What needs to be finished first
A hypothesis list with clearly phrased assumptions is available and sorted by risk.
What needs to be ready before start
Matrix template (assumptions as columns, test formats as rows) as board or printout; assumption list; format description with typical effort and output; examples from Strategyzer/Testing-Business-Ideas.
A discovery lead who walks the matrix, designer and engineer for feasibility judgment, a data analyst for metric questions, and optionally a coach for format choice.
Assumption list with dimensions (Desirability/Feasibility/Viability); list of available test formats with short descriptions; budget and time range for each test.
45-75 min
Set up matrix: assumptions vertically, test formats horizontally (Interview, Survey, Fake-Door, Landing Page, Concierge, Wizard-of-Oz, Prototype, Smoke Test, A/B). Keep format descriptions visible as banners.
The one question this method answers
Which test format best fits each assumption, and which mix of formats tests the open assumption pipeline most efficiently?
Marker: Phase
| Step | Duration | Action | Hint |
|---|---|---|---|
1Phase 1: Categorize assumptions and dimensions | 15 min | Clarify the dimension for each assumption (Desirability = will someone want it, Feasibility = can we build it, Viability = is it economically viable). Mark assumptions on cards. | Assumption dimension drives format choice. Desirability rarely fits smoke test, Feasibility rarely fits survey. Clear assignment avoids format reflexes. |
2Phase 2: Explain format options | 15 min | For each test format explain output, effort, and suitable dimensions. Add examples from Testing-Business-Ideas or previous tests. | Participants often only know two to three formats. Expanding repertoire is part of the workshop value. Present at least seven to ten formats clearly. |
3Phase 3: Fill the matrix | 20-30 min | Discuss suitable format for each assumption. In each matrix cell, note the rationale briefly. A format can cover several assumptions, but one assumption should get a primary format. | Participants who select the same format for every assumption (for example always interview) are not using the repertoire. Ask for alternatives. |
4Phase 4: Derive test plan | 10-15 min | Convert top-five assumptions with assigned formats into a test plan. Define owner, start date, and budget per test. Limit parallel tests to max three. | A test plan without owner and date is only a wish list. The workshop ends with a concrete calendar plan, not the matrix alone. |
What comes out at the end
Matrix as board export or table, plus a Markdown test plan with assumption, format, rationale, owner, start date, and budget. Link to hypothesis prioritization and an experiment canvas per test.
Snapshot per discovery sprint. Document completed tests with outcome and date. Capture format learnings (what worked) in the team wiki as lessons learned.
Compact working template for MVP Test Matrix with context, input, output artifacts, and next step.
| Element | Description | Rating | Evidence | Owner | Next step |
|---|---|---|---|---|---|
| 1 | |||||
| 2 | |||||
| 3 |
What consequence follows from the matrix?
Concrete filled scenario, fictional example
Concrete filled scenario, fictional example
Assumption dimensions
Matrix (simplified)
| Assumption | Dimension | Format | Rationale |
|---|---|---|---|
| H1: EUR 29 monthly pricing | Desirability/Viability | Landing Page with Smoke Button | Fastest test for price reaction. |
| H2: Save more than 5 hours per week | Desirability | Concierge test | Real experience with manually classified backend. |
| H3: 80% match quality | Feasibility | Wizard-of-Oz | We answer manually and compare output against expectation. |
| H4: LinkedIn channel | Desirability | Ads campaign | Direct conversion signal and low setup cost. |
| H5: DATEV without adapter | Feasibility | Tech spike | Code spike, no user test needed. |
| H6: >5000 solo tax advisors in market | Viability | Market analysis + Survey | Secondary data plus quantitative interview. |
Test plan
Recognize symptoms and steer against them
Team uses the same format repeatedly (for example always user interview), regardless of assumption dimension.
The facilitator asks for alternatives. Require at least three different formats in the test plan. Expanding repertoire should be part of the workshop outcome.
A desirability assumption is tested with a tech spike, producing results that do not speak about user desire.
Assign a dimension explicitly per assumption. Format must match the dimension. Use a reference table: "this format works for dimension X."
Test plan lists five tests, each with four weeks of setup, and discovery timeline explodes.
Estimate rough effort per format. Align test plan with discovery capacity. Three small tests in two weeks are better than five tests in eight weeks.
Format is selected but there is no explicit question of what should be confirmed or refuted.
Before format selection, set a threshold per assumption. The format is a tool, the learning goal is the purpose. Without a learning goal it becomes a distraction.
The matrix is built and uploaded to a wiki, but no one starts tests.
End the workshop with test plan and owner list, not with matrix alone. Add calendar entries for each test with kickoff date.
Done signals checkable in under a minute
Go to the profile for purpose, similar methods, and sources or continue to the next method in the catalog.