Citations & Sources
The domains and pages AI assistants cite when answering your prompts — the supply side of everything the other metrics measure.
AI answers are assembled from sources. The Sources leaderboard shows which domains those are for your prompts — the most direct evidence you have of why the other metrics read the way they do.
The columns
| Column | What it counts |
|---|---|
| Citations | How many times the domain was cited across all answers in the period |
| Citation share | That domain's citations as a percentage of all citations counted |
| Coverage | The percentage of answers that cite the domain at least once |
| Domain Rating | A 0–100 estimate of the domain's link authority |
| Type | What kind of source the domain is (see below) |
Source types
Owned and Competitor come straight from your brand setup — your own domain and your tracked competitors' domains. The rest describe what kind of site it is:
| Type | Sites it covers |
|---|---|
| UGC | User-generated content: social networks, forums, Q&A, video sharing, open publishing, review sites |
| Editorial | Professionally edited publications: news outlets, magazines, trade media |
| Reference | Works of record: wikis, encyclopedias, dictionaries, academic databases |
| Institutional | Official bodies: government, education, intergovernmental organisations, NGOs |
| Corporate | A company's own site promoting its products or services |
| Other | Uncategorised: URL shorteners, CDNs, parked domains, unidentified |
The type tells you what kind of work a source needs. Editorial and UGC sources respond to outreach and participation; Reference sources respond to accuracy and citations; Corporate sources are usually competitors you have not tracked yet.
Citations and coverage answer different questions, and the gap between them is the useful part:
- High citations, low coverage — one or two answers lean on the domain heavily. A deep but narrow source.
- Moderate citations, high coverage — the domain is cited a little in most answers. This is the profile of a source assistants treat as a default reference for the topic.
Worked example
Across 36 answers, 300 citations are counted. A review site accounts for 45 of them and appears in 27 answers.
Citation share = 45 / 300 × 100 = 15%
Coverage = 27 / 36 × 100 = 75%
Three quarters of the answers in this topic pass through that one site. Being absent or poorly represented there caps your visibility no matter what you publish on your own domain.
Why sources matter
- They are the actionable layer. You cannot edit an AI answer, but you can work on the pages behind it — making sure your own are accurate and complete, and that the third-party sources shaping your category describe you correctly.
- They explain sentiment. How you are described tends to follow how the top-cited sources describe you.
- They rank the work. Coverage tells you which single source is worth the most effort in a topic.
Reading the table carefully
- Citations are counted per answer, so a period with more answers has more citations. Compare shares between periods, not raw counts.
- Your own domain appearing often is good but not sufficient — assistants weigh third-party corroboration, so a leaderboard consisting only of your own pages is a fragile position.
- Domain Rating is context, not a target. A low-rating niche source with high coverage in your topic matters more than a high-rating site nobody cites here.
- Not every assistant exposes sources for every answer. Answers without citations contribute to visibility metrics but not to this table.
Related metrics
- Visibility — the outcome sources drive
- Sentiment — how those sources describe you
- Sources tab — filtering and drill-down
Updated 2026-08-03