Sharing, Links & Measurement

Referral Traffic

Definition

Referral traffic is the slice of your visitors who came from somewhere else on the web and whose origin the browser disclosed. The disclosure travels in an HTTP request header spelled Referer — a misspelling baked into the 1996 specification and never corrected. Analytics tools group those visits by the sending domain, so you see rows like news.ycombinator.com, a supplier's resources page, or a forum thread you did not know existed. Search engines are usually split out into their own category, and paid clicks into another. What is left over — email clients, chat apps, PDFs, desktop software, QR codes — sends no header at all and is recorded as direct, which is why direct is never simply people typing your address.

Why It Matters

The referrer report is the cheapest way to learn who is talking about your work without being told. A spike from one unfamiliar domain usually means a link in a newsletter or a community post, and the sensible response is to go and read the thread. It is equally important to know what the report cannot see. A campaign that lives entirely in email and WhatsApp can drive 500 visits and show up as a wall of direct, which reads as no traffic to anyone skimming the numbers. That is exactly the gap a UTM parameter closes, and the reason tagged links exist at all.

How It Works

When someone clicks a link, the browser attaches a Referer header naming the page they left, and the receiving server records it. How much of that address arrives is governed by the sending page's Referrer-Policy header. The common default, strict-origin-when-cross-origin, sends only the scheme and domain to another site, so you learn that example.com sent the visit but not which article did. A page can send no-referrer and disclose nothing at all. A link from HTTPS to plain HTTP strips the header entirely, and so does most software that opens a link outside a browser tab. Edge analytics on 99helpers records the referrer alongside the request without setting any cookie, which is enough to rank sending domains but never enough to follow one person between them.

Real-World Example

A studio publishes a 3 MB type specimen at grotesk-specimen.99helpers.site and mentions it once in a Slack community. Over the next week the edge report shows 180 visits from a design newsletter's domain, 60 from a forum thread, and 240 direct. The direct figure is the Slack link, invisible to the referrer header. Adding utm_source=slack to the version posted the following month moves those visits into their own row and turns a shrug into a decision.

Common Mistakes

  • Reading direct traffic as people typing your address — most of it is links opened from email, chat apps and documents that send no referrer
  • Expecting to see the exact page that linked to you — a strict Referrer-Policy on the sending site reduces it to the domain
  • Leaving a self-referral in the report and counting it as new traffic — links from one of your own pages to another are not referrals

Related Terms

Put a file online in seconds

Drop in a document, an image, a page or a whole static website and share the link — free, with no build step and no server to set up.

Host a file free →