Insights ·
ChatGPT traffic didn’t disappear after GPT-6. Much of it is landing in Direct
The short version
Since OpenAI began rolling out GPT-6 on October 7–8, 2026, many links clicked in ChatGPT reach websites without the utm_source=chatgpt.com tag, and often without a referrer, so analytics tools count those visits as Direct. Adobe and Profound data both point to the same pattern. OpenAI has not said whether the change is intentional, so treat a ChatGPT referral drop since October 9 as a possible attribution gap before you treat it as lost visitors.
What changed
OpenAI began rolling out GPT-6 to Plus, Pro, Business and Enterprise users on October 7, 2026, and to Free and Go users on October 8. The release also introduced Intelligent UI, which lets ChatGPT answer with graphics, buttons, forms, charts and interactive tools inside the conversation.
On October 14, Search Engine Roundtable reported that OpenAI appears to have removed or replaced the UTM parameters it had added to ChatGPT links in June, which makes a ChatGPT click harder to tell apart from other visits. The report included no statement from OpenAI.
SEOAST Impact Rating: A3, good to know. The evidence is vendor and practitioner data plus trade reporting, and OpenAI has not confirmed or explained the change. The Impact Rating is SEOAST’s own editorial rating, not a Google or OpenAI severity level.
The evidence so far
Two independent data sets show the same shift. Both are vendor data, not figures from OpenAI.
Adobe (Cedric Huesler) looked at about 500 pages where ChatGPT drove 5% or more of visits and compared October 9–13 with the week before. On those pages, ChatGPT-attributed visits fell by about 31,000 a day while Direct visits rose by about 26,000 a day. Direct went up on two out of three pages that lost ChatGPT visits, where the week before it had been roughly a coin flip. On mobile, the share of Direct visitors seeing a consent banner, a rough proxy for new visitors, moved from 36% to 44%, closer to the 55% seen for known ChatGPT visitors.
Profound tracked the share of ChatGPT answers it collected where at least one cited link carried a tracking tag: about 98% on October 8, 75% on October 9 and 66% by October 14. Over the same days, the share of answers using Intelligent UI went from 2% to about 35%. Profound does not state its sample size.
Practitioners report the same pattern. Sébastien Billard posted that ChatGPT human visits dropped sharply from October 9, alongside fewer ChatGPT-User hits, while OAI-SearchBot activity rose around October 7.
Why it is happening: two explanations, neither confirmed
Until OpenAI comments, both are informed readings of the data, not confirmed causes. If the change is unintended, the tags may come back.
- Link opening: Huesler reports that GPT-6 opens links with window.open(url, "_blank", "noopener,noreferrer") and drops the utm_source=chatgpt.com parameter. With no tag and no referrer, analytics has nothing to attribute, so the visit lands in Direct.
- Intelligent UI rendering: Profound’s assessment is that links rendered inside Intelligent UI often lack the tag, while some links shown after a user opens an extra product or information panel still carry it. If that is right, attribution depends on which link in the same answer a person clicks.
What to do this week
The early numbers suggest the visits still arrive and only the label changed. Plan around that until OpenAI says otherwise.
- Annotate October 7–9 in your analytics. Any trend line across those dates now mixes two attribution regimes.
- Watch Direct on the pages that used to get ChatGPT clicks. A Direct bump on a deep, long-tail article is more likely ChatGPT than a returning customer typing the URL.
- Compare landing-page mix, not just totals. If Direct starts landing on the same pages ChatGPT used to send people to, that is the shift.
- Don’t cut AI search work on one week of attribution data.
- Check server logs for OpenAI’s crawlers and fetchers. Logs do not depend on UTM tags.
What SEOAST checks, and what it does not
SEOAST does not measure traffic or read your analytics, and it does not track whether ChatGPT cites you. SEOAST does not query ChatGPT, Perplexity, Gemini or Google AI Overviews, and it has no AI visibility score, so it cannot tell you how many visits ChatGPT sent.
What it does is audit, from the server HTML, the single page a ChatGPT click lands on: whether it returns a 200, whether robots meta and X-Robots-Tag allow indexing, whether the main content is in the server-rendered HTML, whether the headings are in order, and whether its JSON-LD parses and which @type values it declares. It also reads your robots.txt and lists which AI crawlers, including OAI-SearchBot and ChatGPT-User, are disallowed at the site root. That is reported, not scored.
Questions
Why did ChatGPT referrals drop in my analytics after October 9, 2026?
Since GPT-6 rolled out on October 7–8, many ChatGPT links arrive without the utm_source=chatgpt.com tag and often without a referrer, so analytics files them as Direct. Adobe and Profound data show the shift; OpenAI has not explained it.
Has OpenAI confirmed the change?
No. As of October 16, 2026, OpenAI has not said whether the change is intentional.
Are all ChatGPT links affected?
No. Profound found some links shown in extra product or information panels still carry the tag, so attribution can depend on which link a person clicks.
Can SEOAST show how much traffic ChatGPT sends me?
No. SEOAST audits one page from its server HTML. It does not read analytics, track citations or query ChatGPT.
Sources
- GPT-6 for everyone — OpenAI
- OpenAI Drops UTM Parameters From ChatGPT Results (GPT-6) — Search Engine Roundtable
- ChatGPT missing UTM tags — Profound
- Sébastien Billard on ChatGPT traffic since October 9 — X
- The day ChatGPT traffic “disappeared” from analytics, and where it went — The Prompt Insider