BigQuery export is a GA4 feature that copies your raw, event-level analytics data every day into BigQuery, Google Cloud’s data warehouse. Instead of the summarised figures in GA4’s reports, you get every recorded event with its details, which you can keep for as long as you choose and query in full.
How BigQuery export works
You link a GA4 property to a Google Cloud project under Admin, then Product links, then BigQuery links. From then on, GA4 writes one table per day, holding each event as a row with its date, name, parameters, a pseudonymous user identifier, device, location and traffic source. An optional streaming export adds events within minutes, and BigQuery charges for it.
Three practical points catch people out. The export starts from the day you link it; GA4 does not backfill your history. Standard GA4 properties have a daily export limit, 1 million events a day at the time of writing (October 2026), which a typical small business site will not come near. And the data location is chosen when you create the link: Google Cloud offers a London region, europe-west2, alongside multi-region options such as EU and US, and the location cannot be changed later without starting a new dataset.
Why it matters
The export removes limits built into GA4’s interface. Explorations can apply sampling over large date ranges, reports can apply thresholding, which hides rows when Google judges the numbers could identify individuals, and explorations only reach back as far as your data retention setting, 14 months at most on a standard property. In BigQuery the raw events are all there, unsampled, for as long as you keep them.
It also lets you join analytics with data GA4 never sees. A UK online retailer can match GA4 transactions with refunds from its order system. A lead-generation business can join enquiries with the deals that closed in its CRM, and finally see which campaigns produce revenue rather than form fills.
Under UK GDPR, the copy in BigQuery is data you hold and are responsible for. The pseudonymous identifiers, and any User ID you send, count as personal data, so the project needs proper access controls, a retention period and an entry in your records of processing. Choosing the London region keeps the stored copy in the UK, which makes those records and your privacy notice simpler to write; it does not on its own make the processing lawful.
Common mistakes
- Waiting until the data is needed. With no backfill, the export only pays off if it has been running for months before the question comes up.
- Linking to a BigQuery sandbox project and forgetting about it. Sandbox tables expire after 60 days, so the history quietly disappears.
- Running unfiltered queries across every daily table. BigQuery charges by the amount of data a query scans, so select only the columns and dates you need.
- Accepting a US location by default when the business has told customers their data stays in the UK.
- Giving the whole team owner access to a dataset that holds personal data.
How to act on it
If you run GA4 and expect to want more than its reports, set up the export now, even if nobody will query it for a while. Create a Google Cloud project with billing enabled, choose the London region if UK data residency matters to you, link it from GA4 and select the daily export. Set a budget alert in Google Cloud so a careless query cannot run up a bill.
Then connect a reporting tool such as Looker Studio to a few saved queries rather than to the raw tables. Joining GA4 data with ad spend and CRM outcomes this way is what lets you judge channels on revenue, which is the basis of performance marketing.