A client sent me a screenshot: “GA4 shows 1,100 purchases. BigQuery shows 920. What’s wrong?”
At first glance, everything matched: date ranges, filters, parameters. But the numbers were still different.
Even though GA4 powers both the UI and BigQuery — they often feel like two different systems.
Here’s why:
GA4 UI shows processed, aggregated data. It applies:
• deduplication
• rounding
• sampling (in some cases)
• parameter limits
• UI-specific filters
BigQuery, on the other hand, gives you raw events. That means:
• You may see duplicate events (e.g., page refreshes)
• Delayed events might not show in your timeframe
• GA4 UI’s auto-cleaning rules don’t apply
So no, this isn’t a bug — it’s two different views of the same system. And if you compare them without context, the numbers will never align.
The client began doubting the data. So, which source to trust?
Answer: context decides.
• GA4 UI = quick insights, trend checks, dashboards
• BigQuery = precision, full-funnel analysis, performance reporting
• Business decisions made only from GA4 UI may miss key events
• BigQuery might overcount if filters aren’t precise
Without understanding this, you could kill a valid campaign — or overspend — just because “GA4 said so.”
Here’s how I handle it:
• Use GA4 UI for hypotheses, trends, and executive reporting
• Use BigQuery for final numbers, ROAS, and audits
• Align filters: event_name, event_timestamp,
• Clean tech events in BQ
• Review attribution models used in GA4 based on the raw data
GA4 and BigQuery don’t need to match — they need to make sense. If you can’t explain the gap, that’s the real problem.
Want to get all my top Linkedin content? I regularly upload it to one Notion doc.
Go here to download it for FREE.
