Citation tracking
AI visibility must be measured directly, by sampling the real engines with a fixed intent set and recording brand mentions and citations, for two reasons. Referral analytics see only a sliver of AI engagement: Cloudflare measured OpenAI's crawlers fetching about 1,091 pages per referred visitor and Anthropic's about 38,066 in July 2025. And model outputs are non-deterministic: Anthropic states that even at temperature 0 results are not fully deterministic, and OpenAI states outputs may differ from request to request, with a fixed seed giving only mostly consistent results. Repeated sampling with confidence intervals is therefore required.
Sources
- 1Cloudflare · 2025The crawl-to-click gap: Cloudflare data on AI bots, training, and referralsblog.cloudflare.com
- 2
- 3OpenAI API · 2026Advanced usage: reproducible outputs (seed, system_fingerprint)developers.openai.com
CGrade C: reputable industry data or observation. A grade describes the provenance of the sources above; it is not a probability that a change will work. The effect basis says how the size of the effect is known: measured by a study, taken from an industry report, only documented without a size, or not quantified.
Put this evidence to work
Related entries
More in Measurement, or search the whole base.
All Measurement entries