Glossary · Beginner
What Is a Source Map? Definition and Example
A source map is a ranked list of the domains AI engines cite when answering your category's prompts, grouped by source type. It shows which websites shape what buyers read about your market, and where your brand is present or missing on them. It's one of the most actionable outputs of a GEO audit.
What does a source map include?
| Domain | Type | Citation share | Mentions you? | Action |
|---|---|---|---|---|
| reviewsite.example | Review | 14% | Yes, outdated | Update profile |
| industryblog.example | Media | 9% | No | Pitch data story |
| forum.example | Community | 7% | Yes, mixed | Engage helpfully |
| competitor.example | Competitor | 6% | No | Publish comparison |
| yourbrand.example | Owned | 4% | - | Improve cited pages |
Illustrative example, not real data.
How do you build one?
- Run your prompt set and log every cited URL.
- Reduce URLs to domains and count citations.
- Tag each domain: owned, competitor, review, media, community, reference, other.
- Check whether each top domain mentions you, and whether accurately.
- Split by engine and intent; patterns often differ.
Example
Illustrative. A B2B brand discovers that, on compare prompts, three review sites and one analyst-style blog account for most citations, while its own blog barely appears. It shifts effort to review profiles and a data-led pitch to the blog.
Why it matters for GEO
Engines draw on a limited set of sources per category. A source map tells you which ones matter, so effort goes where it can influence the most answers.
Common mistakes
- Mapping only your own citations and ignoring the domains that hold the rest.
- Treating every high-share domain as an outreach target; some, like competitors' sites, need a content response instead.
- Building the map once and never refreshing it.
Keep the map in the same workbook as your prompt results, so each domain links back to the answers that cited it.
Related
Questions
What data do I need for a source map?
The cited URLs from your prompt set's AI answers, logged with engine, prompt and intent.
How often should I update it?
Monthly or quarterly. Citation patterns shift as engines and content change.
What do I do with it?
Prioritise outreach, reviews, partnerships and content on the domains that influence the most answers.