Product Discovery Research method
Finds the success-critical assumption with the weakest evidence and tests it with the smallest decision-capable experiment.
A Riskiest Assumption Test starts from assumptions rather than a preferred MVP. Priority combines impact if wrong with current evidence. Thresholds, minimum data, and decision rules are set before testing; an interview count alone is not a validity criterion.
Which assumption would damage the idea most if false and currently has the weakest credible evidence?
Priority combines high impact if wrong with weak evidence. A fixed interview count is neither a threshold nor evidence of effect.
Assumptions
Decompose the idea into discrete desirability, viability, feasibility, and adaptability assumptions.
Prioritise
Assess impact if wrong and current evidence separately, then justify one priority.
Test
Set evidence source, metric, threshold, and minimum data before testing.
Decide
Apply the pre-set rule, update evidence strength, and determine the next risk-reduction step.
After each test, update evidence strength and select the riskiest remaining assumption.
Useful before substantial investment in a solution or business model. Clarify assumptions or conduct discovery first when the core question is still open.
Discrete assumptions with risk dimension, impact, evidence strength, and rationale.
Smallest suitable test with population, metric, threshold, result, and updated ranking.
Fictional worked example
Prioritise assumptions by impact and evidence, then connect the top item to a pre-set decision-capable test.
| Assumption | Assumption | Impact if wrong | Current evidence | Smallest suitable test | Metric, threshold, decision |
|---|---|---|---|---|---|
| A1 report exportUse in monthly review | Project leads use exported reports in their monthly review. | Very high: without review use, export misses its core value. | Weak: two internal opinions, no target-segment observation. | Concierge report for 12 suitable project leads with a real review date. | Pass: ≥8 bring it to review and ≥6 use a chart. Result 7/5: missed; revisit value and segment, stop build. |
Participant count is part of design, not a success criterion. Observable outcome, threshold, and evidence strength decide.
Start with the core question: Which assumption would damage the idea most if false and currently has the weakest credible evidence?
When an idea hinges on one critical unknown, it prioritizes testing that exact piece of uncertainty first. It separates problem, assumption, solution, and evidence. The result is captured as a prioritized list of assumptions, a test plan, and a result report.
It identifies the riskiest assumption and builds the smallest possible test around it. Learning energy is concentrated where a mistake would tip the venture over most strongly, instead of testing many comfortable side questions.
The Riskiest Assumption Test, often shortened to RAT, is a stricter version of the MVP idea. Popularized by Giff Constable, Rik Higham, and others, it asks for the one assumption whose falseness would most threaten the initiative. The right experiment is often smaller, cheaper, and faster than a classic MVP.
Collect all assumptions, sort them by uncertainty and leverage, then choose the top one. Define the experiment as small as possible.
The method guide walks execution through 4 phases, timing, 4 pitfalls, and clear stop criteria.
Open method guideYour perspective
Your feedback helps shape the next improvements to methodatlas.
Checklist for critical assumptions by risk and knowledge level.
Short plan for hypothesis, test design, success criteria, and decision.
We believe that ...
Who are we testing for?
What exactly will participants or users do?
We count the test as positive if ...
What can the test not prove?
If positive: ... If negative: ...
...
Plan Riskiest Assumption Test with real scope, suitable roles, and a clear decision objective.
To be confirmed
To be confirmed
Full learning cycle Time: 45 minutes for prioritisation and design; execution is separate
Complete this worksheet on paper or in your own document during the work.
| Assumption | What must be true? | low/medium/high + rationale | source · strength · recency | test · population · duration | metric · threshold · if/then |
|---|---|---|---|---|---|
| A1 report export |
Justify impact and evidence separately; freeze threshold before running.
Strategyzer: Test Your Business Assumptions
Expected artifact: Assumption map
Assumption · risk dimension · evidence source
Entry:
...
Expected artifact: Prioritised assumption
Impact · evidence strength · rationale
Entry:
...
Expected artifact: RAT protocol
Test · population · metric · threshold · decision
Entry:
...
Expected artifact: Decision and new ranking
Result · evidence strength · decision · next assumption
Entry:
...
...
Method guide: https://methodatlas.meierhoff-systems.de/en/methods/riskiest-assumption-test/run-sheet
Fitting alternatives with decision guidance, followed by other related methods.
Instead of Riskiest Assumption Test, choose Assumption Mapping when you want to order assumptions by uncertainty and risk and then test them in a targeted way.
Instead of Riskiest Assumption Test, choose Concierge MVP when you want to deliver the value manually and get learning signals before automating.
Moves customer problem, solution idea, and evidence toward a concrete result through "list assumptions", "map assumptions to dimensions", and "derive the test plan".
Separates discovery and delivery so that validated stories only move into implementation after learning loops.
Separates hypothesis, observation, insight, and next action into four fields after a test.
Condenses hypothesis, test, metric, and success threshold onto one card before execution.
