GA4's BigQuery export copies raw, event-level data from a property into a Google BigQuery dataset you control. It is available to standard and 360 properties. Once the data is in BigQuery, you can query it with SQL, join it with CRM or billing data, and keep it for as long as you choose. It is the usual answer when GA4's interface hits sampling, thresholds or retention limits.
How the export works
You link a GA4 property to a Google Cloud project in GA4's admin. GA4 then writes a daily table per day of data (named events_ followed by the date) and, if you enable it, a streaming intraday table. Each row is one event with nested fields for parameters, user properties, device, geography and traffic source.
Standard properties have a documented daily event limit for the daily export, and streaming has different rules. Check Google's current documentation for the limits that apply to your property.
Costs and setup
The export itself does not have a GA4 fee, but BigQuery charges for storage and queries above its free tier, under your Google Cloud billing account. Partitioned tables, selecting only needed columns, and scheduled aggregate tables keep costs down.
The export starts from the day you link it. It does not backfill earlier data, so link early even if you have no immediate use for it.
SELECT
event_name,
COUNT(*) AS events
FROM `project.analytics_123456.events_*`
WHERE _TABLE_SUFFIX BETWEEN '20260901' AND '20260930'
GROUP BY event_name
ORDER BY events DESC;What you gain
Raw events are not sampled or thresholded and are not subject to GA4's data retention. You can build custom attribution, join product usage with marketing sources, and feed dashboards with exact numbers. Expect some differences from GA4's interface, which uses estimation and modeling in places.
Example
Joining traffic source to revenue
Imagine Ambercrest, a payroll SaaS, exports GA4 to BigQuery and stores the GA4 client ID on each trial account. A scheduled query joins first traffic source to paid conversions in the billing database. The team can now report organic-sourced revenue from its own data instead of estimating it in a spreadsheet.
Common mistakes
Linking late
The export is not retroactive. Every month without it is history you cannot query later.
Running SELECT * on the wildcard table
Scanning every column across every day is the fastest way to a surprising bill.
Questions
Usually close but not identical. GA4 reports use estimation for some counts, modeling with Consent Mode, and filters that raw data does not apply.
