Opportunity Solution Tree (every opportunity quote-backed)

Structure discovery the Teresa Torres way: one measurable outcome at the top, the customer opportunities (needs/pains) beneath it, solutions for the top opportunities, and experiments to validate them. The Evermuse discipline: an opportunity node is only real if a customer said it. Every opportunity carries a verbatim quote; any node you add from intuition is flagged hypothesis — no evidence yet so the tree never launders a guess as a validated need.

Step 0 — Relevance & availability

Confirm this is discovery structuring and Evermuse is connected (see using-evermuse Step 0). If disconnected, you can sketch the tree's skeleton but every opportunity is unverified — label the whole thing ⚠ ungrounded and tell the user to authorize the MCP.

Evermuse Grounding (required)

Follow ${CLAUDE_PLUGIN_ROOT}/skills/using-evermuse/SKILL.md.

Required — one parallel batch of searches. Verify the product (Rule 1), then fire in a single parallel batch:

• 3–4 evidence searches, each worded from a different angle (limit up to 50). Evidence comes back rich and varied — expect large, useful result sets.
• one guidance search and one context search (limit up to 50). These are usually sparse or empty; run them anyway and note when they're thin.

Read each response's digest — it reports how many more results exist. Use judgment on whether a query is worth pulling deeper (raise limit toward the 100 max and/or page with next_offset to avoid repeats), weighing payload size, remaining context, task complexity, and the value of the data. For deep pulls, consider spawning sub-agents — instruct them to return every citation with the same metadata the tools return (url, who_said_it, meeting_name, created_at) so you can still cite.

Optional — considered use. Once grounded, reach for the other tools only when they add value: a quote-angled search for verbatim voice, read_source for a single deep dive, inline citations (references/citations.md), and add_source to save the deliverable. Sources, citations, and saving are optional — not required.

The tree (4 levels)

1. Desired outcome (top) — one measurable metric (e.g. "increase 7-day retention to 40%"). From OKRs/strategy. One outcome per tree.
2. Opportunities — customer needs/pains, framed from their perspective ("I struggle to…", "I wish I could…"). Each must be quote-backed. Prioritize with Opportunity Score = Importance × (1 − Satisfaction), 0–1.
3. Solutions — ≥3 per prioritized opportunity, ideated across PM/Designer/Engineer. Don't commit to the first idea.
4. Experiments — fast, cheap tests per promising solution (hypothesis · method · metric · threshold). Prefer skin-in-the-game over opinions.

Principles: one outcome at a time · opportunities not features · compare/contrast ≥3 solutions · discovery is non-linear (kill what fails, branch anew) · update the tree continuously.

Instructions

1. Define the outcome. Confirm or help articulate a single measurable metric. Refuse to proceed with a vague, unmeasurable one.
2. Map opportunities from evidence. Pull 3–7 real opportunities from the grounded quotes/notes; group related ones; frame each in the customer's words with its quote attached. If you're tempted to add an opportunity with no signal, add it but flag it hypothesis — no evidence yet and note it needs an interview (/evermuse:interview).
3. Prioritize opportunities. Rank by Opportunity Score / demand strength (distinct accounts). Focus on the top 2–3.
4. Generate solutions. ≥3 per prioritized opportunity, across the three perspectives.
5. Design experiments for the most promising solutions (see /evermuse:assumptions for the experiment library).
6. Visualize the tree hierarchically.

Deliverable

🎯 OUTCOME: [measurable metric]
│
├─ OPPORTUNITY: "I struggle to export my data" — 6 accounts, Opp. Score 0.72
│    > "[verbatim]" — [Name], [Meeting], [Date] [`1`](URL)
│    ├─ SOLUTION: Scheduled export  → EXPERIMENT: fake-door · CTR ≥ 8%
│    ├─ SOLUTION: Email-to-share    → EXPERIMENT: prototype task · success ≥ 70%
│    └─ SOLUTION: Public API        → EXPERIMENT: spike · feasible in budget?
│
├─ OPPORTUNITY: [flag] "unified dashboard" — ⚠ hypothesis — no evidence yet
│    → needs an interview before it earns a place on the tree

Further reading

• Adapted from phuryn/pm-skills (MIT) — opportunity-solution-tree. Method: Teresa Torres, Continuous Discovery Habits; Opportunity Score: Dan Olsen, The Lean Product Playbook.