Skip to main content

​
Direct answer

Filter to prompts assigned to the awareness or TOFU stage. Use Coverage to see how broadly your brand appears across answered prompts, Mention Rate to assess eligible answers, and answer text to decide whether the mention is relevant.

​
Data required

  • Answered prompts consistently assigned to the awareness or TOFU stage
  • SELF and DIRECT competitor classifications
  • A fixed provider, model, country, language, topic, tag, and period scope
  • Coverage and Mention Rate raw components
  • Representative answers with and without SELF mentions

​
Workflow

  1. In Workspace > Prompts, verify that awareness prompts are category-neutral and informational.
  2. In Cockpit > GEO Matrix, apply the awareness or TOFU filter and record Coverage and Mention Rate.
  3. Compare topics and providers one at a time while preserving the rest of the scope.
  4. In Monitor > Prompt Tracking, identify prompts driving presence or absence.
  5. Read the matching answers in Monitor > Response Analysis to check whether the brand belongs in the context.
  6. Compare with consideration-stage results only after matching prompt volume and answer coverage.

​
Interpretation

Product Coverage is prompt-level: distinct answered prompts with at least one SELF mention divided by distinct answered prompts. Product Mention Rate is response-level: answers with a SELF mention divided by answers with at least one SELF or DIRECT brand mention. Low brand presence can be expected for broad educational prompts where naming vendors is not relevant. High presence can also be incidental. Use prompt intent and answer context before calling either result strong or weak.

​
Possible next actions

  • Test educational content that answers a recurring category question with verifiable evidence.
  • Remove or rewrite an awareness prompt that does not match the category decision you intend to study.
  • Investigate a topic-specific absence before expanding the prompt set.
  • Track a fixed awareness cohort after the experiment.
Treat each expected metric change as a hypothesis.

​
Limitations

  • Awareness-stage labels depend on your prompt taxonomy.
  • The monitored prompt set does not measure all discovery behavior.
  • Presence does not imply trust, citation, or future purchase.
  • Provider output can vary between runs.