Search Console reports a click, GA4 reports no visit, and the server log gives you a third, different count: requests. Before changing the page or its tracking, inspect the landing-page request and note what each system can actually tell you.

On 12 November 2026, Google announced new documentation on identifying Search traffic in analytics and logs. Inspect the landing-page request, record its Referer header when present, and keep that observation separate from Search Console clicks and GA4 sessions. The header gives a clue about a request; it is not a complete visitor count.

What is a referrer?

When a browser requests a page, it may send an HTTP header named Referer (that spelling is part of the standard header name). It can tell the receiving site where the request came from. Google says Search uses an origin referrer policy: the header normally contains a scheme and host, such as https://www.google.com/, without the search-result path or the person’s query. To find search queries, use the Search Console Performance report.

Google documents several possible values. Web Search may send https://www.google.com/ or a local Google domain. A click from the Google app on Android may send an android-app://com.google.android.googlequicksearchbox/https/www.google.com value when it opens a browser tab, or a Google web origin; the iOS app may send a Google web origin. Google’s source table gives the full wording.

Those origins are not exclusive to ordinary Web Search. Google says other services, including Google Images, can send the same values; Discover uses the value of the surface where it was viewed. A matching header therefore cannot, by itself, tell you “this was a Web Search click.”

Isolate the landing-page HTML request

Ask whoever has authorized access to your site’s request logs for a short, relevant period. Limit the view to requests for the HTML landing page you are investigating. Record the request time, destination path, response status and Referer value if the logging setup captures it. Avoid collecting or sharing unnecessary visitor identifiers.

As a practical filter, start with the specific page path, request method GET, an HTML content type and a 200 response. Inspect 3xx redirects as a separate path through the site; do not count a redirect and its destination as two arrivals. Exclude asset extensions such as .css, .js, images and fonts. Log field names differ by server and CDN, so map these checks to the fields your setup actually records. These filters are practitioner suggestions; Google’s guide says to count landing-page requests rather than resource requests.

After a page loads, the browser may request its images, CSS and JavaScript. Those asset requests often carry the page itself as their referrer. Counting every line with a header would inflate a supposed visit count.

Illustrative example, invented numbers: A guide page receives one HTML request with Referer: https://www.google.com/, followed by five asset requests whose referrer is the guide’s own URL. Inspect the HTML request as a possible Google-origin arrival; do not call the six requests six Search visits. Even the first request cannot identify which Google surface produced it.

In the log What you can say What you cannot conclude
HTML request with a documented Google origin A request reached this page with a Google-origin referrer The query, exact Google surface, person behind the request or a GA4 session
HTML request with no Referer A request reached this page without a recorded referrer Whether it came from direct navigation or a source that removed the header
CSS or image request referring to your own page The browser fetched a page resource That this asset request represents another landing-page arrival

If the header is blank, check the limits

A blank Referer is common enough that you should not label every such arrival “direct.” Google’s new guide lists privacy settings, anti-fingerprinting, some new-tab and app-to-browser paths, and extensions among reasons the header may be missing. It also notes that consent choices and ad blockers can stop JavaScript analytics from recording visits.

The Referer header is supplied by the client and can be absent or forged. A Google-origin referrer is therefore a clue about a request, not proof of a human Search visit.

Check the path between Search and the landing page. A client-side redirect can replace the original referrer with your own page. A redirect with a stricter Referrer-Policy may remove it. If the landing page is unexpectedly routed through another URL, investigate that routing before changing content. Treat these as checks, not assumptions about your site.

Keep three different numbers separate

Search Console records a click when someone selects a result. GA4 records a session after the destination loads and tracking runs. A server log records a request if the server or edge logs it. These are different events, and none should be forced to match one-to-one. Google also notes that Search Console excludes ad clicks while an analytics view may include paid traffic depending on its configuration.

Google lists two more reasons for a mismatch: GA4 excludes known bots while Search Console may not filter them, and Search Console reports on the Google-selected canonical URL while GA4 can record other tracked URLs.

If GA4 calls Google-origin traffic “referral” or “direct,” check channel rules, referral exclusions and filters before declaring a search visibility loss. For the broader clicks-versus-sessions comparison, use the Search Console and GA4 discrepancy guide. If you need to choose a page for a reader-facing update after the measurement check, use the page-prioritisation guide.

End with a short investigation note: the exact dates and timezone, page path, log fields available, count of relevant HTML requests, documented Google-origin examples, missing-header cases, redirect checks and limits. If the evidence points to a measurement problem, fix or escalate that problem. If it points to a page problem, describe the reader issue separately. For help with a wider measurement and content review, discuss an SEO consulting project.

Reviewed 15 November 2026. Sources: Google Search Central's Identify Google Search traffic in your web analytics or server logs; Latest documentation updates; Using Search Console and Google Analytics data for SEO.

Vishnu Sudevan

Senior SEO, AEO & GEO consultant and practitioner. I write about search strategy, the tools I build and lessons from hands-on work.

More about my work →