Key Takeaways
- What it is: BigQuery is Google Cloud's data warehouse; GA4 can export every raw event to it for free, either daily or in real time.
- What it offers: No more sampling, no 14-month retention limit, integration with CRM or Google Ads, and SQL access to all data.
- How much it costs: Exporting is free; BigQuery charges for storage (about $0.02/GB/month) and queries (about $6/TB read), which amounts to 0 to 30 € per month for most small and medium-sized businesses.
- For whom: High-traffic sites, e-commerce sites, and those that want historical data; not necessary for a showcase site that views its reports in GA4.
BigQuery comes up in almost every discussion about Google Analytics 4. The integration is free, can be activated in just a few clicks, and Google promotes it as the natural next step after GA4. However, most small and medium-sized businesses don’t need it. This article explains what the export feature does, what it will actually cost in 2026, and the situations in which it becomes essential.
What BigQuery Is, in Simple Terms
BigQuery is Google Cloud's data warehouse. It stores tables containing billions of rows and queries them using SQL in a matter of seconds, without the need to manage any servers. You pay based on usage: a small fee for storage and a small fee for the volume of data read by each query.
The connection to GA4 is direct: GA4 already stores each interaction as an event (page_view, click, purchase). The interface shows only aggregates. BigQuery receives the events themselves—one per line—along with all their parameters, the session ID, the source, the device, and, if you’ve configured it, the user ID. This is the raw data, unprocessed.
What the export feature actually adds to GA4
| GA4 Interface Limit | What BigQuery Changes |
|---|---|
| Retention of detailed data is limited to 2 or 14 months | Unlimited storage, as long as you pay for it |
| Sampling of explorations exceeding 10 million events | No sampling; the query reads all rows |
| Confidentiality thresholds that hide small lines of data | There are no thresholds, but you are fully responsible for compliance with the GDPR |
| Reports limited to the specified dimensions and metrics | Any combination is possible in SQL, including metrics that GA4 does not calculate |
| Data isolated from other systems | Integration with CRM, billing, Google Ads, Search Console, or a cost file |
| No retroactive correction | Ability to recalculate metrics retroactively (attribution, deduplication) |
The most valuable use for a business is data integration: aligning on-site behavior with actual revenue, margins, or signed quotes. This is something that neither GA4 nor Looker Studio can't do correctly on their own. See also the Pitfalls to Avoid in GA4, many of which stem precisely from thresholds and sampling.
Two Export Modes and Their Limitations
- Daily Export: One table per day (events_YYYYMMDD), delivered within a few hours after midnight. The limit is 1 million events per day for a standard property; beyond that, the daily export is suspended. GA4 360 increases this limit to 20 billion.
- Streaming export: Events are delivered within a few minutes to an intraday table. There is no limit on the number of events, but the streaming cost is approximately $0.05 per GB.
- Fresh daily: An intermediate variant exclusive to GA4 360, with faster delivery of the daily table.
Please note: The export does not include historical data. Nothing that occurred before activation can be retrieved. This is the main reason to activate the link early, even if you don’t use it right away: a company that activates BigQuery in 2026 will have a complete history by 2028—something GA4 alone will never provide.
How much will it cost in 2026?
| Position | Estimated Price | Order of magnitude for an SME |
|---|---|---|
| Export GA4 to BigQuery | Free | 0 € |
| Active Storage | About $0.02 per GB per month, with 10 GB included | A website with 100,000 sessions per month generates 1 to 3 GB per month: a few cents |
| On-Demand Queries | About $6 to $7 per TB read, 1 TB free per month | Common in franchising; €5 to €30 per month if a dashboard queries the raw data tables throughout the day |
| Streaming | About $0.05 per GB inserted | 1 to 5 € per month |
| Human Time | SQL, modeling, maintenance | The actual cost: 2 to 10 days to set up, then a few hours per month |
The classic pitfall isn’t the Google bill, but a Looker Studio report connected directly to event tables that reads hundreds of GB every time it’s opened. The solution is to create aggregated tables (one row per day and per source, for example) using scheduled queries, and to connect the dashboards to them. The bill then drops to just a few euros.
When an SME Really Needs BigQuery
- You have exceeded the thresholds: several million events per month, sampled explorations, lines masked by thresholds.
- Do you have a CRM or an ERP system?: You want to link a session to a signed quote, a delivered order, or a margin.
- You need the history: 3-year comparisons, seasonality, long-term cohorts.
- You are feeding data into another system: lead scoring, calculated advertising audiences, proprietary attribution model.
Conversely, a showcase website or a small online store that follows its GA4 KPIs in standard reports or in a custom report gains nothing from it, except for additional complexity and GDPR obligations. The transfer of raw data to a data warehouse must also be included in your processing record; this topic is discussed in detail in GA4 and GDPR.
Alternatives Before Moving to the Warehouse
- Looker Studio's native GA4 connector: sufficient for most dashboards, with its API quotas; see Export GA4 to Looker Studio.
- A ready-to-use management tool: A platform that integrates GA4, Search Console, and Google Ads without SQL, ideal when the goal is to track rather than explore.
- A one-time export: the GA4 API or the Google Sheets add-on to extract a few tables without setting up any infrastructure.
BigQuery addresses an analytics need; day-to-day monitoring (traffic, conversions, rankings, spending) works just fine without it. The right approach for an SME: enable data export early on to build up historical data, continue monitoring with simple tools, and only create queries when a specific question warrants it.
How GreenRed Can Help You
Instead of juggling multiple tools, GreenRed's GA4 Performance and Traffic Module brings these metrics together in a single dashboard, compares them over time, and shows you which actions to prioritize. You can try it for free—no credit card required—on the page Rates.
Frequently Asked Questions
Is exporting data from GA4 to BigQuery really free?
Yes, the export itself is free, even for a standard GA4 property. You’re billed for the use of BigQuery: storage beyond 10 GB and queries exceeding 1 TB read per month. For an SMB website, the monthly bill is usually between 0 and a few euros if the dashboards read aggregated tables.
Can we retrieve GA4 data from before it was activated?
No. The export begins as soon as the link is created and does not go back in time. Past data remains available in the GA4 interface, subject to its retention period of 2 or 14 months for detailed data. This is the main reason to enable the export early, even if you don’t use it right away.
Do you need to know how to code to use BigQuery with GA4?
You need to know SQL, and more specifically how to work with the nested fields specific to the GA4 schema (event_params, user_properties). AI assistants are very helpful for writing these queries, but validating the results requires an understanding of event structure. Without this in-house expertise, it makes more sense to use a service provider or a ready-to-use tool.
What is the difference between the BigQuery export and the GA4 API?
The API returns data that has already been aggregated by GA4, including its thresholds, sampling, and quotas. The BigQuery export delivers raw events, without aggregation or thresholds, and with no read quota other than the cost of the queries. The API is suitable for standard dashboards; BigQuery is the go-to choice for in-depth analysis, long-term historical data, and joins with other sources.