Metrics Definition
Comprehensive guide to understanding each metric, how they're calculated, and why they matter for your brand's AI search performance.
Core Metrics
Visibility (called Mentioned in the new dashboard) measures whether your brand appears in AI-generated responses. It is a binary metric (0 or 1) indicating presence or absence of your brand in a given response.
Visibility = 1 if your brand is mentioned in the AI response, 0 otherwise. Visibility percentage = (Number of responses where brand appears / Total responses) × 100
Visibility is the foundation of brand awareness in AI search. When your brand appears in AI responses, you gain exposure to potential customers at the moment they are seeking information. High visibility means your brand is part of the conversation when users are making decisions.
Increasing visibility directly correlates with brand awareness and consideration. Brands with higher visibility are more likely to be remembered, researched further, and chosen by consumers. It is the first step in the customer journey within AI-powered search environments.
Share of Voice (SOV) is your share of the tracked-brand conversation, averaged over every answer. An answer that doesn't name you counts as 0, so Share of Voice never exceeds Visibility.
In each answer, Share of Voice = your mentions / the mentions of every brand you track (your brand and your competitors) in that answer. Share of Voice % = (Sum of that share over every answer in the selected period, counting answers that don't name you as 0) / Total answers × 100. In the Share of Voice chart, the grey band is the rest of each day: answers that name none of the plotted brands.
Share of Voice indicates your competitive position in the market. A higher SOV means your brand is mentioned more frequently than competitors, suggesting stronger market presence and mindshare. It helps you understand if you are leading the conversation or being overshadowed by competitors.
Brands with higher share of voice typically enjoy greater brand recall, increased consideration, and higher conversion rates. Monitoring SOV helps you track competitive dynamics and measure the effectiveness of your content and marketing strategies.
Sentiment measures the emotional tone and perception of your brand in AI responses. It is scored from 0-100, where higher scores indicate more positive sentiment.
Sentiment Score = AI-analyzed sentiment score (0-100) based on context, language, and tone of brand mentions. Average Sentiment = Sum of all sentiment scores / Number of responses where brand appears.
Sentiment reflects how your brand is perceived by AI models and, by extension, how information about your brand is being presented to users. Positive sentiment builds trust and credibility, while negative sentiment can damage reputation and deter potential customers.
Positive sentiment in AI responses can significantly influence user perception and decision-making. It acts as a form of social proof, with AI models presenting your brand favorably. Monitoring sentiment helps you identify reputation issues early and measure the effectiveness of PR and content strategies.
Position is the average rank of your first mention within an answer, when you are mentioned (#1 = named first). Lower is better.
Position = your brand's rank among the brands named in the answer (1 = named first, 2 = named second, etc.). Average Position = Sum of positions / Number of answers where your brand appears.
Position matters because earlier mentions in AI responses receive more attention and are more likely to influence user decisions. Being mentioned first or early suggests your brand is the primary or most relevant option for the query.
Brands mentioned earlier in AI responses benefit from primacy effect—users are more likely to remember and consider them. Improving position can lead to increased brand recall, higher click-through rates, and better conversion outcomes.
Top Sources identifies the domains and URLs that are most frequently cited by AI models when generating responses. These sources represent the content that AI models trust and reference most often.
Top Sources are ranked by how many times each domain appears in the selected period: as a cited (linked) source with "Cited Sources Only", or as a retrieved source with "All Sources". Sources the pipeline has not classified yet are left out.
Understanding which sources AI models reference helps you identify authoritative content in your industry. If your content appears in top sources, it indicates high authority and trustworthiness. If competitors dominate top sources, it reveals content gaps and opportunities.
Having your content in top sources increases your brand's visibility and credibility. It positions your brand as a trusted authority and can drive organic traffic. Analyzing top sources helps you identify partnership opportunities and content strategies to improve your authority.
Domain Types categorize sources by their nature and origin: Corporate, Competitor, Editorial, UGC (User-Generated Content), Reference, Institutional, and Other. Your own domains keep their type and are highlighted in green.
Domain Types come from the pipeline's classification of each source: Corporate (company websites), Editorial (news/media), UGC (forums/social), Reference (encyclopedias), Institutional (educational/government) and Other. Competitor = a domain owned by one of your tracked competitors, worked out from your competitor list (their websites and owned domains); it replaces the stored type.
Different domain types carry different levels of authority and trust. Editorial and institutional sources are often perceived as more credible, while UGC sources may indicate community sentiment. Understanding which types AI models prefer helps you tailor your content strategy.
Having your content appear across diverse domain types, especially high-authority types like Editorial and Institutional, enhances your brand's credibility and reach. It also helps you understand where your brand fits in the information ecosystem and identify opportunities to expand your presence.
URL Types categorize individual pages by their content format and purpose: Article, Listicle, How-to guide, Document, Comparison, Alternative, Homepage, Category page, Product page, Discussion, and Other.
URL Types come from the pipeline's classification of each page: Article (blog posts, news articles), Listicle (ranked lists), How-to guide (tutorials, guides), Document (PDFs and other documents), Comparison (X vs Y pages), Alternative (alternatives-to pages), Homepage (a site's home page), Category page (section or listing pages), Product page (product/service pages), Discussion (forum and community threads), and Other.
Different URL types serve different purposes in the customer journey. How-to guides may attract users in the awareness stage, while product pages are crucial for consideration. Understanding which types perform best helps you optimize your content mix.
Optimizing your URL type distribution can improve visibility across different stages of the customer journey. A diverse mix of URL types ensures your brand appears for various query types and user intents, maximizing your reach and relevance.
Mentions count the total number of times your brand is explicitly mentioned across all AI responses. This metric tracks the frequency of brand references.
Mentions = Total count of brand name occurrences across all AI responses in the selected time period. This includes all instances where your brand name appears in the response text.
Mentions provide a raw measure of brand presence and frequency. Higher mention counts indicate stronger brand presence and suggest that your brand is top-of-mind for AI models when generating responses.
Increasing mentions can lead to improved brand recall and recognition. Frequent mentions reinforce your brand in users' minds and can influence their decision-making process. Tracking mentions helps you measure the effectiveness of your brand-building efforts.
The "Used" metric is the share of all citations ("Cited Sources Only") or of all sources ("All Sources") in the selected period that came from this domain or URL.
Used % = (Times the domain or URL appears as a cited source, or as a retrieved source with "All Sources") / (All cited sources, or all retrieved sources) × 100. On a domain's page, a URL's Used % is its share of that domain's total.
The "Used" metric reflects how authoritative your content is perceived to be and whether AI models trust your content as a reliable source. It tracks when your domain or content is utilized by AI models to generate responses, regardless of whether your brand is explicitly mentioned or your content is directly cited.
When your content is included among the sources AI models use, it can enhance the depth and accuracy of AI-generated information, contribute unique perspectives or data that shape the final response, and increase your content's visibility and authority within the broader ecosystem of referenced materials. Usage analysis helps you identify which content is most influential, understand your content's market position, and track improvements in content authority over time.
Average Citations is the number of citations (or, with "All Sources", retrieved sources) a domain gets per prompt run in the selected period.
Average Citations = Times the domain appears as a cited source (or as a retrieved source with "All Sources") / Prompt runs in the period. A prompt run is one prompt on one day, however many AI providers answered it.
Average citations reflect how authoritative your content is perceived to be, whether AI models trust your content as a reliable source, and your content's ability to serve as a comprehensive reference. Higher average citations indicate that your content covers topics comprehensively, AI models find your content relevant for multiple aspects of a topic, and you have strong content depth and breadth.
Citation analysis helps you identify which content is most influential, understand your content's market position, and track improvements in content authority over time. Content with higher average citations is more likely to be seen as authoritative and trustworthy, which can significantly impact brand perception and user decisions.