Is this public measurement methodology a certification or rating?
No. It explains measurement conditions, scope, and limitations; it is not a certification, rating, or recommendation guarantee.
Public measurement methodology
How GEO and AI visibility evidence is measured
This public measurement methodology explains how GrayRhino AI presents fixed questions, explicit sources, same-condition evidence, failures, and retest limitations across public pages, product reports, and customer communication.
This page states the public measurement rules, scope, and limitations. It is not a certification, rating, or guarantee.
Eight public measurement rules
These rules constrain public wording and report interpretation. They define how evidence is created, how comparisons are made, and which conclusions must not be overstated.
01
Fixed questions and explicit measurement sources
Each measurement should show the actual questions, model/API surface, measurement source, and time instead of selecting questions after the fact.
02
Brands and competitors use the same method
Brand mentions, competitor mentions, and citation judgments should use the same rules for every entity.
03
Raw answers, citations, and failures are retained
Reports should retain question-level answers, citations, missing states, failures, and unavailable states instead of presenting failures as success.
04
Question-set or measurement-source versions (Prompt Set / Provider Plan) do not create continuous trends
When the question set, measurement-source version, model, or API surface changes, the report should state that the data is not directly continuous instead of stitching it into a false trend.
05
Facts, analytical judgments, and Actions are layered
Original answers and citations are facts; findings are analysis; Actions are follow-up work. The three layers should remain distinguishable.
06
No advice is generated without evidence
When no completed answer exists, or the evidence cannot be linked to a specific question, measurement source, and answer, the system does not generate Actions or advice. A completed answer with no displayable citations is treated only as a checkable “missing verifiable sources” fact and links back to that answer.
07
Retests describe change, not causality
Retests should describe observed before/after changes under the same method, not claim that an Action necessarily caused an AI answer to change.
08
Plans and customer relationships do not change measurements
Payment status, plan tier, or customer relationship should not change Provider results, mention judgments, or evidence presentation.
Scope and limitations
- The current scope is Mainland China and Simplified Chinese only.
- New projects use 4 base questions; confirming a competitor adds 1 competitor-comparison question.
- Current measurement sources are only Volcengine Ark API web search and Baidu Qianfan Intelligent Search Generation API.
- This page is a public measurement methodology, not consumer ranking certification, authority rating, or recommendation guarantee.
- Current product screens, reports, and help pages are the reference for available functionality.
Related pages
GEO visibility monitoring
Review product scope, report contents, continuous monitoring value, and plan matrix.
Home
See GrayRhino AI positioning and current capabilities.
Help
Read about the app, reports, Actions, retests, and plans.
Experiments
Review current API measurement sources and consumer-side boundaries.
Frequently asked questions
Why use fixed questions and measurement sources?
Fixed questions and explicit measurement sources reduce methodology drift, so brand, competitor, citation, and missing-evidence changes can be reviewed under the same method.
Can retests prove optimization causality?
No. Retests only describe observed changes under the same method; they do not claim that an Action necessarily caused an AI answer to change.
Can a plan change measurement results?
No. Plans and customer relationships do not change measurement results, mention judgments, or evidence presentation.