The essential vocabulary
The few terms you need to talk GEO confidently — and explain it to a client.
You’ll come across these terms in the app and in reports. Here’s what they mean, simply.
Prompt
The question someone asks an AI assistant (“office fit-out in Manchester”). MarquePhare picks the prompts to track based on their estimated demand: we don’t measure what nobody asks.
Citation and mention
A mention is when the AI names a brand in its answer. A citation is when it relies on a site as a source (often with a link). Both matter; the report keeps them apart.
Share of voice
The share of answers that cite your client, compared with competitors, on the same prompts. 40% share of voice: out of 10 answers, your client appears in 4.
Margin of error
AI answers vary from one run to the next. So every percentage comes with its sample size (“n”) and its margin. A change smaller than the margin isn’t progress: it’s noise.
GEO readiness
The site audit’s score out of 100: technical, content, mentions and authority. It tells you what still needs strengthening to get cited — not whether you already are. The measured answers tell you that.
Content score
An article’s score out of 100 in the editor: direct answer, structure, sourced data, FAQ, entities… The higher it is, the more likely AI is to reuse the text.
E-E-A-T
Experience, Expertise, Authoritativeness, Trust: the signals that make an AI judge a source reliable. Named authors, proprietary data and mentions on recognized sites strengthen them.