Scene–User–Demand Research

Turn heterogeneous feedback into an auditable evidence chain. Treat collection volume as an input constraint, not as proof of demand.

Choose the operating mode

• Design: create the research contract, scenario universe, sampling matrix, schema, and acceptance gates.
• Execute: implement or run collectors, normalization, deduplication, balanced views, labeling, and audit jobs.
• Audit: inspect an existing corpus and state exactly which claims it can and cannot support.
• Synthesize: produce opportunity theses, counter-evidence, validation queues, and product decisions from an audited corpus.

For an end-to-end study, read references/research-protocol.md. When creating tables or collectors, also read references/data-contract.md. Copy and complete assets/research-contract-template.md at the start of a new study; use assets/opportunity-card-template.md for final synthesis.

Apply the SURE model

Use this analysis key for every important observation:

User role × Scene/trigger × Task/outcome × Current substitute × Friction/cost × Consequence × Evidence level