Skip to content

Product Discovery

  1. Identify the user, current problem and desired change in behavior. Reuse existing research and constraints before asking for more.
  2. Separate known facts, assumptions and open questions. Choose the uncertainty that would most change the decision.
  3. Define a small experiment or end-to-end slice with an observable success signal. Include a failure or stop condition where relevant.
  4. Describe the primary flow, edge cases and necessary interface/API contract. Prefer the existing stack and one complete slice over speculative infrastructure.
  5. Add acceptance criteria and tasks to the current work record. Link designs or research at their existing location rather than duplicating them.
  6. Collect user feedback, preserve its wording and record what changed. Verify the outcome before broadening scope.

Output

Produce a concise problem statement, success signal, scope, acceptance scenarios and next action. Add options and tradeoffs only when a real decision needs them.

Guardrails

  • Do not treat a brainstorm as evidence of demand or invent interview findings and metrics.
  • Do not add a new roadmap, proposal and task spreadsheet when one work record will do.
  • Ask only for missing information that changes a consequential decision; make routine reversible choices independently.

Canonical source: skills/product-discovery/SKILL.md.