Editorial starter: a measurement framework to adapt to your own research. No platform results are asserted here.
Start with a question set
Choose questions that reflect the tasks your audience is trying to complete. Separate informational, comparison, and recommendation questions. Preserve the exact wording so future observations remain comparable.
Distinguish mentions from citations
A brand mention and a linked citation answer different questions. Log whether the brand is named, whether your domain is linked, and whether the answer recommends a product. Do not combine these into a single score without explaining the weighting.
Build an observation sheet
Record the prompt, platform, observation date, brand mention, linked domains, and any uncertainty. Save enough context to let someone else review the classification.
A simple citation rate
Divide observations with a citation to your domain by the number of eligible observations. Always report both the numerator and denominator. A result based on five prompts should not look as reliable as a result based on five hundred.
Interpret changes carefully
Repeat prompts over a defined window. Changing the prompt set or platform mix can change the aggregate even when underlying behavior has not changed. Compare like with like and preserve the original logs.
Next steps
Pilot the framework on a small set of questions, review ambiguous cases, then freeze the definitions before expanding the study.