What Is Dark Traffic?

Dark traffic is website traffic whose true source cannot be identified, so analytics files it under "direct" or "unassigned." It has always existed — from apps, email clients, and HTTPS downgrades — but AI assistants have enlarged it sharply, because answers are consumed inside chat interfaces and outbound links frequently strip the referrer. The result is a growing gap between the influence AI has on demand and the traffic your dashboard can trace to it.

Why do AI assistants generate so much dark traffic?

Three mechanisms compound. First, many AI answers never produce a click at all — the value is the mention. Second, in-app and mobile browsers often omit the Referer header. Third, some platforms route citations through redirects or apply strict referrer policies. A user can read your brand in a Perplexity answer, screenshot it, and buy a week later with zero traceable path.

How does dark traffic distort reporting?

• It inflates the "direct" channel, masking real acquisition sources.
• It makes last-click models systematically undercredit AI discovery.
• It hides rising demand behind flat referral numbers, so teams underinvest in citation tracking.

How do you shrink the blind spot?

Layer three tactics. Capture raw server-side referrers before they are lost. Add a self-reported "How did you hear about us?" field at sign-up — the single most reliable correction for AI-era attribution. And triangulate: correlate spikes in direct traffic to deep pages with known citation events in AI answers.

Example

A DTC brand watches direct traffic to a specific product page jump 3x the week after ChatGPT started recommending it for a category query. No referrer confirmed the link, but the sign-up survey showed 22% of new buyers naming "ChatGPT" — dark traffic made visible only by asking.