Metrics & the Visibility Score

What’s being measured

Every metric below is measured over a set of runs — one prompt executed on one platform (see Prompts, runs & answers). That set can be your whole project’s history, or narrowed to one platform, one prompt category, or one day — the way each number is calculated doesn’t change, only the runs it’s measured over do.

A run counts toward the total even when the platform never answered — a Google search where no AI Overview appeared still counts as a run “in scope,” just one where you weren’t mentioned or cited. That’s a deliberate choice: a platform going silent on a question is a real visibility outcome, not a data gap to be quietly excluded (see Answer engines we track).

Mention Rate

The simplest metric, on a 0–100% scale: of every run in scope, what percentage mentioned your brand anywhere in the text? This is pure text-level presence — it doesn’t care whether you were also linked as a source.

Citation Rate

The same idea, one level down the funnel and also on a 0–100% scale: of every run in scope, what percentage linked to one of your pages as a source? This is link-level presence, and it’s tracked completely separately from Mention Rate — see why the two differ.

Position & prominence

Being mentioned is good; being mentioned first is better. Every mention is recorded with a position — whether your brand was the first name in the answer, the second, and so on. From that, your dashboard shows a Position Score on a 0–100 scale that combines two things into one number: how early you tend to be named, and how often you’re absent altogether.

Read it directionally: being named early and consistently drives the score toward 100, while being absent most of the time — or usually named last — pulls it toward 0. Folding absence into the same number is what makes it more honest than a raw “average position,” which would have to quietly ignore every run where you didn’t appear at all.

Sentiment Index

Each run where you’re mentioned contributes one sentiment judgment — positive, neutral, or negative (see how sentiment is judged). The Sentiment Index rolls those up onto a 0–100 scale that centers on 50 for neutral: consistently positive coverage pushes it toward 100, consistently negative coverage toward 0. Because it’s scored once per run rather than once per individual name-drop, a brand mentioned three times in one glowing answer doesn’t get an outsized boost over a brand mentioned once in three separate glowing answers.

Visibility Score

The headline number, on a 0–100 scale. It’s a single blend of the four signals above — whether you’re mentioned, whether you’re cited, how prominently you’re placed, and the tone you’re described with. The most weight sits on simply being present — named and cited at all — with prominence and tone as smaller refinements on top. In practice you can treat it as your one-glance answer to “how visible am I in AI answers right now?”

Every brand you track — yours and each competitor — gets a Visibility Score computed on the exact same rules. That’s what makes the number comparable across your competitive set instead of just being a self-referential index.

Share of Voice

Where Visibility Score tells you how you’re doing in isolation, Share of Voice tells you how you’re doing relative to the field. On a 0–100% scale, it’s your share of all the brand mentions across you and every tracked competitor combined — the number that answers “of all the attention AI answers give to brands like mine, how much of it is mine?” There is a citation-based version too, measuring your share of all the citations rather than mentions.

Reading the score over time

A single day’s answers can be noisy — the same prompt asked twice won’t always come back identical. OptimizeCamp’s alerting smooths this out by comparing today’s score against a trailing 7-day average rather than against yesterday alone, so a real, sustained drop stands out from ordinary day-to-day variance before it gets flagged (see Alerts).

One more thing worth knowing: editing a tracked prompt’s wording, or changing your brand profile’s aliases, changes what actually gets measured going forward. That’s a methodology change, not a real shift in AI visibility — keep it in mind if a metric jumps right after you make one of those edits.

Every competitor, same rules

Nothing about these metrics is brand-specific — the exact same calculation runs for your brand and for every competitor in your brand profile. That’s what makes head-to-head comparisons on the Competitors screen meaningful: every score in that table was produced by the identical rule set, just pointed at a different brand.