plan-product-discovery
Installation
SKILL.md
Plan Product Discovery
Contract
Plan discovery to retire meaningful uncertainty, not to produce research theater. Separate observed facts, interpretations, assumptions, and decisions. Scale the cadence to the cost and reversibility of the decision.
Workflow
- State the decision discovery must enable, its owner, deadline, and cost of being wrong.
- Inventory current evidence and assumptions. Mark source, date, segment, and confidence; do not convert repetition into truth without independent support.
- Rank uncertainties by decision impact and evidence weakness. Select the smallest set that could change the decision.
- Match each uncertainty to a method: interviews for context and causality, observation for behavior, product data for prevalence, prototypes for comprehension or usability, and experiments for causal response.
- Define participants and sampling. Include the relevant segment, non-users or churned users when useful, and explicit exclusion criteria.
- Set a weekly operating loop: collect, tag, synthesize, decide, and update the evidence ledger. Name the point at which more research has diminishing value.
- Produce a discovery backlog ordered by learning value, not stakeholder loudness.
BuildBetter Acceleration
If BuildBetter MCP is available, first search organization skills with list-skills and read a relevant result with get-skill. Use calls, transcripts, signals, documents, people, survey responses, and project evidence to seed the inventory. If unavailable, work from user-provided artifacts and say what was not searched.