Understanding Metrics
Complete guide to every metric on your Sill dashboard: what they measure, how they're calculated, and what good scores look like.
Your Sill dashboard tracks how AI engines perceive and recommend your brand across several key metrics. This guide explains each one, how it's calculated, and how to interpret your scores.
Topline Metrics
These five metrics appear as cards at the top of your dashboard. Each is a percentage (0–100%) with a trend indicator showing change from your previous monitoring run.
AI Visibility
What it measures: Your brand's overall Share of Voice across AI responses. This is the single most important metric, answering "how prominently do AI engines recommend me?"
How it's calculated: Each time an AI engine mentions your brand, Sill scores that mention based on two factors:
- Positioning: Was your brand the primary recommendation, a secondary option, or just mentioned in passing?
- Sentiment: Was the mention positive, neutral, or negative?
These two scores are multiplied together and averaged across all monitored prompts to produce your AI Visibility score.
How to interpret it:
| Score | Meaning |
|---|---|
| 60%+ | Exceptional: Consistently recommended as a primary or secondary choice with strong sentiment |
| 45–60% | Strong: Frequently recommended, mostly positive sentiment |
| 30–45% | Moderate: Mentioned regularly, mix of primary/secondary/mentioned positioning |
| 15–30% | Weak: Mentioned but rarely recommended prominently |
| Below 15% | Minimal: Rarely appears in AI responses |
How to improve it: Strengthen your content with clear product positioning, structured data, and authoritative sources. See your report's recommendations for specific actions.
Mention Rate
What it measures: The percentage of AI responses that mention your brand at all, regardless of how favorably or prominently.
How it's calculated: For each prompt Sill monitors, it checks whether your brand name appears anywhere in the AI response. Mention Rate is simply:
(Responses that mention you / Total responses) x 100
How to interpret it:
| Score | Meaning |
|---|---|
| 80%+ | Very High: AI engines almost always include you in relevant discussions |
| 50–80% | Good: Mentioned in most relevant queries |
| 20–50% | Moderate: Mentioned selectively, depends on query type |
| Below 20% | Low: Rarely mentioned, low AI awareness of your brand |
How to improve it: Ensure your brand appears on platforms AI engines reference: Wikipedia, review sites, Reddit, industry publications. A higher off-site presence directly correlates with mention rate.
Citation Share
What it measures: The percentage of URLs that AI engines cite that belong to your domain. This is independent of mentions, so you can be mentioned without being cited, and vice versa.
How it's calculated: When AI engines provide source links alongside their responses, Sill collects those URLs and checks how many point to your domain:
(URLs pointing to your domain / Total cited URLs) x 100
How to interpret it:
| Score | Meaning |
|---|---|
| 40%+ | Strong: Your content is a go-to source for AI engines |
| 20–40% | Good: Solid citation presence |
| 5–20% | Emerging: Appearing in citations but competing heavily |
| Below 5% | Minimal: Rarely cited directly |
How to improve it: Create comprehensive, well-structured content that AI engines want to link to. Use schema markup, include original data and statistics, and maintain strong topical authority.
Content Score
What it measures: How well your website is optimized for AI discovery and citation. This audits your actual content against the patterns AI engines favor.
How it's calculated: Sill audits your primary domain and up to four additional pages, scoring them on factors including:
- Word count: Comprehensive content (target: 2,900+ words for key pages)
- Answer capsules: Direct answers placed immediately after headings
- Statistics and data points: Original numbers AI engines can reference
- Expert quotes and citations: E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) signals
- Comparison tables: Structured data AI can easily parse
- Schema markup: JSON-LD structured data that helps AI understand your content
- Author information: Clear attribution and expertise signals
How to interpret it:
| Score | Meaning |
|---|---|
| 80%+ | Excellent: Your content follows AI-friendly best practices |
| 60–80% | Good: Strong foundation with some optimization opportunities |
| 40–60% | Moderate: Multiple areas to improve |
| Below 40% | Needs Work: Significant gaps in AI-scannable content |
How to improve it: Follow the specific content recommendations in your AI Visibility report. Focus on adding structured data, answer capsules, and original statistics.
Off-site Presence
What it measures: Your brand's visibility across external platforms that feed AI training data and inform AI responses.
How it's calculated: Sill scans five platform categories and scores each one:
- YouTube: Do you have video content about your brand?
- Wikipedia: Does your brand have a Wikipedia page?
- Review platforms: Are you listed on G2, Capterra, TrustPilot, etc.?
- Reddit: Is your brand discussed in relevant subreddits?
- Press: Do major publications cover your brand?
The scores across all five categories are averaged to produce your Off-site Presence score.
How to interpret it:
| Score | Meaning |
|---|---|
| 80%+ | Established: Strong multi-platform authority |
| 60–80% | Good: Present on most key platforms |
| 40–60% | Present: Visible on some platforms, gaps exist |
| Below 40% | Limited: Minimal off-site visibility |
How to improve it: Prioritize the platforms where you're absent. Getting listed on review sites, creating YouTube content, and earning press coverage all contribute to how AI engines perceive your brand authority.
Sentiment & Position
These metrics appear in the Competitor Rankings table and the Sentiment tab.
Brand Sentiment
What it measures: The average tone when AI engines mention your brand. Displayed as a hero metric in the Sentiment tab.
How it's calculated: Each AI response mentioning your brand is classified into one of five sentiment levels:
| Level | Score |
|---|---|
| Very Positive | 100 |
| Positive | 75 |
| Neutral | 50 |
| Negative | 25 |
| Very Negative | 0 |
Your Brand Sentiment is the average of these scores across all mentions.
Key insight: Sentiment measures tone, not frequency. A high sentiment score with low visibility means AI engines speak favorably about you when they do mention you, but they don't mention you often enough.
Sentiment Score
What it measures: The same as Brand Sentiment, but shown per-brand in the Competitor Rankings table. This lets you compare how favorably AI engines discuss you versus competitors.
A score of 50 is neutral. Above 75 indicates consistently positive mentions. Below 25 suggests AI engines highlight weaknesses or concerns.
Position Score
What it measures: Where your brand appears in AI recommendation hierarchies. Are you the top pick, an alternative, or just mentioned?
How it's calculated: Each mention is classified by its structural position:
| Position | Score | Meaning |
|---|---|---|
| Primary | 100 | Main recommendation ("Best Overall", "Our Top Pick") |
| Secondary | 66 | Strong alternative (top 2–5, "Best for...") |
| Mentioned | 33 | Referenced but not recommended ("Also consider", "Others include") |
| Negative | 0 | Explicitly discouraged ("We don't recommend") |
Your Position Score is the average across all responses.
Key insight: Position Score measures structural ranking, not sentiment. A brand can be the primary recommendation but with caveats (high position, moderate sentiment), or be mentioned favorably but only in passing (low position, high sentiment).
Competitor Rankings
The Competitor Rankings table shows your top competitors ranked by AI Share of Voice (Visibility). Each competitor shows:
- Visibility (%): Their Share of Voice score, calculated the same way as yours
- Sentiment: A score bar (0–100) showing how positively AI engines discuss them
- Position: A score bar (0–100) showing their average recommendation positioning
Your brand is always shown in the table (highlighted) even if it falls outside the top 5. You can sort by any column to compare different dimensions.
Charts & Trends
Competitor Visibility
A line chart showing your Share of Voice over time compared to your top competitors. Each line represents a brand, with your brand highlighted in green. Use this to track whether you're gaining or losing ground relative to competitors.
Citation Share Trend
An area chart showing how your citation share percentage has changed over time. Rising trends indicate AI engines are increasingly linking to your content.
Sentiment Trend
A line chart comparing your brand's sentiment score against competitors over time. The dashed line at 50 marks neutral, with scores above being positive and below being negative.
Sentiment Distribution
A stacked bar showing the breakdown of all AI responses mentioning your brand by sentiment level, from Very Positive to Very Negative. This gives you a quick visual of the overall tone distribution.
Sentiment Keywords
A word cloud showing the most frequent sentiment-related keywords AI platforms associate with your brand across all responses. Larger words appear more frequently. Use this to understand the themes and attributes AI engines emphasize when discussing your brand.
Top Sources
A table of the most frequently cited domains when AI engines discuss brands in your category. This shows which websites AI engines trust and reference most, helping you identify where to seek coverage or backlinks.
Quick Reference
| Metric | Excellent | Good | Moderate | Needs Attention |
|---|---|---|---|---|
| AI Visibility | 60%+ | 45–60% | 30–45% | Below 30% |
| Mention Rate | 80%+ | 50–80% | 20–50% | Below 20% |
| Citation Share | 40%+ | 20–40% | 5–20% | Below 5% |
| Content Score | 80%+ | 60–80% | 40–60% | Below 40% |
| Off-site Presence | 80%+ | 60–80% | 40–60% | Below 40% |
| Sentiment | 75+ | 50–75 | 25–50 | Below 25 |
Need help interpreting your specific scores? Reach out to our team at daniel@trysill.com.