Key Takeaways
- Default method: Looker Studio's native GA4 connector—it's free, takes just five minutes to set up, and is sufficient for most SME dashboards.
- Its limitation: GA4 API quotas (tokens per hour and per property) that cause errors in reports that are resource-intensive or receive a high volume of views.
- Two saves: Data Extract, which caches lightweight tables, and BigQuery, which eliminates quotas and sampling at the cost of a little SQL.
- Discrepancies in figures: almost always due to thresholds, the comparison period, mismatched metrics (sessions vs. users), or a forgotten filter.
Connect GA4 to Looker Studio It takes five minutes. Getting a report that remains reliable and displays correctly six months later requires a bit more effort. This article reviews the three ways to export GA4 data to Looker Studio, the setup steps for each, their limitations, and solutions to the most common issues.
Method 1: The native GA4 connector
This is the default method and the only one needed for most websites. Steps:
- In Looker Studio, create a data source and select the "Google Analytics" connector.
- Select the account, then the GA4 property (there are no longer any views, unlike in Universal Analytics).
- Validate: Looker Studio loads all available dimensions and metrics, including custom dimensions declared in GA4.
- Create a report or apply a template; see The Best Looker Studio Templates for GA4.
The connector queries the GA4 Data API with every page view. It therefore adheres to GA4’s rules: confidentiality thresholds that mask rows that are too small, a 14-month limit for detailed data, and API quotas. A standard report (traffic by source, pages, conversions) easily stays within these limits.
The Issue of Quotas, Explained
Every standard GA4 property has a token budget: approximately 200,000 per day and 40,000 per hour, with a limit of 10 concurrent requests. Each chart consumes tokens—more so if it covers a long period, includes many dimensions, or displays unaggregated data. A report with 30 charts opened by ten people on Monday morning can exhaust the hourly quota; the charts will then display “quota exceeded” or remain blank.
| Factor | Effect on consumption | Parade |
|---|---|---|
| Number of charts per page | Each chart is a query | Group them together; limit to 10 or 15 per page |
| Long-Term Period and Comparison | Cost Doubling | Default period of 28 days; comparison available upon request |
| High-cardinality dimensions (page, query, city) | Resource-intensive queries | Filter, limit rows, prioritize extraction |
| Number of readers | Multiply everything | Data extraction, PDF scheduling |
| Data Freshness | Request on Every Startup | Set the freshness to 12 hours in the source |
Method 2: Data Extraction (Data Extract)
The "Data Extraction" connector copies a subset of GA4 data into a table cached by Looker Studio (limited to approximately 100 MB per extraction), which is automatically refreshed daily, weekly, or monthly. You select the dimensions, metrics, and a time period, and then connect the charts to this extract rather than to the live source.
- Benefits: No more quota errors, instant display, and the ability to retain data for longer than 14 months if the data is aggregated into another table.
- Limitations: No data for the day; scope fixed at the time of creation; size capped; multiple extracts must be created for different perspectives of analysis.
- Typical use: a monthly management report, viewed by many people, covering 10 to 20 consistent metrics.
Method 3: BigQuery
The native export from GA4 to BigQuery delivers raw events, without API quotas or sampling. Looker Studio then connects to an aggregated table built using SQL. This is the robust solution for large volumes, long time series, and joins with other data. How it works and the associated costs are detailed in BigQuery and GA4. Note: Connecting Looker Studio directly to event tables results in a high number of queries; always use aggregated and scheduled tables instead.
| Criterion | Native connector | Extraction | BigQuery |
|---|---|---|---|
| Setup | 5 minutes | 30 minutes | 1 to 3 days, SQL required |
| Cost | 0 € | 0 € | 0 to 30 € per month |
| Freshness | Near-real time | Daily at best | Daily or via feed |
| API Quotas | Yes | No to reading | No |
| History dating back more than 14 months | No | Partial | Unlimited after activation |
| Thresholds and Sampling | Yes | Yes, inherited from the API | No |
| Profile | All | Shared and Stable Reports | High-traffic, e-commerce, analyst |
Discrepancies in the numbers between GA4 and Looker Studio
This is the first question that comes up after setup: “The report doesn’t show the same numbers as GA4.” In almost all cases, one of the following causes is at play:
- Confidentiality thresholds: GA4 hides rows that are too small when Google signals are enabled; Looker Studio inherits this hiding behavior, but not at the same level of aggregation.
- Mismatched metrics: Compare "active users" in GA4 with "total users" in Looker Studio; the definition of sessions and users must be the same.
- Date and time zone: A different default period or an incorrectly set property time zone causes the totals to be off.
- Missing filters: A filter at the report or page level excludes a source without the reader noticing.
- Yesterday's data is still being processed: GA4 finalizes its figures within 24 to 48 hours; a report viewed in the morning displays preliminary data.
To validate a report, always compare a simple metric (sessions over 7 full days, unfiltered) between the standard GA4 report and Looker Studio. If they match, any discrepancies observed elsewhere are due to the report’s structure, not the connection. The guide How to Read a GA4 Report helps define the right benchmarks, and the article on the GA4 KPIs specifies which metrics to display first.
Take It Further: Integrating GA4 with Search Console and Google Ads
Looker Studio also offers native connectors for Search Console and Google Ads. Combining all three in a report is done using the “data blend” feature, with a common key such as date or page. The method, its limitations, and alternatives that don’t require custom code are described in Connect GA4 and Search Console in a dashboard.
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
Why does my Looker Studio report show "quota exceeded"?
Because the native connector queries the GA4 API with every view, and the property has a limited token budget per hour and per day. Too many charts, long comparison periods, detailed dimensions, or a large number of viewers will exhaust this budget. Reduce the number of charts, set the refresh interval to 12 hours, or switch to data extraction.
Do you need BigQuery to connect GA4 to Looker Studio?
No. The native connector is sufficient for most small and medium-sized business websites. BigQuery becomes useful when traffic exceeds several million events per month, when historical data beyond 14 months is important, or when you need to cross-reference GA4 with a CRM or cost data. It requires SQL skills and a small monthly Google Cloud bill.
Can the numbers in Looker Studio be the same as those in GA4?
Yes, if we compare the same metrics over the same time period, without any filters, and once the data has been finalized—that is, 48 hours after the end of the period. The discrepancies stem from different definitions, confidentiality thresholds applied at different levels of aggregation, or filters applied in the report and subsequently forgotten.
Can you store more than 14 months of GA4 data in Looker Studio?
Not with the native connector, which tracks data retention. There are two options: data extraction, by compiling monthly extracts into a spreadsheet or table, or the BigQuery export, which retains everything from the time it is enabled. In both cases, no data is retrievable for the period prior to implementation.