How to read your scan evidence
Pillar diagnostics explain the evidence behind a scan. Coverage describes what we could verify; it does not promise rankings, AI citations or business outcomes.
Overlapping pillars, unchanged headline
Pillars group related checks and can overlap: one check may inform more than one pillar. Do not add pillar scores together. The stored headline score, grade and existing scoring weights remain unchanged; pillar diagnostics are a separate interpretation.
Weighted evidence coverage
Coverage is verified observed weight divided by applicable weight. Explicitly not-applicable checks are excluded from both numerator and denominator. Missing, failed, unknown or unverifiable collection stays in the denominator and adds no observed coverage. Ambiguous partial assessment does not count as fully verified. If applicable weight is zero, evidence is insufficient, never a perfect score.
Coverage gates before rounding
Coverage below 0.67 suppresses the numeric pillar score and shows insufficient evidence. From 0.67 to below 0.85 the score is provisional. At least 0.85 is scored. These decisions use unrounded weighted coverage, even if a displayed percentage rounds up to a boundary. A scored pillar is not a guarantee of complete collection.
Method versions and historical results
Newly calculated diagnostics record their methodology version. Stored legacy snapshots retain their original version and are not rewritten. Historical diagnostics recalculated using a newer method are labelled as recalculated; old scans without collection evidence cannot be treated as verified observations. Compare method versions before interpreting differences.
Collection and comparison limits
Provider outages, blocked requests and missing collection can reduce coverage. Evidence displayed in results is sanitized and subject to access controls. Current redacted, origin-only evidence cannot substantiate an improvement comparison: it does not establish comparable before-and-after measurements. Estimated impact remains an estimate.