What needs to be finished first
A clearly formulated outcome or North Star Metric with target value exists that the tree can anchor to.
At least 5-10 fresh user interviews from the target segment have been conducted from which opportunities can be derived.
What needs to be ready before start
Visual board (Miro, FigJam, Mural) with hierarchical tree structure; sticky note colors for outcome (top), opportunities (middle), solutions (bottom), experiments (floor); link to interview data repository.
Product trio (Product Manager, Designer, Engineering Lead) as core; optional researcher; one owner for the tree, updated across sessions.
Current outcome with target value and baseline; raw interview notes or highlights from recent interviews; existing solutions/features in backlog; learning status from ongoing experiments.
1-2 h initial, then 30-60 min weekly
Set outcome at the top of the tree and make it bold. Prepare three to five opportunity lanes. Define sticky-note convention. Set tree owner and review cadence (for example Tuesday weekly).
The one question this method answers
Which opportunity should the team address next with which solution, and which experiment provides the next learning signal?
Marker: Phase
| Step | Duration | Action | Hint |
|---|---|---|---|
1Phase 1: Outcome anchor | 10 min | Place the outcome at top of tree as a clear measurable statement with baseline and target. Make time window explicit (for example, next 12 weeks). | Outcome must be a user or business metric, not an output ("launch feature X"). A wrong outcome creates wrong opportunities. |
2Phase 2: Opportunities from interview data | 30-45 min | Distill customer needs, pain points, and desires from interview highlights. Phrase each opportunity as a user-view sentence ("I want ... so that ..."). De-duplicate and cluster into 3-5 lanes. | Opportunities are not solutions. "Faster checkout" is solution language, "I lose trust when checkout takes too long" is opportunity language. |
3Phase 3: Pick one opportunity | 10 min | Team selects ONE opportunity to deepen. Selection criteria: outcome impact, number of mentions in interviews, addressability. | Focusing on more than one at a time fragments focus. The tree grows over time, not in one session. |
4Phase 4: Diverge solutions | 20-30 min | For each selected opportunity, collect 3-5 solutions. First divergent ideation (including outlandish ideas), then convergence. Keep solutions short as actions ("inline validation," "hide optional fields"). | If only one solution appears, team bias is too strong. Force at least three solutions, otherwise the best option is missed. |
5Phase 5: Experiments under solutions | 20-30 min | Define at least one experiment per solution that tests the riskiest assumption. For each experiment: hypothesis, method, success criterion, estimated effort, owner. | An experiment is not "build the solution and watch." It is a targeted assumption test (prototype test, fake door, interview). Otherwise it is delivery. |
What comes out at the end
Collaborative whiteboard living tree diagram with outcome, opportunities, solutions, experiments, and status markers (open, in progress, learned, discarded), plus a learning log text list with date and result per experiment.
Tree is a living document; take monthly snapshots as photo export for history. Mark discarded opportunities and solutions as "discarded because ...", rather than deleting them. This prevents repeated debates.
Structure for outcome, opportunities, solutions, and experiments.
Outcome
Opportunities
Solutions
Experiments
Concrete filled scenario, fictional example
Concrete filled scenario, fictional example
Outcome: Activation rate for session 1 from 22% to 38% by 30.09.2026.
Recognize symptoms and steer against them
Opportunity layer contains feature proposals ("improve dashboard") instead of user needs.
Convert by asking: "Why does the user need this feature?" Response belongs in opportunity, feature moves into solution layer.
Outcome is "launch feature X" or "deliver roadmap item Y."
Shift outcome to user or business metric. What changes for user when output is delivered? That is the real outcome.
Team works on 5 opportunities at once, each with 2 experiments.
Focus on one opportunity per iteration. Depth in solutions beats breadth in opportunities. Limit active experiments to 2.
Tree is set up once per quarter and then ignored, no updates from experiments.
Run weekly refinement sessions with trio. Tree updates must be required content of each review, otherwise it becomes irrelevant.
Roadmap items are listed under solutions and then executed and "tracked."
Experiment definition must be strict: hypothesis, test method, and success criterion before build. If the result is irrelevant to build decisions, it is not an experiment.
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.