What counts as a citation
A citation is any URL an AI engine links to as part of its answer for a tracked prompt. Prefer stores the citing engine, the prompt, the position in the answer, and the destination URL. Deduplication is per (run, prompt, engine, URL).How Prefer classifies sources
Each cited URL is tagged with a source type so you can filter by intent rather than raw domain:- Owned: your own domain and known subdomains.
- Earned: press, podcasts, and third-party writeups.
- Third-party: general reference sites and neutral publishers.
- Competitor: a tracked rival’s domain.
- Listicle: “best of” and comparison round-ups.
- Review: G2, Capterra, TrustRadius, and category equivalents.
- Forum: Reddit, Hacker News, Stack Overflow, and similar community threads.
- Directory: category directories and marketplaces.
- Community: Discord, Slack communities, and specialist forums.
Where you read source data
- Monitor > Overview > Top cited content: your owned URLs ranked by citation share.
- Monitor > Competitors > Sources: which domains cite which brands, split per engine.
- Prompt detail > Top Domains / Top Citations: per-prompt source lists with links out.
Find pages you should own
1
Open Competitors > Sources
Filter to a single competitor.
2
Sort by citation share
The top rows are the sources that competitor “owns”: pages that engines cite for their brand and not for yours.
3
Send to Action Center
Use the row action to create an Action for outreach, guest content, or an owned page that can compete for the same slot.
Related
Competitors
Compare source ownership across every tracked rival.
Action Center
Turn source gaps into ranked outreach and content work.
