Skip to content

borghei/Claude-Skills / project-management/discovery/opportunity-solution-tree

opportunity-solution-tree

Verified

Maps one outcome to customer opportunities, 3+ solutions each and assumption tests, then audits the tree. Use when prioritizing discovery work, mapping solutions to a problem, or checking whether a roadmap moves outcomes.

Opportunity Solution Tree (OST)

Teresa Torres' framework from Continuous Discovery Habits. An OST visualizes the path from a desired outcome to the assumption tests that will validate or invalidate candidate solutions.

When to use this skill

  • Prioritizing discovery work for a quarter
  • Structuring weekly customer touchpoints
  • Mapping multiple solutions to one problem (vs jumping to solution)
  • Auditing whether roadmap actually moves outcomes
  • Coaching a team into continuous discovery rhythm
  • Pivoting discovery away from a dead-end branch

The tree structure

                  [Outcome]
                      |
        +-------------+-------------+
        |             |             |
   Opportunity   Opportunity   Opportunity
        |             |             |
     +--+--+      +--+--+      +--+--+
     |     |      |     |      |     |
   Solution Solution ...
        |
   +----+----+
   |         |
 Assumption  Assumption
   Test        Test

Levels

  1. Outcome — a single, specific, measurable business / product outcome
  2. Opportunities — customer needs/pains/desires that, if addressed, drive the outcome
  3. Solutions — candidate ways to address each opportunity
  4. Assumption tests — experiments validating that the solution will deliver

Clarify First

Before building the tree, settle the inputs below. When one is missing or vague and would change the result, ask about it; when it wouldn't, use a sensible default and say which one you used.

  • The one outcome: a single, measurable, bounded outcome (the tree root; "ship X" or "make users happy" produces an invalid tree).
  • Customer evidence source: interviews / tickets / analytics that populate the opportunity layer (opportunities must come from research, not the team's imagination).
  • Engagement type: net-new tree vs auditing an existing roadmap (net-new builds top-down; an audit maps current solutions back onto outcomes).

Ask about the two or three inputs that matter most for this request, not the whole list. If the user wants a result straight away, produce it and state your assumptions at the top.

Workflow

Layer definitions, good and bad examples, the weekly rhythm and stuck-tree symptoms are in references/ost-fundamentals.md; this section is the procedure.

  1. One outcome. Measurable, behavioral or business, bounded to a quarter or half, and within the team's influence ("Lift week-1 activation from 28% to 40% by end of Q3"). "Ship X" is an output; "hit $20M ARR" is too high for one team.
  2. Opportunities from evidence. Pull them from interviews, tickets, sales objections, analytics and survey text, phrased as the customer's need ("Admins want to invite their whole team at once"), never as a build ("Add CSV import"). Cluster to 3-7 per outcome.
  3. Three or more solutions per opportunity. A single solution means the team is defending a favourite; ask "what if we couldn't build that?" and look at analogous problems.
  4. Assumption tests. For each solution, name the riskiest of value, usability, feasibility and viability, and the cheapest test with a time box and a pass criterion. Climb the ladder only as needed: interview, landing page or smoke test, Wizard-of-Oz or concierge, low-fi prototype, high-fi prototype, production A/B. AI-built prototypes make the prototype rungs cheap, not the earlier rungs unnecessary.
  5. Audit the tree with ost_validator.py (below), and fix every fail.
  6. Iterate weekly. Add opportunities from new interviews, record test results, kill failed solutions, promote validated ones to the roadmap.

To audit an existing roadmap, map each roadmap item to an opportunity and the outcome; items with no opportunity are orphans, and opportunities with no item are gaps.

Tool

ost_validator.py reads a JSON tree (schema in its --help; JSON is the default output, --format markdown for a table, --output to write a file).

python3 project-management/discovery/opportunity-solution-tree/scripts/ost_validator.py \
  --input project-management/discovery/opportunity-solution-tree/assets/sample_ost.json --format markdown         # exit 0
python3 project-management/discovery/opportunity-solution-tree/scripts/ost_validator.py \
  --input project-management/discovery/opportunity-solution-tree/assets/sample_ost_flawed.json --format markdown  # exit 1

It fails a tree with no outcome or opportunities, an outcome not marked measurable or behavioral, or no update in 30+ days (measured from as_of, so a fixed file gives a fixed answer). It warns on output language in the outcome, opportunities that start with a build verb (add, build, create, implement and similar, checked against the text whatever the is_customer_phrased flag says), missing evidence, fewer than 3 solutions, solutions without tests or pass criteria, and no update in 14+ days. Exit codes: 0 no fail, 1 at least one fail, 2 bad input.

Anti-Patterns

Outcome that is an output

Mistake: The root reads "Ship the new onboarding". Why it happens: Roadmaps are written in features, and the tree gets copied from the roadmap. Instead: Name the behavior or business number that the onboarding should move, with a baseline and a target. The validator fails an outcome not marked measurable and behavioral.

Solutions in the opportunity layer

Mistake: Opportunities read "Add bulk CSV import", "Build SSO". Why it happens: Teams think in solutions, and a solution feels more concrete than a need. Instead: Ask "why would a customer want that?" and write the answer as the opportunity; the import becomes one of three solutions under it.

The static tree

Mistake: The tree was built in a planning offsite and not touched since. Why it happens: Building it felt like the deliverable. Instead: Update it weekly from new interviews and test results; a tree older than 30 days fails the audit.

More failures and fixes: references/ost-anti-patterns.md.

Scope & Limitations

In Scope:

  • Building an Opportunity Solution Tree (Teresa Torres) from one outcome through opportunities and solutions to assumption tests
  • The opportunity-versus-solution test, opportunity sizing and the rule of 3+ solutions per opportunity
  • Value, usability, feasibility and viability assumptions per solution, with an assumption test ladder from interview to production A/B test
  • A tree audit for outcome quality, opportunities written as solutions, single-solution branches, missing tests and stale trees (ost_validator.py)
  • Mapping an existing roadmap onto the tree to find orphan solutions and gaps

Out of Scope:

  • Running interviews or synthesizing them into opportunities (see project-management/discovery/customer-interview-script/ and project-management/discovery/interview-synthesis/)
  • Designing the experiments in detail (see project-management/discovery/brainstorm-experiments/)
  • Defining the company-level North Star (see project-management/execution/north-star-metric/)
  • Building prototypes for assumption tests (see project-management/discovery/ai-prototyping/)

Important Caveats:

  • ost_validator.py reads a hand-built JSON tree. Flags such as is_customer_phrased and has_evidence are self-reported, so the tool checks structure, not the quality of the evidence.
  • Opportunity scoring (impact, frequency, strategic fit) is an estimate the team makes. The tool does not compute it from data.
  • The tree goes stale without weekly updates. The validator warns when last_updated is more than 14 days before as_of and fails it past 30.

References

  • references/ost-fundamentals.md — layers, outcome/opportunity/solution examples, assumption tests, weekly rhythm, stuck trees
  • references/ost-anti-patterns.md — common failures + fixes

Related skills

  • project-management/discovery/identify-assumptions — assumption surfacing
  • project-management/discovery/brainstorm-experiments — test design
  • project-management/discovery/customer-interview-script — interview prep
  • project-management/discovery/interview-synthesis — turn interviews into opportunities
  • project-management/execution/north-star-metric — outcome definition
  • project-management/strategy-frameworks/lean-canvas — strategic context
  • product-team/research-summarizer — interview synthesis

Files

FilePurpose
scripts/ost_validator.pyAudits a JSON tree for outcome quality, opportunity phrasing, solution diversity, test coverage and freshness
assets/sample_ost.json, assets/sample_ost_flawed.jsonRunnable trees: sound (exit 0) and flawed (exit 1)
assets/ost_template.mdTree worksheet to fill in
references/ost-fundamentals.md, references/ost-anti-patterns.mdFramework detail; failures and fixes
rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit.Stars: 468kUpdated Claude CodeCodexVerifiedInstall
rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure.Stars: 468kUpdated Claude CodeCodexVerifiedInstall
rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics.Stars: 468kUpdated Claude CodeCodexVerifiedInstall
ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first.Stars: 194kUpdated Claude CodeCodexVerifiedInstall