Complete firstAssumption Mapping
A critical assumption or hypothesis is identified that is crucial for the initiative and testable.
Without: Without a clear assumption, the canvas becomes an activity checklist instead of a learning instrument.
RunsheetProduct Discovery
Experiment Canvas
ComplexityLow
Time30-60 min
Participants1-5
FormatWorkshop + async
MaturityEstablished
01
Prerequisite
What needs to be finished first
02
Preparation
What needs to be ready before start
Materials
Canvas template (Strategyzer, Lean UX, or custom); pens and stickies; shared document or Miro board; list of assumptions to test; examples of metrics and thresholds from prior experiments.
People / roles
A Discovery lead or Product Manager as canvas owner; a designer for test setup; an engineer for technical feasibility; optionally a Data Analyst for metrics sources.
Pre-read
Assumption list; known customer or user-segment data; available test tools (landing page, fake-door, prototype); budget and timeframe; existing analytics metrics.
Time needed
30-60 min
Setup
Use canvas as printout or board. Visible fields: hypothesis, riskiest assumption, test design, metric, success threshold, learning goal, decision rule, prerequisites. Use a previous experiment canvas as reference.
03
Core question
The one question this method answers
Which assumption are we testing with which experiment, at what threshold is the hypothesis confirmed, and what decision follows from the result?
04
Flow
Marker: Sektion
| Step | Duration | Action | Hint |
|---|---|---|---|
1Section 1: Learning goal and assumption | 10 min | Formulate learning goal as a question ("Do we want users to buy Feature X?"). Name the riskiest assumption behind it. Fill both fields. | If the learning goal is stated as "we want to build feature," it is an implementation intent, not a learning goal. The canvas is the wrong tool; use roadmap item instead. |
2Section 2: Hypothesis and success metric | 10 min | Write hypothesis as an if-then statement. Name success metric with source (analytics tool, survey, manual analysis). Define success threshold. | A hypothesis without threshold is wishful thinking. Set threshold before test, not after. Without source, a metric is not measurable. |
3Section 3: Test design and setup | 15 min | Choose test type (landing page, fake-door, prototype, concierge, Wizard of Oz). Describe setup: what is built, target group, distribution. Estimate effort and duration. | Prefer smallest working test. If a prototype is built where a landing page is enough, time is wasted. Effort should match assumption size. |
4Section 4: Decision rule | 10 min | Define before test: what happens on success, what on failure. Options: Pivot, Persevere, Stop, next experiment. List risks and prerequisites. | Without a decision rule, results are interpreted instead of applied. Define success and failure path before test, this is methodological protection. |
5Section 5: Review and approval | 5-10 min | Align canvas with stakeholder or sponsor. Confirm budget and owner. Set test start date. | If a sponsor does not accept threshold, clarify beforehand. Threshold changes after the test start devalue the experiment. |
05
Artifact
What comes out at the end
Form
Completed Experiment Canvas as document or board export with all sections, plus test plan with setup details, date plan, and owner. Linked to hypothesis backlog.
Versioning / ownership
One canvas per experiment with ID, date, status (Planned, Running, Completed). Add result section after test, do not overwrite. Link to follow-up experiments.
Tool alternatives
- Strategyzer test card template
- Miro or FigJam with canvas template
- Notion template with sections
- Confluence page with canvas structure
- Productboard or Avion with experiment feature
experiment-canvas-working-template.md
Compact working template for Experiment Canvas with context, input, output artifacts, and next step.
Experiment Canvas Canvas
Context
What is this method used for?
Core question
Which question should be answered at the end?
Input
Which data, observations, or materials are available?
Working area
- Area 1:
- Area 2:
- Area 3:
- Relationships / patterns:
Output artifacts
- Filled Experiment Canvas:
- Success metric:
Open questions
- ...
Next step
Owner, date, success signal.
06
Example output
Concrete filled scenario, fictional example
experiment-canvas-beispiel.md
Concrete filled scenario, fictional example
Experiment Canvas - Concierge Test for pre-classifying invoices, 2026-05-18
Learning goal: Do Solo tax advisors really save time if vouchers arrive pre-classified?
Riskiest assumption: Manual pre-classification by tax advisors takes >30 seconds per voucher; automated suggestions would reduce it by >50%.
Hypothesis: If 10 Solo tax advisors receive pre-classified vouchers for 1 week, they save on average >5 hours compared to the previous week.
Metric: Self-reported processing time per voucher batch (before/after), source: notebook tracking via Notion template. Threshold: median savings >5 h/week.
Test design: Concierge test. Recruit 10 Solo tax advisors (Recruiter: Respondent.io, EUR 80/person). We classify vouchers manually in backend (1 person, 2 h/day), users see suggestions in existing UI mock.
Decision rule: Median savings >5 h => Persevere, build automated version (3 sprints). 3-5 h => re-test with another target group or setup. <3 h => Pivot to another value proposition.
Prerequisites: Recruit 10 users by 2026-05-25, manual backend classifier @lisa, UI mock @marcus, recruiting budget EUR 800 approved.
Test window: 2026-05-27 to 2026-06-03 (data collection), 2026-06-04 analysis.
07
Pitfalls
Recognize symptoms and steer against them
Trap
Learning goal is implementation intent
Symptom
Learning goal is "introduce Feature X" instead of "test assumption Y."
What to do
Formulate a question as learning goal. If implementation is already intended, assumption truth is assumed. Use sprint planning instead of canvas.
Trap
Missing or late threshold
Symptom
Hypothesis says "higher conversion" but no number, success is interpreted after the test.
What to do
Fix a concrete number before test (" >5%", ">50 signups"). Threshold is protection against confirmation bias. Without a threshold, no test.
Trap
Overbuilt test design
Symptom
Prototype is built for 6 weeks where a 2-day landing page would test the assumption.
What to do
Choose the smallest sufficient test format. Rule of thumb: effort proportional to assumption size. Use Riskiest Assumption Test as reference, not MVP.
Trap
No decision rule
Symptom
Result is interpreted, stakeholders look for evidence of preferred path.
What to do
Before test, fix decision for each result category (success, partial, failure). Do it in writing. Post hoc interpretation is confirmation bias.
Trap
Test without valid target group
Symptom
Participants are colleagues, friends, or random people, not actual target users.
What to do
Define target group explicitly before test. Recruit via platforms or own channels. With wrong target group, result is invalid, not just "better than nothing."
08
Stop criteria
Done signals checkable in under a minute
Assumption is not falsifiable (too vague or too broad); test would be theater.
No threshold can be defined, success/failure is not measurable.
Stakeholder is not willing to accept decision rule beforehand, test is re-interpreted later.
No recruitable target group, data basis would be distorted.
Test effort exceeds assumption weight (e.g. 8-week prototype for micro-hypothesis).
No metrics source exists or cannot be built during test window.
Finished the runsheet?
Go to the profile for purpose, similar methods, and sources or continue to the next method in the catalog.