What needs to be finished first
At least three documented data points from user interviews or analytics are available so the Users and Outcomes sections do not emerge from assumptions.
What needs to be ready before start
Lean-UX-Canvas template (8 fields) as whiteboard, Miro, FigJam, or A1 printout; sticky notes in two colors (content, assumption); pens; visible timer; research excerpts as pre-read.
One facilitator (PM, designer); cross-functional team (PM, designer, engineer, possibly business); sponsor for reviews and outcome setting; stakeholders as input providers.
Current business problem in one sentence; strategy or goal anchor; research snippets (interviews, data); known assumptions from earlier initiatives; desired workshop outcome (hypothesis list, first experiment).
2-3 h
Display the canvas template with fields 1-8: (1) Business Problem, (2) Business Outcomes, (3) Users, (4) User Outcomes, (5) Solutions, (6) Hypotheses, (7) Critical Assumptions, (8) Experiments. Keep sequence.
The one question this method answers
Which testable hypothesis links business problem, desired outcomes, users, solution idea, and critical assumptions so that the next experiment is clear?
Marker: Phase
| Step | Duration | Action | Hint |
|---|---|---|---|
1Phase 1: Business Problem | 15-20 min | State the current business problem in one to two sentences. Who is affected, what happens, how much does it cost? Back up with data or observations. | If the business problem is phrased as “we need more users,” it is too vague. Require a concrete, measurable pain point. |
2Phase 2: Business Outcomes | 15-20 min | Desired business outcomes: measurable changes (revenue, conversion, cost, retention). Maximum 3 outcomes. For each outcome define baseline and target value. | Outcomes without metrics are wishful thinking. For each outcome define a number and source. If a metric is missing, mark it as assumption and plan measurement setup. |
3Phase 3: Users | 20-25 min | Which user segments are affected? For each segment provide concrete description with behavior and context. If personas exist, reference them. | Anyone writing “all users” has not segmented. Lean-UX-Canvas requires narrowing to one to three segments, otherwise User Outcomes are not specific. |
4Phase 4: User Outcomes | 20-25 min | What do users want to achieve (Jobs to be Done)? Which behavior change would produce the business outcome? Connect Users and Business Outcomes. | User Outcomes and business outcomes must be logically linked. If users reach their outcomes, business outcomes should naturally follow. If not, test the model. |
5Phase 5: Solutions | 20-25 min | Collect multiple solution ideas that enable user outcomes. At least 5 options. Use 1-2 sentences per idea, no feature-level detail. | If only one solution is written in Phase 5, Phase 4 was too narrow. Optionality is the value, the experiment decides later. |
6Phase 6: Hypotheses | 15-20 min | Formulate hypotheses for top solutions: “We believe solution X at user Y will produce outcome Z.” Maximum 3 hypotheses for the first test. | Hypotheses without clear subject-solution-outcome structure are not testable. Keep the format strict, otherwise Phase 8 cannot be derived. |
7Phase 7: Critical Assumptions | 20-25 min | Per hypothesis, what assumptions must be true for it to work? For each assumption define risk (what happens if wrong?) and validation need. Mark the riskiest. | Critical assumptions are often missed. Ask what all has to be true for this to work. At least 3 assumptions per hypothesis. |
8Phase 8: Experiments | 15-20 min | First experiment for the riskiest assumption. Method (interview, prototype, fake door, data analysis), sample, success criterion, owner, timeline. Schedule immediately. | An experiment without date and owner is intent, not action. In the workshop, schedule it. If not, the canvas dies in the drawer. |
What comes out at the end
Completed Lean-UX-Canvas with all 8 fields, hypothesis list, assumptions backlog with priority, first experiment as spike with owner and date, linked to research sources.
Date and initiative in the header. New canvas for each iteration with reference to prior version. Track hypothesis status (open, in test, confirmed, disproven). Run experiments in a separate experiment log.
Compact working template for Lean UX Canvas with context, input, output artifacts, and next step.
What is this method used for?
Which question should be answered at the end?
Which data, observations, or materials are available?
Owner, date, success signal.
Concrete filled scenario, fictional example
Concrete filled scenario, fictional example
1 Business Problem: 78% of first-time installers leave the app before step 5 in onboarding, which costs us 14 EUR CAC per user lost.
2 Business Outcomes:
3 Users: Solo freelancers in DACH, revenue 100k-500k EUR, no in-house tax advisor. Devices mainly iPhone and iPad.
4 User Outcomes:
5 Solutions (excerpt):
6 Hypotheses:
7 Critical Assumptions:
8 Experiments: A/B test wizard vs. slideshow for step 3, 50% traffic starting 26.05., evaluation on 09.06. Owner @lisa. Success criterion: step-3 dropout <25% in wizard variant.
Recognize symptoms and steer against them
Team jumps directly to Solutions (field 5) before clarifying problem and outcomes.
Keep phase order strict. Facilitator stops jumps. If solutions arise early, park them and return in Phase 5.
“More engagement” or “better user experience” with no number.
For each outcome define metric, baseline, target. If metric is missing, mark as assumption and include measurement setup as critical assumption.
Solutions field has one idea everyone wants and other options are not collected seriously.
At least 5 Solutions are mandatory. Cross-compare before selecting hypotheses. Anyone allowing only one option has decided too early in Phase 5.
Canvas appears certain, no one knows what was tested and what was guessed.
Use two colors (content, assumption). At least three assumptions per hypothesis. Mark the riskiest as the first experiment.
Phase 8 ends with "we will test soon", no owner, no date.
Schedule in the workshop. Name owner explicitly. Put review date in calendar, not "soon". Otherwise the canvas leaves the workshop and dies.
Canvas is filled once and never updated; learning dissipates.
Update canvas after each experiment evaluation or create new version. Maintain hypothesis status. Quarterly review the full canvas history.
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.