SpyderBot Documentation
Analysis Scope, Coverage, and Comparability
This page answers what scope and coverage mean in SpyderBot, and when two results can be compared safely.
Scope
Scope defines what a result represents. Depending on the product, scope may include prompts, AI systems, search surfaces, market, language, competitors, sources, website, time, filters, and method version.
Always read the scope displayed by the product before interpreting a number, finding, or trend. A result outside its declared scope should not be generalized to the entire AI ecosystem. In Brand Insights, the scope strip on a report view is the boundary for interpreting every metric and finding in that run.
Coverage
Coverage is not one universal metric. Use the qualified term that matches the question:
- prompt-system cell coverage;
- valid-response coverage;
- cell or observation depth;
- AI-system or search-surface coverage;
- time coverage and freshness;
- request or session measurement coverage;
- source or attribution coverage;
- audit target coverage.
A large sample concentrated in one prompt or one source may provide depth without broad coverage. Broad coverage does not guarantee consistency or correctness. Validity, strategic relevance, and evidence quality matter in addition to volume. Unavailable or insufficient states are not the same as a measured zero.
Comparability
Two results are comparable only when the differences do not invalidate the intended interpretation. Review:
- prompt set and prompt versions;
- AI systems, surfaces, and model policies;
- brand and competitor scope;
- market and language;
- source and connection health;
- date range, timezone, and incomplete periods;
- method or taxonomy version;
- eligibility rules and filters;
- material collection gaps or corrections.
If comparability is not supported, establish a new baseline instead of calculating or narrating a trend. Provider updates, Prompt Set changes, method versions, competitor changes, integration changes, source corrections, and incomplete periods may invalidate a direct comparison.