Innovation 101
Ideation & Prototyping

Assumption Mapping

Surfacing the beliefs a concept secretly depends on and sorting them by importance and uncertainty, so you test the one that could kill the idea first, before building.

Every idea rests on a stack of assumptions. One or two are both critical and completely untested. Find those, and test them before you build anything.

What it is

The beliefs a concept secretly depends on — sorted by how dangerous they are.

Assumption mapping is the discipline of surfacing the beliefs an idea secretly depends on, and then sorting them by risk so you test the dangerous ones first. Every concept rests on a stack of assumptions: that customers want this, that they will pay this price, that you can build it, that regulators will allow it, that partners will cooperate. Most of them feel safe. But usually one or two are both critical — if wrong, the whole idea collapses — and unknown — you have no real evidence either way — and those are the ones that quietly kill projects when a team builds for months before discovering that a foundational belief was false.

The method sorts assumptions along two axes. IMPORTANCE asks how much the concept depends on this assumption being true — would the idea survive if it turned out false? UNCERTAINTY asks how much evidence you actually have — do you know this, or are you just hoping? Plotting every assumption on this importance-by-uncertainty grid reveals the one quadrant that matters most: high importance and high uncertainty, the assumptions that are both critical and untested. These are the leap-of-faith assumptions, and they are exactly where your testing energy should go first.

The whole point is prioritisation of risk. It is tempting to validate the assumptions you are confident about (they are comfortable to confirm) or the ones that do not much matter (they are easy to check). Assumption mapping forces the harder, more valuable move: identify the belief that is both most consequential and least proven, and design the cheapest possible test for it, before committing to build. Test the thing that could kill the idea first, while it is still cheap to be wrong.

Place each assumption. The dangerous corner is critical and untested.

The dangerous corner is critical and untested. Test that first.

Each card is an assumption the concept rests on. Click it to see which quadrant it sits in and what to do about it — discovering that the high-importance / high-uncertainty corner is where your testing energy belongs.

When to deploy it

For de-risking a specific concept before building. Not for vague ideas or risk you will not act on.

Use Assumption Mapping when

  • You have a concept and are deciding what to test or prototype first, before committing to build.
  • You want to de-risk an idea by finding the belief most likely to be both critical and wrong.
  • A team is confident about an idea and you suspect that confidence rests on untested assumptions.
  • You are prioritising a backlog or set of experiments and need to aim testing energy at the highest-risk beliefs.

Do not lean on it when

  • ×The idea is too vague to have concrete assumptions yet — shape the concept first, then map what it depends on.
  • ×You are not actually willing to test (or kill the idea over) the leap-of-faith assumption. Mapping risk you will ignore is theater.
  • ×The relevant uncertainty is trivial or already well-evidenced across the board (rare, and worth double-checking).

The honest limit: assumption mapping identifies and prioritises risk; it does not resolve it. Placing an assumption in the leap-of-faith corner tells you what to test, not what the answer is — the test still has to be run. Its most common failure is dishonesty about the axes: rating a shaky assumption as “known” because admitting the uncertainty is uncomfortable, which quietly moves the real risk out of view. The method is only as good as the team’s honesty about how little it actually knows.

How it works

Six moves, from surfacing every assumption to testing the one that matters most.

01

Surface every assumption the concept rests on.

Brainstorm the full set of beliefs the idea depends on, across desirability (will people want it?), feasibility (can we build it?), and viability (will it work as a business?). Push past the obvious to the buried assumptions no one has said out loud. The dangerous belief is often a buried one.

02

Rate each on importance.

For each assumption, ask: if this turned out to be false, would the idea survive or collapse? An assumption the idea cannot survive without is high importance. This places it on the vertical axis.

03

Rate each on uncertainty — honestly.

Ask how much real evidence you have: do you actually know this, or are you assuming it? This is the axis teams fudge. Be rigorous about admitting what is genuinely untested, because comfortable overconfidence here hides the real risk. For each "known" assumption, ask what the actual evidence is.

04

Plot them and find the leap-of-faith corner.

Place every assumption on the importance-by-uncertainty grid. The high-importance / high-uncertainty quadrant holds the leap-of-faith assumptions: the critical, untested beliefs that should be tested first. This corner is the entire point of the method.

05

Design the cheapest test for the riskiest assumption.

Take the top leap-of-faith assumption and design the smallest, fastest experiment that could prove it wrong. Favour a rough prototype, a concept test, a fake-door, or a manual simulation over an expensive build. The Zappos model — photograph shoes, build a simple site, buy at retail and ship manually — is the benchmark: answer the belief in days with no engineering investment.

06

Test, learn, and re-map.

Run the test, update your evidence, and re-plot. A tested assumption moves out of the uncertainty zone; the next-riskiest belief becomes the priority. Assumption mapping is iterative: it continually points testing at the highest remaining risk.

Best practices

What good looks like — and the mistakes that prevent it.

When it goes well

  • The team surfaces the buried assumptions, not just the obvious ones, across desirability, feasibility, and viability.
  • Uncertainty is rated honestly, with real rigour about admitting what is genuinely untested.
  • Testing energy goes to the leap-of-faith corner — critical AND unknown — not to comfortable or trivial assumptions.
  • The riskiest assumption gets the cheapest possible test, run before committing to build.
  • The map is treated as living: after each test, evidence is updated and the grid re-plotted so testing keeps aiming at the highest remaining risk.

The mistakes, and how to avoid them

Being dishonest about uncertainty.

Rating a shaky belief as "known" because the uncertainty is uncomfortable moves the real risk out of view. Be ruthlessly honest about what you actually have evidence for. For each "known" assumption, ask: what is the actual evidence?

Validating the comfortable assumptions.

Teams love confirming what they are already confident about. Aim at the critical-and-unknown corner instead — the belief that could kill the idea, not the one that will confirm it.

Testing the trivial.

Spending scarce experiments on low-importance assumptions (however uncertain) wastes energy. Prioritise by both axes, not just uncertainty. Nice-to-know is never the priority.

Listing only the obvious assumptions.

The dangerous belief is often a buried one no one has said aloud. Push past the surface to the assumptions the idea silently depends on — the beliefs that feel too obvious to question.

Mapping risk you will not act on.

Identifying a leap-of-faith assumption and then building anyway, without testing it, is theatre. Map only if you are willing to test — and possibly kill the idea.

Logistics

How to run it well — and the discipline that makes the map trustworthy.

Assumption mapping works best with a cross-functional group, because different people hold different assumptions and see different risks, and because a team that maps its own idea alone tends to be blind to its own leaps of faith. Explicitly invite the sceptic’s view: “what would have to be true for this to work, and how do we know it is?”

Separate surfacing from rating

First get all the assumptions on the table (diverge), then rate them on importance and uncertainty (converge). Mixing the two — judging assumptions as they are named — suppresses the buried ones that matter. Name everything first, evaluate nothing.

Push hard on the uncertainty axis

Because this axis is the one teams fudge, build in a deliberate challenge: for each "known" assumption, ask what the actual evidence is. If the honest answer is "we just think so," it belongs in the uncertain column. This single discipline is what makes the map trustworthy.

Design cheap, fast tests for the top assumptions

The output of the map is a test plan for the leap-of-faith corner. Favour the smallest experiment that could disprove each critical assumption — a rough prototype, a concept test, a fake-door, a manual simulation — over an expensive build. The Zappos-style cheap test is the model: answer the belief in days with no engineering investment.

Use common tools as examples, not recipes

Whiteboards and sticky notes are the classic format; digital canvases (Miro, Mural, FigJam — named as common examples, not endorsements) work well for distributed teams. The tool is not the method; the discipline of honest placement on the two axes is.

Keep it visible and re-map after tests

Maintain the grid where the team can see it, and update it as tests return evidence. A tested assumption moves; the next-riskiest becomes the focus. Assumption mapping is a living document, not a one-time exercise.

AI and this method

AI will list your assumptions in seconds. It cannot tell you which one is the leap of faith — because that depends on what you actually know.

Toggle between modes to see where AI helps build the map — and why judging the two axes stays human.

In-depth example

The same scenario. Two approaches — one finds the leap of faith, one misses it.

A founder mapping the assumptions behind an online shoe store. The genuine leap of faith: will people buy shoes without trying them on? Toggle between the traditional approach and a hypothetical AI-first approach to see what each finds — and what each misses.

Shared scenarioA founder wants to build an online shoe store. The concept rests on many assumptions — but which one is the genuine leap of faith? Will people actually buy shoes online without trying them on first? That is the Zappos origin question: critical (the whole business hinges on it) and untested (no evidence exists either way).

How the mapping worked

The team surfaced every assumption the online shoe store depended on — across desirability (do people want this?), feasibility (can it be built?), and viability (do the economics work?). Most assumptions were important but well-evidenced. Shoes can be photographed and displayed online. People already buy things online. Delivery and returns are operationally viable.

But one assumption sat alone in the leap-of-faith corner: critical, and genuinely untested. Everyone already agreed that shoes had to be tried on before buying. Whether people would actually purchase them online, sight unseen, was a belief nobody had tested — and the entire business model depended on the answer.

The assumptions mapped

MONITOR

Shoes can be sold online

Important, and reasonably evidenced — e-commerce was already established for other goods. Safe to proceed on, but not the dangerous bet.

MONITOR

Delivery and returns are operationally viable

Known. Other categories already shipped and accepted returns. Safe.

LEAP OF FAITH

People will buy shoes without trying them on

Critical — the concept collapses without this being true — and completely untested. No evidence existed either way. This is the one.

MONITOR

Customers prefer a wide online selection

Plausible from other categories; important but not as unknown as the core question.

IGNORE

Website can be built and run reliably

Known and not a differentiator. Move on.

The cheapest test for the riskiest assumption

Rather than build a full e-commerce platform to find out whether people would buy shoes online, Nick Swinmurn designed the smallest possible test for exactly that assumption. He photographed shoes in local shoe stores, posted them on a simple website, and when someone ordered, bought the shoes at retail price and shipped them himself.

He lost money on every sale. That was not the point. The test answered the one leap-of-faith assumption in days, with no engineering investment: people would buy shoes online, without trying them on first. The Zappos business model was validated by a test aimed at the belief that could have killed the idea — not at the comfortable ones.

Testing the thing that could kill the idea first, while it was still cheap to be wrong, is the whole discipline.

Frameworks

Where Assumption Mapping shows up.

A de-risking method that sits between having an idea and testing it, Assumption Mapping maps to the build-and-test framing moments of frameworks — where a concept is selected and the team must decide what to test before committing.

Related methods

What to combine with Assumption Mapping.

A close cousin at a different altitude — worth distinguishing. Assumption Mapping surfaces the risky assumptions behind a SPECIFIC concept (what does this idea depend on, and which belief is the leap to test?). Orthodoxies surfaces the unquestioned assumptions of an entire INDUSTRY (what does everyone in this field take for granted, and what if the opposite were true?). Assumption mapping de-risks an idea; orthodoxies breaks an industry convention to find opportunity.
The natural next step: the leap-of-faith assumption is exactly what a rapid prototype should be built to test first, cheaply. Assumption mapping identifies the target; the prototype is the test instrument. The output of the map is a test design brief for the prototype.
The disciplined way to test a desirability assumption from the map with real users. If the leap-of-faith assumption is about whether people want the thing, concept testing is the appropriate experiment to run.
Upstream: the ideas generated in ideation are what you then map for their riskiest assumptions before building. Crazy 8s generates eight concepts in eight minutes; Assumption Mapping identifies which of those concepts rests on the least dangerous bet.
Complementary: the Balanced Breakthrough names the three lenses (desirability, feasibility, viability). Assumption Mapping surfaces and prioritises the specific untested beliefs within each lens for a given concept. They work together — the Breakthrough frame tells you which lens is missing; Assumption Mapping finds the specific critical-and-unknown belief within it.

Sources & further reading

The work behind this method.

Testing Business Ideas

David Bland and Alexander Osterwalder (2019)

The definitive source for the assumptions map — importance by evidence — and cheap experiment design. Bland and Osterwalder's framework structures the full cycle from assumption to experiment to learning, making this the closest companion to the method as described here. Their assumption map is an importance-by-uncertainty grid; their experiment library shows how to design the cheapest possible test for each type of assumption.

The Lean Startup

Eric Ries (2011)

The source of the "leap-of-faith assumption" concept and the discipline of testing the riskiest belief first. Ries argues that the job of an early-stage team is not to build a product but to test whether the foundational beliefs behind it are true — and that the MVP is an instrument for testing the leap-of-faith assumption, not a minimum version of the final product.

The Right It

Alberto Savoia (2019)

On testing whether an idea is worth building before you build it. Savoia's "pretotype" approach is the practical toolkit for the cheap experiment design that assumption mapping calls for: the smallest, fastest, cheapest test that could answer the leap-of-faith question before any engineering investment.