Linking GA4 to BigQuery is one of the most useful things a business can do with its analytics data. It's also the point at which a free tool starts sending a bill, which worries people more than it usually needs to.
The export from GA4.
GA4 offers two kinds of export, and you can turn on either or both.
- The daily export writes the previous day's events to a table named
events_YYYYMMDD. It costs nothing to send. - The streaming export makes the current day's data available within a few minutes, in an
events_intraday_YYYYMMDDtable that's deleted once the daily table is complete. Google charges $0.05 per GB for it, at the time of writing. It's best effort, may have gaps, and leaves out some attribution data for new users.
For most businesses, the daily export is enough. Streaming is worth paying for only when someone needs today's data today.
The limit on standard GA4.
Standard GA4 properties can export up to 1 million events a day in the daily export. Analytics 360 properties can export up to 20 billion. The streaming export has no volume limit.
If a standard property goes over the limit, Google pauses the daily export, and earlier days aren't reprocessed. A property far over it can be paused straight away. To stay under, exclude events you don't need, such as noisy scroll or timing events, or exclude data streams, in the export settings.
The export also needs a valid payment method on the Google Cloud project. If the export stops because payment fails, Google says the missed data can't be exported again.
BigQuery storage and queries.
Once the data is in BigQuery, Google Cloud charges for two things. These are the US prices at the time of writing. They vary by region, so check the page for yours.
- Storage. The first 10 GiB each month is free. Beyond that, Google's own example puts 100 GiB of active storage for half a month at $1.15. Tables not changed for 90 days move to long-term storage, which is about half the price.
- Queries. On-demand queries cost $6.25 per TiB of data scanned, and the first TiB each month is free. Each query is billed for at least 10 MB, and so is each table it reads.
How big the export grows depends on how many events the site sends, so check the dataset's size after the first month. Storage tends to grow slowly and predictably. Queries are where bills surprise people.
The sandbox.
The BigQuery sandbox lets you try BigQuery without a billing account. It has a lifetime limit of 10 GiB of storage, which isn't given back when you delete data, and 1 TiB of queries a month. Tables expire after 60 days, and streaming isn't supported.
That makes it fine for a trial, and poor for keeping history. Upgrade the project with billing before you rely on the data, and remove the 60-day expiry from the dataset.
How to keep queries cheap.
- Select only the columns you need. BigQuery charges for the columns a query reads, so
SELECT *on the events table is the costly habit. - Filter on
_TABLE_SUFFIXto read only the days you need. Google notes the filter must use constant values to limit what's scanned. - Don't count on
LIMIT. On a table that isn't clustered, it doesn't reduce the data read. - Preview tables instead of querying them. The preview tab is free.
- Check the estimate first. The console shows how much a query will process before you run it.
- Set a maximum bytes billed on queries, and custom daily query quotas on the project, so a mistake can't run up a large bill
- Build smaller tables for reporting. Summarise the raw events once a day, and point Data Studio dashboards at the summary rather than the raw export.
A dashboard connected straight to the raw export, re-running its queries for every viewer, is a common cause of an unexpected bill.
Marc Alexander sets up the export, cost controls and reporting tables as part of connecting GA4 to BigQuery. The GA4 mini audit checks whether the data is worth exporting first.
Sources.
- Google Analytics Help: BigQuery Export
- Google Analytics Help: Set up BigQuery Export
- Google Analytics Help: BigQuery Export schema
- Google Cloud: BigQuery pricing
- Google Cloud: Try BigQuery using the sandbox
- Google Cloud: Estimate and control costs
Checked on 31 October 2026.