Methodology

Information is gathered per field, using a defined field catalogue with stable field keys. AI may help collect candidate values. Candidate values are validated against their source before they are stored as facts, and each fact keeps its own evidence, source and confidence.

When information cannot be established, it is shown as Unknown. Unknown is never treated as a negative finding, and it is never replaced with a guess or an inferred value.

Items listed under "Needs attention" are produced by methodology rules applied to validated data. They point to concerns, inconsistencies, missing critical information or material policy gaps. They are not a written verdict about the store.

The score will be shown as a percentage with an accompanying status and confidence. Score weights and thresholds are part of the methodology, not the design.

The scoring methodology is not approved yet, so no score is calculated or displayed during beta. Current methodology version: 0.1-draft.

External reviews are contextual signals, never verified company facts. Cross-platform review aggregation is deferred and is not part of any evaluation yet.