TL;DR Summary:
Tracking Signals Vanished: After GPT-6 launched, ChatGPT links reportedly stopped carrying UTM tags and referral data, making visits harder to identify in analytics.Traffic Didn’t Disappear: ChatGPT referrals dropped by roughly 31,000 daily visits while direct traffic rose by about 26,000, suggesting misclassification rather than fewer visitors.Audit Before Reacting: Compare direct and ChatGPT traffic on affected pages, document the shift, and avoid changing AI marketing plans until OpenAI clarifies the cause.Why did my ChatGPT referral traffic suddenly drop? If you manage a website and watch your analytics, you may have noticed this exact pattern in the past week. OpenAI appears to have removed ChatGPT UTM tracking from outbound links in its chatbot, and the change is making it harder for site owners to know when a visitor actually came from ChatGPT.
What changed with ChatGPT UTM tracking after GPT-6
OpenAI released GPT-6 in October 2026. Shortly after, marketers noticed that links from ChatGPT stopped carrying the utm_source=chatgpt.com tag, a short code in a web address that tells analytics tools where a visitor came from. OpenAI added that tag back in June. Adobe’s Cedric Huesler reported that ChatGPT now opens links in a way that also strips out referral data, the background information a browser normally sends to show which site sent the visitor. Without either signal, your analytics software has no way to know the visit came from ChatGPT.
The numbers behind the ChatGPT UTM tracking disappearance
Huesler studied about 500 pages where ChatGPT previously made up at least 5% of visits. He compared October 9 to 13 against the prior week. ChatGPT traffic on those pages fell by roughly 31,000 visits per day. At the same time, direct traffic rose by about 26,000 visits per day on those same pages. Two out of three pages that lost ChatGPT referrals saw a matching rise in direct traffic. Search traffic also dropped 7% during this period, but Huesler linked that to broader search trends, not the ChatGPT UTM tracking issue.
Why this looks like a tracking problem, not fewer visitors
The pattern points to an attribution issue rather than an actual drop in people using ChatGPT to find websites. Huesler found that the newly classified direct visitors behaved much like the ChatGPT visitors you used to see in your reports. This matters because direct traffic, visits with no identifiable source, is usually the hardest category to explain. If ChatGPT visits are landing in that bucket, your reports will understate how much traffic the platform sends you. This is exactly the kind of blind spot that tools like Measuremate are built to surface, by auditing your GA4 setup to flag where referral data is quietly being misclassified as direct traffic.
What’s still unconfirmed about ChatGPT UTM tracking
OpenAI has not confirmed whether removing UTM tracking was intentional. The company has not issued a statement explaining the change. Huesler’s analysis also does not prove that every missing referral traces back to this one cause. You should treat the connection as a strong pattern, not a settled fact, until OpenAI responds or more analytics providers confirm the same trend.
What this means for your website reporting
If you track AI referral sources, check your direct traffic numbers against your past ChatGPT referral data. Look for pages that match the pattern Huesler described, a drop in ChatGPT visits alongside a rise in direct visits on the same pages, during the same window. Review landing-page performance alongside both metrics before you conclude that ChatGPT is sending you less traffic. A dashboard showing a decline in ChatGPT referrals right now most likely reflects a tracking gap, not a change in how people use the platform.
Watch your direct traffic closely over the next few weeks and compare it against historical ChatGPT referral patterns on your top pages. Hold off on any decisions about AI-driven marketing strategy until OpenAI clarifies whether this change was deliberate. The most useful thing you can do right now is document the shift in your own data, so you have a clear before-and-after record once more information becomes available. If this kind of attribution confusion sounds familiar, tools like Measuremate can help you audit your GA4 setup and catch tracking gaps before they distort your referral data.
