GA4, Server Logs and Traffic Counts: Compare the Same Unit

GA4 sessions, server requests, CDN requests and tracked button clicks represent different events. Their totals can all be internally correct while differing substantially. A gap is a question to investigate, not an automatic estimate of fake traffic.

Write down what each source measures

• GA4: analytics events and processed user or session metrics under its collection rules.
• Server records: requests reaching the logged server, which can include assets and APIs.
• CDN records: requests observed at the edge, including cache behavior and other response types.
• SearchBastion tracking: interactions observed by the installed tracking implementation.

Check timezone, date coverage, property, hostname and filters before comparing. Keep missing data separate from measured zero. Where the sources offer different granularity, acknowledge the limitation instead of forcing a match.

Worked example: one page load creates many requests

Imagine a browser opens one article that fetches images, styles and scripts. The server or CDN may record many requests, GA4 may record a session and page-view event, and button tracking may record nothing until a button is used. These are illustrative relationships, not fixed conversion ratios. Dividing the request total by sessions cannot identify how many people were bots.

Investigate an unexpected change

1. Select one URL group and comparable dates.
2. Check tag and consent changes near the start of the difference.
3. Separate page requests from static resources and API activity where possible.
4. Verify reporting freshness and synchronization failures.
5. Review unexplained bursts with application behavior and accepted outcomes.

Avoid false precision

Do not label a request-to-session difference as confirmed waste, human visitors or a fraud percentage. Known bots can be excluded by GA4, and legitimate visits can be absent because analytics code was blocked or consent was withheld. A useful report shows definitions and remaining uncertainty alongside each number.

Use Analytics Compare for an initial implementation check. Google's GA4 bot exclusion guidance explains one source of measurement differences.