Module 03 · Lesson 3 of 4
Citation Source Analysis
Citation source analysis answers one question: in this niche, who does AI search actually use, and why? Not who you assume. Who the log says. The method is grouping your logged citations by source type and domain, then interrogating the repeat winners.
Build the source-type table
Take every cited domain from your last two or three runs and count it. Then label it. The labels differ by niche, and discovering your niche's labels is half the point. Common ones:
| Type | Why it wins | Can you compete? |
|---|---|---|
| Official documentation | Primary for platform facts | Only for facts you are primary for |
| Big publishers | Breadth, brand, speed | Rarely head-on; via coverage instead |
| Niche specialists | Depth and trust in one field | Yes: this is your lane |
| Comparison / data pages | Structured, current, complete | Yes: build the better structured page |
| Communities and forums | First-hand experience, honesty | By participating honestly |
| Video channels | Demonstration on screen | If you can show, not just tell |
Interrogate the winners
For each recurring domain, open the exact cited page and ask four questions. What claim is it primary for? How is the passage built (structure, length, placement)? What corroborates it (links in, brand mentions, author credentials)? What does it have that your equivalent page lacks? Answer in writing, one line each. Patterns appear fast: most niches have two or three source types doing most of the winning, and the reason is usually a mixture of primary status and structure, not magic.
Then split the response
Analysis ends in a decision per recurring winner: own it or earn it. Own it means the winning slot is winnable with your own content: you can be primary, better structured, more specific (Module 4 builds those pages). Earn it means the slot belongs to a third party: a publisher, a comparison site, a community, and your move is to earn coverage there, which is Module 6's work. Trying to own every slot produces mediocre pages; trying to earn every slot leaves your own site hollow. The log tells you which is which.
- Cluster: "ai seo course" queries, three runs, eleven citations logged.
- Counts: publisher-A four times, comparison-site-B three times, two one-off blogs, vendor-C twice, community-D twice.
- Labels: publisher-A big publisher, comparison-site-B comparison page, vendor-C vendor, community-D community.
- Interrogation: publisher-A wins on brand and speed; comparison-site-B wins on structure and completeness; community-D wins on first-hand course experience.
- Decisions: comparison page: own it (build a better one, Module 4). Publisher-A: earn it (coverage, Module 6). Community-D: earn it (honest participation). One-off blogs: ignore; not recurring.
- Export all cited domains from your last two runs. Count and label them.
- Build the source-type table with a count column for your niche.
- Pick the three recurring domains that cost you the most visibility.
- For each, write one line: own it or earn it, and why.
- Star every "own it" line: those are your Module 4 briefs, already researched.
Why count only recurring domains?
What does "own it or earn it" mean?
Should you try to be cited on every query?
Sources used in this lesson
Google Search Central: AI features and your website
Google Search Central: helpful, reliable, people-first content