How to Track Visitors From ChatGPT in Google Analytics
Use ChatGPT's referral source in Google Analytics to inspect sessions, engagement, and conversions without confusing traffic with citations or recommendations.
When someone clicks a link from ChatGPT to your website, the visit may appear in Google Analytics 4 with chatgpt.com as the source. This gives founders a practical way to inspect sessions, engagement, and key actions from at least some ChatGPT referrals.
Referral reporting is useful, but it answers a narrow question: who clicked through and was measured on your site? It does not tell you every time your business was mentioned, cited, compared, or recommended inside an AI answer.
What ChatGPT referral tracking measures
Google Analytics measures website activity after the visitor reaches a page carrying your analytics tag and consent conditions allow measurement. In the session-scoped acquisition report, the source describes where the session was acquired.
Use this data to understand landing pages, engaged sessions, and business actions. Do not use it as a complete count of AI visibility. A user can read an answer without clicking, type your URL later, use a privacy tool, decline consent, or move between devices.
The chatgpt.com source signal
OpenAI says referral URLs from ChatGPT search include utm_source=chatgpt.com. When the parameter reaches your site and your analytics setup records it, GA4 can attribute the session source to chatgpt.com.
Historical visits and different ChatGPT surfaces may not behave identically, so inspect the actual values in your property. You may see chatgpt.com in Session source or a related value in Session source / medium. Keep the rule broad enough to catch the genuine source without grouping unrelated traffic into it.
Find ChatGPT traffic in Google Analytics
In a standard GA4 property, open Reports, then Acquisition, then Traffic acquisition. The exact navigation can vary if your property administrator has customized the report collection.
Set a useful date range, such as the last 28 or 90 days.
Use the table's primary dimension selector and choose Session source or Session source / medium.
Search or filter for chatgpt.com.
Review sessions, engaged sessions, average engagement time, key events, and total revenue where those metrics are configured.
Add Landing page as a secondary dimension, or open a landing-page report and apply the same session-source comparison.
Build a reusable comparison
Create a report filter, comparison, or exploration that includes Session source containing chatgpt.com. Name it clearly and use the same definition each month. Compare it with all users or other relevant sources, but avoid conclusions from a tiny sample.
A simple monthly table can show sessions, engaged-session rate, key events, and top landing pages. Add notes when you publish a major service-page change or run a campaign, so later changes have context.
Track meaningful actions
Traffic alone is not a business result. Configure the actions that matter—such as a qualified form submission, booked call, purchase, or trial start—as GA4 key events. Test each event before using it in a decision.
Then ask whether ChatGPT-referred sessions reach those actions and which pages help. A small number of well-matched visitors can be more valuable than a larger number of unqualified sessions, but the data should support that conclusion.
Do not confuse traffic, citations, and recommendations
A referral means a measured visit reached your site from a source.
A citation means an AI answer linked to a source in the observed response.
A mention means the business or page appeared in an answer, with or without a link.
A recommendation is a stronger answer-level judgment and still may not produce a click.
These signals overlap, but none is a substitute for the others. A citation can generate no clicks, while a person can visit later through search or a direct URL after first learning about the business in ChatGPT.
Why the numbers may be incomplete
Analytics can undercount because of consent choices, tracking prevention, blocked scripts, redirects that strip parameters, cross-domain configuration, and users who do not click. Attribution rules can also assign a session differently from the way a founder intuitively remembers the journey.
Check your own implementation before diagnosing a visibility problem. Confirm that the analytics tag loads on the landing page, referral parameters survive redirects, key events fire once, and internal referrals or payment domains are handled correctly.
A monthly founder checklist
Use the same reporting date range and chatgpt.com session-source rule.
Record sessions, engaged sessions, key events, and top landing pages.
Check whether unusual spikes came from a campaign, one popular page, or tracking changes.
Compare conversion quality, not only traffic volume.
Test a small set of buyer questions separately to observe mentions and citations.
Write down limitations instead of presenting referral traffic as complete AI visibility.
Primary sources used
OpenAI, Publishers and developers FAQ: https://help.openai.com/en/articles/12627856-publishers-and-developers-faq
Google Analytics Help, Traffic acquisition report: https://support.google.com/analytics/answer/12923437
Google Analytics Help, Dimensions and metrics: https://support.google.com/analytics/answer/14731736