Count a target mention when the answer text names the target, product, or a defined alias. Then classify prominence: lead recommendation, shortlist, incidental mention, or absence. A link is not required for a mention.
Citation
The answer supplies a source
Count a citation when the grounded response returns a source URL or explicit attribution. Classify whether it belongs to the claimed domain or a third party. The source may support a category claim without naming the target brand.
The four possible outcomes
Mention and citation are independent answer events
Mentioned + owned domain cited
The brand enters the answer and its site appears in the source set. Inspect whether the cited page supports the specific claim.
Mentioned + no owned citation
The brand is visible, but the answer may rely on third-party corroboration, model knowledge, or another source. This is still a valid mention.
Not mentioned + owned domain cited
The site contributes evidence to the answer without winning brand inclusion. The cited page may be useful but weakly connected to the entity or buying decision.
Neither mentioned nor cited
The question produced no direct target visibility. Review which competitors and sources appeared before choosing an action.
How to score the signals
Preserve the event before calculating the rate
Discovery mention rate
Neutral eligible answers naming the target ÷ neutral eligible answers.
Prominence rate
Lead or shortlist target mentions ÷ answers that mention the target.
Claimed-domain citation rate
Eligible answers citing the submitted domain ÷ eligible answers.
Third-party source frequency
Answer-event frequency plus the number of distinct buyer questions reached by each domain.
Worked example
One answer, three separate observations
“For a small marketing team, consider Acme for simple audits and Beacon for continuous monitoring.” Sources: an independent category roundup and beacon.example/pricing.
Acme: shortlist mention, no owned citation
Beacon: shortlist mention, owned-domain citation
Category roundup: third-party citation, not a brand mention
Illustrative answer only. Real tests must preserve the provider, model, timestamp, question, full response, and returned source URLs.
Act on the observed gap
A mention problem and a citation problem need different work
Mention gap
Clarify category and audience language, answer the exact buyer question, connect the page to the brand entity, and earn relevant third-party corroboration.
Prominence gap
Strengthen the proof, comparison detail, use-case specificity, and independent authority needed to move from incidental inclusion to a shortlist.
Owned citation gap
Publish an original, directly relevant source page with extractable facts, clear authorship, supporting evidence, and stable crawlable URLs.
Third-party citation gap
Identify the domains repeatedly used for the buyer question, then pursue legitimate coverage, data contributions, listings, or expert references where editorially appropriate.
These actions support both AEO and GEO: answer-ready pages help selection, while corroborating sources help generative engines verify the entity and its claims.
See both signals together
Measure mentions, prominence, citations, and competitors across four providers.
The $9 first audit preserves every answer and source behind the result.
A brand mention occurs when the answer names the target brand, product, or accepted alias. The answer may mention a brand without linking to or citing its website, so mention rate and citation rate must be scored separately.
What is a ChatGPT citation?
A citation is a source link or attribution supplied with the answer. For brand measurement, distinguish the claimed domain from third-party domains. A citation can support the answer without naming the target brand, and a mention can appear without an owned-domain citation.
Are ChatGPT mentions or citations more important?
They answer different questions. Mentions show brand inclusion in the response; citations show which sources supported the grounded answer. Commercial discovery may prioritize mentions and prominence, while content authority and referral opportunity may prioritize citations.
How do you measure ChatGPT brand mentions and citations?
Freeze neutral buyer questions, use consistent web-search conditions, preserve each answer and source, and label target mentions, prominence, selected competitors, and source ownership. Report eligible-answer denominators and failures explicitly.
Apply this resource
Put chatgpt citations vs brand mentions into a measured workflow
Preserve the questions, conditions, answers, citations, and failures so the next decision rests on inspectable evidence.