Assumption Mapping

What happens

List beliefs about users, value, feasibility, and viability as testable statements.

Facilitator does

Write the first assumption on the board yourself before asking the room to contribute — it lowers the activation energy and models the format. Each assumption should be written as a belief that could be false: "Users will pay more for faster processing" not "users want speed." When participants write conclusions, ask: "What would have to be true for this to hold?" That turns a conclusion into an assumption.

Participants do

Generate beliefs quickly, one per note, written in a form that could be tested or disproven. Speed matters here — get raw material on the board before the room starts editing itself.

Output / signal of success

15–30 distinct beliefs are visible, each written in a form that could be tested or disproven. No two notes describe the same belief. The board contains uncomfortable items, not just safe ones.

What happens

Place each assumption on a 2x2 by impact on outcomes and certainty of evidence.

Facilitator does

Place the first assumption yourself and explain your reasoning aloud: "I put this here because we have no data on it and if we're wrong it changes the business model." This models the two-axis judgment explicitly. For contested placements, don't resolve by consensus — ask: "What evidence would move this?" If nobody can name evidence, it belongs in the high-uncertainty quadrant regardless of how confident the room feels.

Participants do

Place items and challenge placements with evidence, not opinion. If you want to move an item, name what evidence you have for a different placement.

Output / signal of success

The matrix has real differentiation — a cluster in high-impact/high-uncertainty, some items clearly low-impact, some clearly high-certainty. The map is not flat. If everything sits at moderate certainty, the room is avoiding the judgment.

What happens

Select high-impact, high-uncertainty items as test priorities by asking what happens if each is wrong.

Facilitator does

For each assumption in the high-impact/high-uncertainty quadrant, ask: "If this is wrong, what would we stop building or do differently?" Items where the answer is "nothing much" can be parked — they're not critical. Items where the answer forces a pivot need test plans. If the room resists moving assumptions to high-uncertainty, ask: "What single piece of evidence would make us less certain about this?" If nobody can name one, the certainty rating isn't evidence-based — it's hope.

Participants do

Apply the consequence test honestly — not "could this be wrong in theory" but "if this is wrong, what changes about the plan." Items that wouldn't change the plan don't need testing.

Output / signal of success

3–6 assumptions are selected as test priorities, each with a stated consequence if wrong. The room agrees these are the beliefs that would most change the plan if they failed.

What happens

Assign specific experiments and owners for the top critical assumptions.

Facilitator does

For each selected critical assumption, require three things before closing: who will test it, what a confirming result looks like, and what a disconfirming result looks like. Vague plans ("we'll talk to users about this") are not accepted. Push until the test is specific: "Which users, what question, what answer would change your decision about this assumption?"

Participants do

Name a test that is lightweight enough to run in two weeks. If the right test would take months, ask whether a proxy test — something that gives directional signal quickly — can be run first.

Output / signal of success

Each critical assumption has a named owner, a test method, and explicit criteria for what confirming and disconfirming results look like. The room can describe what they will do in the next two weeks without additional planning.