The 3RD glossary: 15 metrics behind your dashboard

Most GEO glossaries explain the industry's shared vocabulary: GEO, AEO, citations, hallucinations. This one is different. These are the 15 metrics and terms you'll actually run into inside 3RD, translated into plain language with a real-world example for each.

Visibility Score

The core measure of how often a brand appears across its tracked prompts, shown as an overall figure and broken down by prompt type. A snack brand tracking 90 prompts might have an overall Visibility Score of 42, but see that score drop sharply on its must-win prompts specifically, revealing that the brand shows up reasonably often in general conversation but disappears on the exact questions its category leaders compete hardest on.

Prompt types: must-win, core and foundation

The three tiers 3RD uses to classify tracked prompts by strategic weight: must-win prompts are the ones a brand cannot afford to lose (direct category leadership questions), core prompts are everyday buyer questions, and foundation prompts are broader awareness-level queries. A bank might score well on foundation prompts like "what is a savings account" while scoring poorly on the must-win prompt "best savings account Denmark," which is the gap that actually matters commercially.

Opportunity Radar

A five-axis view of a brand's position: Visibility, Ranking, Sentiment, Citations and Crawlability, plotted together so strengths and weaknesses show up side by side rather than buried in separate reports. A retailer might score strongly on Sentiment, meaning AI models speak well of it when it comes up, but weakly on Crawlability, meaning technical barriers are quietly stopping AI crawlers from indexing its site in the first place, a fix that's cheap relative to the visibility it unlocks.

Competitor ranking

A ranked leaderboard, shown on the dashboard as the Competitor rank card, placing a brand against its named competitors by visibility score over a chosen period and showing exactly where it sits and by how much. An insurance brand might discover it ranks third out of five tracked competitors, comfortably ahead of two smaller players but meaningfully behind the category's top two, which reframes "we have a visibility problem" into "we have a specific gap to two specific competitors."

Attribution share

The share of a brand's citations that come from its own website versus third-party sources, broken down per AI provider and shown on the dashboard as the Attribution by provider card. A furniture brand might find that 60 percent of its ChatGPT citations point to its own product pages, while on Perplexity that figure drops to 20 percent, with the rest going to retailer listings and review sites, meaning the two models are effectively building their answers from different parts of the web.

Prompt coverage

The share of a brand's tracked key prompts where it gets mentioned at all, shown on the dashboard as the Key prompts coverage card, alongside the specific prompts it wins and the ones it's absent from. A software company might have strong overall coverage but discover its three highest-intent prompts, the ones closest to a purchase decision, are exactly the ones where it's never mentioned, which is a more useful finding than the aggregate number alone.

Sentiment distribution

The positive, neutral and negative share of how AI models talk about a brand when they do mention it, plus an overall sentiment score. A bank might show 70 percent positive sentiment overall, but on closer inspection find the negative share is concentrated almost entirely around one recurring complaint about a specific fee, a narrow, fixable issue rather than a broad reputation problem.

Citation sources (Citations Top Sources)

The detailed, per-source list of exactly which pages cite a brand, shown on the dashboard as the Citations Top Sources view, with each source tagged as owned, earned or paid and given a citation count and share. A skincare brand might see that a single ingredient-explainer article on a third-party beauty site accounts for a disproportionate share of its total citations, making that one page worth actively maintaining and worth understanding.

Source category

A classification of every citing source into types such as Editorial, Blog, Social/UGC, Commercial, Institutional or Aggregator, making it possible to see the shape of a brand's citation mix at a glance. A supplement brand might notice its citations skew heavily toward Social/UGC sources like forum threads, and barely register on Editorial sources, a pattern that suggests where outreach effort would do the most good.

Outreach opportunities

Third-party sources that get cited often for a brand's relevant topics, where the brand itself isn't mentioned yet, grouped by source with when they were first and last seen. A travel brand might find a specific "best hotels in Copenhagen" listicle gets cited repeatedly by AI models answering exactly the questions it cares about, without the brand appearing anywhere in it, turning a vague "we should do outreach" into a specific, addressable target.

Topic mentions

Visibility broken down by topic rather than by individual prompt, shown on the dashboard as the Topics leaderboard, making clear which subject areas a brand owns and which it doesn't. A snack brand might score strongly on the topic of recipe ideas but weakly on healthy snacking, even though both topics get similar volumes of relevant prompts, pointing to a content gap that's specific rather than general.

Content gaps

A per-prompt view, matching the dashboard's Content Gaps page, of whether a brand's own pages actually cover a given prompt, with a coverage percentage, a count of matching pages, and an opportunity score for closing the gap. An electronics brand might see a high-value prompt with a strong opportunity score but zero matching pages on its own site, meaning no one has to guess what content to write next.

Visibility trend

The brand's visibility score plotted daily over a chosen window, shown on the dashboard as the Visibility development chart, making movement and inflection points visible rather than just a single snapshot number. A retailer might spot a sustained dip in its trend line starting the same week a competitor launched a major campaign, connecting a metric change to a specific, real-world cause instead of treating it as unexplained noise.

Readability and technical health

An average readability, on-page SEO and Core Web Vitals score across a brand's own analyzed pages, along with which specific pages perform best and worst. A B2B software brand might find its strongest piece of content, by far the most thorough on its topic, also has the worst Core Web Vitals score in its entire library, a plausible reason a page that should be winning citations isn't.

Playbook

The set of recommendations and tasks generated for a brand, translating what the other metrics show into a concrete to-do list rather than leaving the analysis to speak for itself. A bank sitting on a large gap between its foundation and must-win prompt scores might see a playbook task suggesting comparison content for a specific high-value prompt, turning a diagnosis into a next action rather than another number to interpret.

Using this list

These 15 terms map onto the actual cards and views inside the 3RD dashboard, so the next time a number moves, you'll know exactly what it's measuring and where to look for the reason behind it.