Complete firstAssumption Mapping
An assumption map with all assumptions critical for the idea, sorted by uncertainty and leverage, is available.
Without: Without map, a random assumption is tested and the biggest uncertainty lever remains hidden.
RunsheetProduct Discovery
Riskiest Assumption Test
ComplexityMedium
Time1-2 Wochen pro Iteration
Participants2-6
FormatWorkshop + async
MaturityEstablished
01
Prerequisite
What needs to be finished first
Complete firstExperiment Canvas
Team can structure experiment with hypothesis, metric and threshold (Canvas method known or applied in parallel).
Without: Without experiment structure, measurable success criterion is missing and test produces anecdotal data.
02
Preparation
What needs to be ready before start
Materials
Assumption list with uncertainty/leverage scoring; Experiment Canvas for structure; minimal test setup (landing page builder, prototype tool, concierge script); tracking tool for metrics; budget approval.
People / roles
One discovery lead (Product Manager or founder); one designer/builder for test setup; one engineer for technically complex tests; one data analyst for metric evaluation.
Pre-read
Idea or initiative with value proposition; assumption map with top-3 candidates; target group and recruiting channel; available test tools; budget and iteration timeframe.
Time needed
1-2 weeks per iteration
Setup
Assumption map as quadrant (axis uncertainty, axis leverage). Top-right (high uncertainty, high leverage) marked as RAT candidate. Experiment Canvas for selected assumption. Choose test tool according to assumption type (value, usage, technology, market).
03
Core question
The one question this method answers
Which assumption, if false, would endanger the initiative most, and what is the smallest possible test to examine this in 1-2 weeks?
04
Flow
Marker: Phase
| Step | Duration | Action | Hint |
|---|---|---|---|
1Phase 1: Collect and rate assumptions | 60-90 min | Collect all assumptions critical for idea (value, usage, feasibility, market). Rate uncertainty (0-3) and leverage (0-3) per assumption. Place in quadrants. | Anyone finding no assumptions is ignoring them. Common categories: will customers pay, will customers find us, can we build it, is market large enough. |
2Phase 2: Choose top assumption | 15 min | Choose one assumption from top-right quadrant (high uncertainty, high leverage) as RAT. Document rationale. Clear selection statement: "We test this assumption now." | Maximum one assumption per test iteration. Parallel tests make causes inseparable. Other assumptions wait in pipeline. |
3Phase 3: Design minimal test | 1-3 days | Smallest working test (landing page, fake door, Wizard of Oz, concierge, survey). Fix threshold before build. Setup time under 3 days ideal, max 1 week. | Temptation: too-large test design. Rule of thumb: if test needs more than 1 week, assumption is too large or design too fat. Cut smaller. |
4Phase 4: Run test | 3-7 days | Test live. Collect data. Daily short evaluation, final evaluation after period. Watch assumption-killer signals (stop early if clearly invalidated). | Do not stop test earlier than planned unless assumption is clearly refuted (for example 0 signups at 1000 visits). Then early stop and pivot discussion. |
5Phase 5: Pivot, persevere or stop | 30-60 min | Result against threshold: success = Persevere (test next assumption), mixed = Re-Design, failure = Pivot (change value proposition) or Stop. Document decision. | Pivot is victory, not defeat. Early invalidation saves months. Persevere without new tests is dangerous; keep processing assumption pipeline. |
05
Artifact
What comes out at the end
Form
Assumption map snapshot, selected RAT assumption with rationale, Experiment Canvas, test setup documentation, raw data, evaluation against threshold, decision with date.
Versioning / ownership
Own ID and date per RAT. Assumption map with version state, mark processed assumptions. Archive test result (do not delete) even on pivot, as later iterations need reference.
Tool alternatives
- Miro or FigJam for assumption map
- Strategyzer Test Card for RAT
- Notion template Experiment + RAT
- Carrd or Webflow for landing page tests
- Maze or UserTesting for prototype validation
assumption-map-checklist.md
Checklist for critical assumptions by risk and knowledge level.
- Formulate assumption as a testable statement
- Assess risk
- Assess level of knowledge
- Add data source or test idea
- Mark riskiest assumption
- Define success criterion
- Set owner and date
experiment-plan-markdown.md
Short plan for hypothesis, test design, success criteria, and decision.
Experiment Plan
Hypothesis
We believe that ...
Target audience
Who are we testing for?
Test design
What exactly will participants or users do?
Success criterion
We count the test as positive if ...
Risks and limits
What can the test not prove?
Decision afterward
If positive: ... If negative: ...
Owner and date
...
06
Example output
Concrete filled scenario, fictional example
riskiest-assumption-test-beispiel.md
Concrete filled scenario, fictional example
Riskiest Assumption Test - Receipt Preclassification, Iteration 1, 2026-05-18
Idea: SaaS tool for solo tax advisors with AI-supported receipt preclassification.
Assumption map (Top 5)
- Solo tax advisors pay >29 EUR/month for it. (U=3, L=3) <- RAT
- AI preclassification has >80% hit rate. (U=3, L=2)
- DATEV interface works without custom adapter. (U=2, L=3)
- Target group reachable through LinkedIn Ads. (U=2, L=2)
- Receipts are mostly submitted as PDF. (U=1, L=2)
Selected RAT: Assumption 1 (willingness to pay). Rationale: highest uncertainty leverage; without this assumption business model breaks.
Test: Landing page with value proposition, price 29 EUR/month visible, smoke button "Reserve now". Traffic source: LinkedIn Ads, budget 500 EUR, target group tax advisors 1-3 employees.
Threshold: >3% conversion on reserve button with at least 500 visitors.
Test period: 2026-05-20 to 2026-05-27.
Result: 612 visitors, 24 reservations = 3.9% conversion. Threshold exceeded.
Decision: Persevere. Next RAT: Assumption 2 (AI hit rate) through concierge test with 10 reserve users. Owner: @anna. Start: 2026-06-03.
07
Pitfalls
Recognize symptoms and steer against them
Trap
Multiple assumptions tested at once
Symptom
Test mixes value proposition and distribution; on failure it is unclear which assumption was false.
What to do
One assumption per test. If distribution and value must both be tested, split test into two phases. Clean assumption isolation is methodological core.
Trap
Wrong RAT choice
Symptom
Test checks assumption with high leverage but low uncertainty (for example database works).
What to do
Rate uncertainty honestly: what do we really know, what do we guess. Low uncertainty needs no test. RAT is always in top-right of quadrant.
Trap
Threshold set afterward
Symptom
Result arrives, stakeholder finds rationale why 1% is already success.
What to do
Fix threshold in writing before test. After-test discussion: was threshold wrong, yes or no. If yes, re-test with corrected threshold, no reinterpretation.
Trap
Overengineered test
Symptom
Instead of 2-day landing page, 6-week prototype built, iteration speed collapses.
What to do
Test effort proportional to assumption. Rule: if setup time > test runtime, setup is too large. Choose smallest sufficient format.
Trap
Persevere without pipeline
Symptom
One assumption confirmed, team jumps into full build mode, other assumptions remain unconfirmed.
What to do
Work assumption pipeline. Persevere means next RAT, not build sprint. After 3-5 confirmed RATs, MVP build is defensible.
08
Stop criteria
Done signals checkable in under a minute
Idea has no critical assumptions left (all confirmed), RAT is not next step.
Top assumption is not falsifiable (too vague), test would be speculation.
Test effort exceeds investment budget for discovery phase.
Target group for test not reachable, result would not be meaningful.
Stakeholder does not want outcome-open test, decision already made.
Already in build phase, RAT would only provide proof, no course correction.
Finished the runsheet?
Go to the profile for purpose, similar methods, and sources or continue to the next method in the catalog.