Pretotyping · Built at Google, taught at Stanford

What is pretotyping?

Pretotyping validates a product idea in days, not months — testing whether people actually want it before you build it. Here's how it works, how it differs from prototyping, the 10 methods, and the results enterprises see.

Coined by Alberto Savoia at Google, 2010 · $30M+ saved

Banked

$68M+

$38M generated $30M saved

Fund the best, save on the rest.

Definition · Answer first

What is pretotyping?

Pretotyping is a set of techniques for validating a product idea quickly and cheaply — testing whether people actually want it before you build it. Coined from “pretend” + “prototype” by Alberto Savoia at Google in 2010, it answers “should we build this?” using real customer behaviour, not opinions. It’s the fastest route to idea validation: enterprises using it (Google, AGL, PEXA) have saved $30M+ by killing bad ideas before development.

90%

of new products fail in the market

Harvard Business School / Christensen

88%

of AI pilots never reach production

MIT, 2025

$30M+

saved by enterprises killing bad ideas early

Exponentially client results

Why it works · The Law of Market Failure

The Law of Market Failure

Most new ideas fail in the market — even when they’re built well. Alberto Savoia calls this the Law of Market Failure, and the only defence is to test demand objectively and fast, before you invest in building. That means trusting YODA (Your Own DAta) on what people actually do over the confident opinions of “Thoughtland,” and insisting on skin in the game — real clicks, sign-ups or money — as proof. It’s the same discipline behind validating any business idea.

Pretotyping vs prototyping

Pretotyping tests demand. Prototyping tests the build.

The two are easy to confuse, but they answer different questions. Pretotyping comes first — it tells you whether an idea is worth building at all. Prototyping comes later, once you already know people want it. Read the full pretotyping vs prototyping breakdown →

DimensionPretotypingPrototyping
Question it answersShould we build this? Will anyone want it?Can we build this? Does it work?
What you testReal market demand and customer behaviourFunctionality, feasibility and usability
What you buildLittle or nothing real — a fake door, a façade, a stand-inA working approximation of the actual product
Typical costHours to days, often near-zero spendWeeks to months, real engineering budget
When you use itBefore you commit to buildingAfter you've decided it's worth building
The process

How to run a pretotype

Five steps take you from a fuzzy idea to a Stop, Pivot or Proceed decision backed by real behaviour.

01

Frame the Key Market Hypothesis

Name the market belief you're actually betting on — who the customer is and what you expect them to do.

02

Say it with numbers (the XYZ hypothesis)

Rewrite it as 'At least X% of Y will do Z.' A measurable bar turns a hope into a test you can pass or fail.

03

Set your pass/fail threshold

Decide the number that means proceed — and the number that means kill — before you run it, so the data can't be argued away.

04

Shrink it to a Hypozoom

Build the smallest, cheapest believable version that still produces real behaviour, then pick a method — Fake Door, Mechanical Turk, and so on.

05

Run it, then Stop, Pivot or Proceed

Collect real skin-in-the-game data — clicks, sign-ups, money — and let the threshold, not opinion, decide the idea's fate.

Watch

Pretotyping in two minutes

Leslie Barry on the one question pretotyping answers — should we build it? — and why it beats prototyping for testing demand.

From the free Learn Pretotyping series. More on Leslie Barry's channel →

Worked examples

Pretotyping in the wild

Six pretotypes that answered 'should we build it?' before anyone spent big — the moves behind Palm, IBM, Airbnb and more. See all examples →

01

Palm Pilot

Jeff Hawkins carried a block of wood for weeks, pretending to check his calendar on it, to prove he'd actually use a pocket organiser before building one.

02

IBM speech-to-text

IBM faked a working dictation machine with a hidden typist. Buyers loved the demo but quit within hours — sore throats, too much noise — saving years of wasted build.

03

Second Day Sushi

Our own test: 'at least 20% of sushi buyers will buy day-old sushi at half price.' Almost nobody did, so we killed the idea in days, not quarters.

04

Airbnb

The founders rented out air mattresses in their own apartment to test whether strangers would really pay to sleep in a spare room.

05

Google Glass

A concept video stood in for the finished product, gauging real demand and sign-ups long before any mass manufacturing.

06

Skincare AI

A beauty brand tested an AI makeup recommender with a Fake Door, then a Mechanical Turk, adding cost only after each result proved demand.

The methods

The 10 pretotyping methods

Each method is a fast, cheap way to collect real behavioural evidence — proof people will act, not just say they like the idea. See the full methods guide →

01

The Mechanical Turk

Replace expensive technology with humans behind the scenes to test demand before you automate — the core of a Wizard of Oz test.

02

The Pinocchio

Build a non-functional, lifeless version and use it as if it were real to test how often you'd reach for it.

03

The Fake Door

Advertise the offering and measure clicks or sign-ups for something that doesn't exist yet.

04

The Façade

Stand up a convincing front-end or landing page and count the people who try to buy or sign up.

05

The YouTube

Demonstrate the idea in a short video and gauge interest from views, shares and sign-ups.

06

The One-Night Stand

Offer the service for real, but only once or for a short window, to see if anyone actually shows up.

07

The Infiltrator

Slip the product into a real retail or usage environment and watch whether people engage with it.

08

The Re-Label

Put a new label on an existing product to test whether a new positioning or use-case lands.

09

The Provincial

Test in one small, local market first before assuming the idea will scale everywhere.

10

The Minimum Viable Pretotype

Strip the idea to the single riskiest assumption and test only that, as cheaply as possible.

FAQ

Pretotyping, answered

What is pretotyping?

Pretotyping is a set of fast, cheap techniques for validating whether people actually want a product idea before you build it. Created by Alberto Savoia at Google in 2010, it uses real customer behaviour — clicks, sign-ups, purchases — to answer 'should we build this?' rather than relying on opinions or surveys.

What's the difference between pretotyping and prototyping?

Pretotyping tests demand: will anyone want this? It comes first and costs almost nothing. Prototyping tests the build: can we make it work? It comes later, once you already know there's demand, and requires real engineering effort. Pretotyping stops you from prototyping ideas nobody wants.

How do you write a pretotyping hypothesis?

Use the XYZ format — 'At least X% of Y will do Z' — so success is a number, not an opinion. For example: 'At least 20% of packaged-sushi buyers will buy day-old sushi at half price.' Setting that threshold before you run the test means the result can't be argued away afterwards.

What is the Law of Market Failure?

It's Alberto Savoia's observation that most new ideas fail in the market even when they're competently built — the market simply doesn't want them. The only reliable defence is to test demand quickly and objectively, with real behavioural data (YODA — Your Own DAta), before investing in development. Pretotyping exists to fight this law.

What is a Hypozoom?

A Hypozoom is the smallest, cheapest believable version of your idea that still produces real customer behaviour. Instead of building the whole product, you zoom in on the single riskiest assumption and test only that — often with a Fake Door or Mechanical Turk — so you get a genuine signal for almost no cost.

Who invented pretotyping?

Pretotyping was coined and developed by Alberto Savoia, a former Engineering Director and Innovation Agitator at Google, around 2010. He wrote the book 'The Right It', and the method is now taught at Stanford and used by enterprise innovation teams worldwide.

What are the pretotyping methods?

The classic pretotyping methods include the Mechanical Turk, the Pinocchio, the Fake Door, the Façade, the YouTube, the One-Night Stand, the Infiltrator, the Re-Label, the Provincial, and the Minimum Viable Pretotype. Each is a different way to collect real behavioural evidence quickly and cheaply.

How is pretotyping different from an MVP?

An MVP (minimum viable product) is still a real, working product you build and ship — it tests whether people will use and pay for a functional version. A pretotype is built before any product exists and often isn't functional at all; it tests raw demand. Pretotyping happens earlier, faster and cheaper than building an MVP.

Bring us one idea. We'll tell you if it's worth building.

Twenty minutes with Leslie Barry. We look at one idea on your roadmap, the evidence you already have, and whether pretotyping is a useful next step. No pitch.