Google Search referrals: check what your report counts

Printed charts, notebooks and part of a laptop viewed from above.

An SEO report can show a drop in Google visits even when the first problem to investigate is the channel definition. Google’s new referral-identification guidance, updated 12 November, gives analytics teams a reason to check their measurement rules before changing content.

The guidance describes how the HTTP Referer header can identify arrivals from Google. It also explains the limits: the header doesn’t supply search queries, and matching these visits won’t reproduce Search Console click totals exactly.

Define the number you need

Write a sentence above the report describing what it counts. For example: landing-page requests with a recognised Google referrer during the selected period. Keep that definition alongside the filter settings so the next analyst can reproduce it.

Use a small sample to review the current classification. Choose dates that include a normal working day and an unusual traffic day. Keep the original output before making changes; otherwise a corrected rule can erase the evidence needed to explain the earlier discrepancy.

Inspect the classification rules

Google says most analytics systems recognise its referrer values automatically. Where visits appear as direct or referral traffic, check custom channel rules, exclusions and filters. Its documentation includes local Google domains and an Android app format, so a rule covering only one web origin deserves review.

Ask the analyst to show matched and unmatched examples with identifying information removed. Read the full rule rather than accepting a screenshot of the channel name. Document which values it accepts and who owns future changes.

Keep the limits beside the chart

Google notes that other services sharing these origins, including Images, can send the same values. Browsers may omit the header. In server logs, count landing-page requests rather than the page’s supporting resources. Use Search Console for query information.

Keep a short limitations note in the dashboard. If a stakeholder asks for a precise split the evidence can’t provide, mark it as unavailable. Avoid filling the gap with a guessed allocation that later becomes a target or a budget assumption.

Investigate differences before assigning causes

Put the analytics view and Search Console report side by side with matching date labels. Record each system’s definition. Google’s explanation points to privacy controls, redirects and measurement methods as possible reasons for disagreement; these are investigation leads rather than a diagnosis of your site.

Assign one person to trace a representative journey and another to review the report settings. Agree what observation would support each suspected cause. Keep a separate list of checks that need developer access, so an incomplete investigation doesn’t become an unsupported conclusion.

For teams building that handover, our article on turning SEO learning into a site diagnosis offers a related approach to evidence and ownership. Close this measurement review with the revised definition, the comparison result and any remaining uncertainty attached to the report.

Want to rank higher and drive more organic traffic?

Technical SEO, content strategy, and performance optimization—we help businesses get found. Let's discuss your SEO goals and create a roadmap.