I'm trying to get an overview of all page views based on session source using BigQuery, but I see very large differences compared to the user interface.
What I am wondering, does Google use last non-direct click for this metric (that explains the differences)? I know it uses it for calculating conversions but I just want to understand how it assigns the source and medium at the start of a session.
Hi @nv93 the differences you’re seeing between BigQuery and the Google Analytics 4 (GA4) UI likely come down to how GA4 assigns source and medium to a session.
Here’s the key thing to know:
Why This Matters in BigQuery
If you’re analyzing data at the event level instead of the session level, you might notice that some events have different session_source and session_medium compared to what you see in the GA4 interface.
💡 To get a more accurate view, make sure you're pulling these values from the session_start table rather than from each individual event.
A Simpler Alternative
If you’d rather avoid dealing with complex SQL logic, tools like Windsor.ai can help by extracting GA4 data with the correct attribution directly into Power BI, Looker Studio, or Google Sheets. This can help you:
✅ Reduce discrepancies
✅ Save time on data processing
✅ Ensure consistency across reports
Hope this helps!