "will this fit into someone's daily habits," a dead prop you carry around answers it for the price of an afternoon and some cardboard.

The Fake Door: measure who tries to walk through

The Fake Door pretotype puts up an entrance to a product that doesn't exist and counts how many people try to enter. The door can be an ad, a landing page, a pricing button, or a menu item; clicking it leads to an honest "coming soon" or a short waitlist form rather than a real purchase.

Use it to size raw interest quickly and cheaply. A fake door measures attention, which is a genuine signal but a weak one — a click is low on the skin-in-the-game ladder. Pair it with something that asks for more commitment before you trust the result.

The Façade and One-Night Stand: deliver the experience manually

The Façade and One-Night Stand pretotypes deliver a real experience to real customers without building any lasting infrastructure. A Façade presents a working storefront, then fulfills each order by hand behind the scenes — the customer gets the product, you get proof of real transactions. A One-Night Stand offers the full service just once or briefly, the way the founders behind a now-famous lodging marketplace first rented out an air mattress in their apartment for a single event.

Use these when you need to see people pay, not just click. They generate high-caliber evidence — money changing hands — precisely because you carry the operational burden manually instead of automating it away.

The YouTube, Infiltrator, and Relabel: borrow reality to test demand

These three pretotypes borrow an existing surface to make your idea feel real. A YouTube pretotype is a short video that shows the product working — even if the "working" is staged — and measures the response; a well-known file-syncing tool famously used a demo video to drive a waitlist surge before the product was finished. An Infiltrator smuggles your product onto a real shelf inside someone else's store to see if strangers buy it in a genuine purchase environment. A Relabel puts your intended label on an existing product to test whether demand hangs on your specific positioning.

Use them to test demand inside a real context you didn't have to build. The realism of the surrounding environment is what makes the resulting data trustworthy.

Read the TRI Meter: turning evidence into a build-or-kill decision

The TRI Meter — Savoia's "The Right It" meter — is the gauge that turns a pile of scattered experiments into a single go/no-go read. Instead of asking "do I feel good about this idea," it asks "which way is the accumulated evidence pointing," and it moves only in response to real data.

Start skeptical and let the data move the needle. Each pretotype you run produces a result that nudges your confidence toward The Right It or the Wrong It. One encouraging signal doesn't flip the verdict; a consistent pattern of skin-in-the-game evidence does. The meter is a discipline against two failure modes at once — abandoning a good idea after one soft result, and clinging to a bad one because you're emotionally invested.

The needle should reflect the caliber of your evidence, not its volume. A hundred waitlist emails barely move it; ten paid pre-orders move it a lot. Weighting your experiments by how much skin was in the game keeps the gauge honest. A tool like Edmired can hold each experiment's result in one place so the pattern — not your mood on a given morning — is what you're reading.

The output is a decision, not a feeling. When the meter points clearly toward The Right It, you've earned the right to start building It right and to move on to the rest of the startup idea validation process. When it points the other way, you've saved yourself the most expensive mistake in the startup playbook. Either outcome is a win, because both were bought cheaply.

Common ways founders fool themselves while pretotyping

Most bad validation isn't dishonest — it's motivated reasoning dressed up as evidence. Knowing the traps in advance is the cheapest way to avoid running experiments that were rigged to say yes from the start.

• Asking leading questions. "Would you use an app that saves you time?" only ever gets a yes. You're back in Thoughtland collecting opinions, not observing behavior. Design tests that watch what people do, not what they say they'd do.
• Mistaking weak signals for validation. Likes, shares, and "cool idea" comments feel like traction but sit at the bottom of the evidence ladder, as the breakdown of how strong different validation evidence really is makes clear. Counting them as proof is how founders convince themselves a Wrong It is a Right It.
• Leaning on other people's data. A market report about an adjacent category is not YODA. It describes a world that already exists, not the new thing you're proposing. Only your own fresh data answers your own question.
• Testing a claim you never scaled down. A grand market-wide hypothesis can't be tested today, so it never gets tested at all. If you haven't hypozoomed to a local, immediate xyz version, you don't have an experiment — you have a wish.
• Falling in love with "It right" too early. The pull to start building, designing, and polishing is strong, and it feels productive. But effort spent making the Wrong It beautiful is the single most expensive mistake the Law of Market Failure predicts.
• Running one test and stopping. A single result — good or bad — is noise. The TRI Meter moves on a pattern of evidence, so a lone data point should update your view a little, not settle it.

Key Takeaways

• Pretotyping answers "should we build this," while prototyping answers "can we build it." Do them in that order, because it is far cheaper to discover the Wrong It in a day than after a year of flawless engineering.
• The Law of Market Failure says most new products fail even when built and marketed competently. Treat your own idea as unproven until real evidence says otherwise, rather than assuming success is the default.
• Escape Thoughtland by rewriting beliefs as an XYZ Hypothesis — "at least X% of Y will Z" — then hypozoom it down to a small, local version you can actually test this week.
• Trust only YODA: Your Own DAta, collected fresh and first-hand from real people. Other people's market reports describe a world that already exists, not the new thing you are testing.
• Weight every signal by its skin in the game. A paid pre-order predicts behavior; a "like" or a verbal "I'd buy that" costs the person nothing and predicts almost nothing.
• Match the pretotype to the riskiest belief. Mechanical Turk, Pinocchio, Fake Door, Façade, YouTube, One-Night Stand, Infiltrator, and Relabel each fake a different layer so you can test demand without building the product.
• Let the TRI Meter, not your emotions, decide. Move the needle only with real evidence, and count a clear "no" as a win, because it saved you the most expensive mistake in the playbook.

Frequently Asked Questions

What is the difference between pretotyping and prototyping?

Pretotyping tests whether an idea is worth building; prototyping tests whether you can build it well. A pretotype fakes the product's core experience in hours or days to measure demand, while a prototype is a real, rough working version that takes weeks or months and tests feasibility and quality. Pretotype first.

How do I write an XYZ hypothesis for my idea?

Fill in the sentence "at least X% of Y will Z," where X is a percentage threshold you commit to in advance, Y is a specific target market, and Z is a concrete action like buying or pre-ordering. Then hypozoom it down to a smaller, local, testable version you can run right now.

What counts as skin in the game in pretotyping?

Skin in the game is anything of value a person risks to signal real interest: money, time, personal information, effort, or reputation. A paid deposit or completed purchase is strong evidence; a "like," a survey answer, or a verbal "I'd use that" is weak, because it costs the person nothing to give.

How much time and money does pretotyping cost?

Far less than building the product — that is the entire point. Savoia frames pretotypes as tests that take hours or days and the smallest possible investment, versus the weeks or months a prototype demands. Many pretotypes, like a fake door or a Pinocchio, cost close to nothing beyond your attention.

Is pretotyping the same as building an MVP?

No. An MVP (minimum viable product) is still a real, working product you build, usually over weeks. A pretotype comes earlier and cheaper: it fakes the experience to test demand before you commit to building anything at all. Pretotype to decide whether the MVP is even worth starting.

Why do so many well-built products still fail?

Because of the Law of Market Failure: most new products fail in the market even when competently executed and skillfully marketed. The usual cause of death is a lack of genuine demand — a Wrong It — which no amount of engineering, design, or marketing can fix. Validating demand first is the only real safeguard.