What Is a Referrer? How Traffic Sources Are Detected
The HTTP Referer header explained: how analytics classifies sources, why referrers go missing, and what referrer policy changes.
The referrer is the URL of the page a visitor came from, carried in the HTTP Referer header (misspelled in the 1996 spec, immortal ever since). It is how analytics knows a visit came from Google, a newsletter, or a forum thread — when it survives the trip, which is decreasingly often.
How classification works
Analytics buckets traffic by referrer domain: search engines → organic, social platforms → social, everything else → referral, and no referrer at all → direct. UTM parameters, when present, override the referrer — explicit tagging beats inference.
Why referrers go missing
- Apps and email clients: links opened from native apps (mail, WhatsApp, Slack) usually carry no referrer — the engine of dark traffic.
- Referrer-Policy headers: the modern default (strict-origin-when-cross-origin) strips paths on cross-site navigation — you see news.ycombinator.com but not which thread.
- HTTPS → HTTP: browsers send nothing downgrade-ward (rare now, but historic).
- Privacy browsers and extensions that trim or remove the header outright.
Reading referrer data correctly
- Treat referrer-based source numbers as floors, not totals — the stripped share landed in direct.
- Origin-only referrers mean you cannot know the page that linked you; if the source matters, ask the linker or watch for traffic-shape clues.
- For links you control, never rely on referrers at all: UTM-tag everything and the classification becomes deliberate instead of forensic.
Privacy note: the referrer arrives with a normal page request and is processed as aggregate source data by cookieless tools — no cross-site profile required. It is the rare tracking signal that predates, and survives, the entire cookie era.