Sources — how citations are attributed and de-duplicated

TL;DR — Every URL an AI engine cites is captured, normalised, grouped by domain, and bucketed as owned / third-party / competitor-owned. That's how we tell you where the AI's attention actually lives for your category.

What "source" means here

When an AI engine answers a prompt, it often (not always) shows the URLs it drew from. Some engines expose these openly (Perplexity, Google AI Overviews), some show them on hover (ChatGPT with browsing), some only surface them via structured API responses.

For every response we capture, we log the citations. Over the course of a week, that adds up to hundreds or thousands of URL captures for your category.

Normalisation

Raw citations are messy — the same page shows up as example.com/post?utm=x, www.example.com/post/, example.com/post, and so on. We normalise every URL by:

• Stripping tracking parameters (utm_*, fbclid, gclid, session IDs)
• Removing trailing slashes and fragment identifiers
• Lowercasing the host
• Collapsing www. variants

So a single page counts as one source, no matter how the engine formatted the citation.

Grouping by domain

The Sources tab shows normalised URLs grouped under their root domain. You can drill from reddit.com into the specific threads, or from a competitor's blog into the specific posts.

The three buckets

Every source is one of:

• Owned — a URL on your primary domain
• Third-party — a URL that's neither yours nor a tracked competitor's (Reddit, Wikipedia, roundup blogs, PR sites, YouTube, etc.)
• Competitor-owned — a URL on one of your tracked competitors' domains

The bucket assignment is deterministic — no fuzzy matching. If a domain isn't in your competitor list, it's third-party by default.

How to read the tab

Two useful mental models:

1. "Where is the attention?" — the top third-party sources tell you which forums, blogs, and directories AI engines trust for your category. That's your PR / guest-post / community-participation target list.
2. "Where is my attention?" — your owned sources tell you which of your pages are actually earning citations. Not every page ranks equally in AI eyes; the ones that do earn more citations should get more of your content investment.

What we don't do

• We don't guess citations. If an engine didn't expose the URLs for a given response, we log the response as "no citations shown" rather than fabricate a guess.
• We don't count self-links as third-party corroboration.
• We don't inflate counts by adding partial-URL matches — everything is on the full normalised URL.