Loading tutorials…
Loading tutorials…
The native GA4 connector looks like three clicks — but the choices you make in those clicks lock in sampling, scope, and refresh patterns for the life of the report. Here's the right way to wire it up.
Who this is forAnyone reporting on GA4 data through Looker Studio. About 60% of DIY setups select the wrong property type or skip the sampling check, ending up with reports that disagree with GA4 by 10-15% for months.
What you'll need
Step 1
Looker Studio offers a "Google Analytics" connector that handles both UA and GA4. Inside it, you pick GA4 explicitly. Don't pick the deprecated UA path.
In Looker Studio, click Create → Data source. In the connector gallery, search "Google Analytics."
Click the Google-authored connector (blue G icon, says 'By Google'). Authorize access if prompted.
On the account picker screen, you'll see two tabs at the top: "GA4" and "Universal Analytics." UA is end-of-life since July 2024 — pick GA4 unless you have an archived UA property you specifically need.
Drill: GA4 Account → Property → Data Stream. Pick the production web stream (usually named after your domain). If you have multiple streams for staging vs. production, pick production carefully.
Click Connect (top-right). You'll see the GA4 field schema — every dimension and metric, around 200+ fields.
Step 2
In the field schema view (right after Connect), rename, hide, or aggregate fields once at the source level — not in every chart you build.
Look at the field list. Each row has Name, Type, Aggregation, and Description columns.
Find any dimension you'll use frequently with a confusing name. For example, "Session source / medium" is the standard channel field — rename it to "Channel" if your team uses that vocabulary.
Find any metric defaulting to SUM that should be AVERAGE (Engagement rate, Bounce rate, Conversion rate). Click the aggregation cell and switch.
Hide fields you'll never use to reduce the schema noise — click the eye icon on rows like "Audience name" or technical fields.
Add calculated fields here once (not per-report) — e.g., a custom 'Cost per session' that requires blending Google Ads cost data isn't done here, but a simple ratio of Conversions / Sessions can be (see tutorial 6).
Step 3
Pick Owner's credentials for shared marketing dashboards. Refresh defaults to 12 hours — leave it unless you have a real need for faster.
After saving the field config, click Edit Connection (or back into the data source from Home). Find the Credentials section.
For most marketing dashboards, set to Owner's credentials. This means everyone who views your report sees data through your access — you don't need to grant every viewer GA4 access.
For internal BI where access control matters at the row level, use Viewer's credentials. Each viewer needs GA4 viewer access on the property.
Refresh frequency: GA4 sources default to 12 hours. Lower to 1 hour only if you actually need it — faster refresh consumes more API quota and rarely changes the decisions made from the dashboard.
Save the data source. Name it clearly: "GA4 — Acme.com (Production)". Future-you will appreciate the explicit naming when there are five sources to pick from.
Step 4
Create a throwaway report, drop in 3 charts (Users, Sessions, Top sources), and confirm data loads without errors before building real dashboards.
From the data source, click Create Report (top-right). Looker Studio creates a new report with your GA4 source pre-attached.
Add a Date range control (Insert → Date range). Set default to Last 28 days.
Add a Scorecard with Metric = Total users. The number should appear within 5 seconds.
Add a Table with Dimension = Session source / medium, Metric = Sessions. The top row should be a recognizable channel (google / organic, direct / none, google / cpc).
If any chart shows an error ("Configuration incomplete," "Failed to fetch data," etc.), the connection is broken. Most common cause: GA4 viewer access was revoked on the underlying property.
Step 5
Compare Looker Studio numbers to the same metrics in GA4 Reports for the same date range. Tolerate 1-3% drift; investigate anything beyond.
Open GA4 in a separate tab → Reports → Acquisition → User acquisition. Set date range to Last 28 days (matching your Looker Studio report).
Note the Total users count in GA4.
Switch to Looker Studio. The Total users scorecard should be within 1-3% of GA4. Anything beyond suggests sampling, threshold filtering, or a wrong property selection.
Repeat with Sessions. Then with one specific channel row (e.g., google / organic Sessions).
Document the reconciliation in a Methodology page or a note in your team's shared doc. If numbers ever drift later, this baseline tells you what changed.
Step 6
If you see the yellow shield icon on any chart, that chart is sampled. Reduce date range or aggregate higher to remove sampling.
Sampling happens when GA4 estimates results from a subset of data instead of all data. It's normal for date ranges over 60 days or queries with many dimensions.
Look for the yellow shield icon at the top-right of a chart. Hover for sampling rate (e.g., 'Based on 38% of sessions').
For sampled charts: reduce the date range (Last 28 days rarely samples; Last 90 days often does), drop a dimension, or use the GA4 native UI for exact-match reporting on small slices.
For mission-critical exact numbers (revenue reconciliation, board reports), connect Looker Studio to BigQuery export of GA4 instead of the GA4 connector. Unsampled, but adds a per-query BigQuery cost.
Document any sampled charts on the report itself — add a text note like "Dimensions exceed 10K combinations; sampling applies to this chart." Honest dashboards stay trusted.
Step 7
Final check: confirm credentials are set as intended, confirm the data source name is clear, and confirm no embedded sources exist in the report.
Open Resource → Manage added data sources. You should see one entry, named clearly (e.g., "GA4 — Acme.com"), with credentials matching your intent.
If you see two GA4 entries — one named and one with a generic name — you accidentally created an embedded source while building. Delete the embedded one and reconnect charts to the named source.
Click Share (top-right). For client dashboards, use email-based sharing only. For internal team dashboards, "Anyone with the link" within your Workspace domain is fine.
Send the share link to one stakeholder first. Confirm they can open it without errors. Only then broadcast to the team.
Common mistakes
Connecting to the wrong GA4 property
What goes wrong: Most clients have a Test or Demo GA4 property in their account list alongside Production. Pick the wrong one and your dashboard reports data from a property with 2% of real traffic. The team makes decisions on a phantom dataset for weeks before someone catches it.
How to avoid: Verify property selection by checking the user count against GA4 UI within the first 10 minutes of connecting. A 90% discrepancy means you picked the wrong property.
Using Viewer's credentials when you mean Owner's
What goes wrong: Every new stakeholder you share the dashboard with gets a permission error and pings you for help. You spend 20 minutes per viewer granting them GA4 access. Compounds badly with team turnover.
How to avoid: Set the GA4 data source to Owner's credentials. One person (the owner) needs GA4 access; everyone else just needs Looker Studio share access.
Ignoring the sampling indicator
What goes wrong: You report 1.8M sessions to leadership. The actual unsampled number is 2.4M. Annual revenue projections based on this dashboard are 25% too low. Budgets get cut. Real money lost on a phantom shortage.
How to avoid: Hover every chart with a shield icon. If sampling rate is below 50%, that chart is unreliable. Reduce date range, drop a dimension, or switch to BigQuery export.
Hard-coding date ranges in chart settings
What goes wrong: You set 'Last 28 days' inside every individual chart instead of using a page-level Date range control. Now when the team wants to look at Q1, they can't filter — each chart has its own date setting they have to find and edit.
How to avoid: Add ONE Date range control per page. Set chart-level dates to Auto (inherits from control). Now the dashboard is filterable in one click.
Not renaming the data source clearly
What goes wrong: Six months in, you have data sources named "Data Studio Sample," "Untitled Data Source," "Google Analytics" — and no idea which belongs to which client. Building a new dashboard becomes a 30-minute archaeology project.
How to avoid: Rename every data source on creation: "[Tool] — [Property/Account Name]" (e.g., "GA4 — Acme.com Production"). Apply retroactively to existing sources in 5 minutes.
Recap
Done — what's next
How to set up Looker Studio and build your first report
Read the next tutorial
Hand it off
GA4 → Looker Studio is foundational. A bad connection makes every downstream report wrong. If you'd rather have a specialist set this up correctly and audit the reconciliation quarterly, that's typically $80-200 for setup + $50-100/month for ongoing audits.
See specialist rates
Three usual causes: (1) the date range in Looker Studio includes today, which is incomplete; (2) sampling kicked in on the chart (look for the yellow shield); (3) you connected to a sub-property or filtered view instead of the main data stream. Reconcile with Yesterday or Last 7 days excluding today as your test baseline.
Most marketing teams: the native GA4 connector is fine. Switch to BigQuery export when you need: (1) unsampled data, (2) custom event-level joins, (3) data retention beyond GA4's 14-month limit, or (4) cross-property aggregations. BigQuery adds cost (~$5/month/property for typical traffic) and SQL skill requirements.
Default is 12 hours. You can lower to 1 hour for free, or 15 minutes with Looker Studio Pro. GA4 itself updates data on a 24-48 hour delay for some dimensions (especially attribution-related), so faster Looker Studio refresh doesn't always show fresher numbers.
Yes — add multiple GA4 data sources (one per property), then build charts that pull from each. For cross-property aggregation in a single chart, you'll need to either use BigQuery to merge upstream or use Looker Studio's data blending (see tutorial 5).
Looker Studio
Looker Studio is free and surprisingly powerful — but the first report is where most owners stall. This is the walk-through that gets you from blank canvas to a shareable dashboard without making the rookie structure mistakes you'll regret later.
Looker Studio
Connecting Google Ads to Looker Studio is two clicks. Connecting it *correctly* — with the right scope, the right conversions, and a layout that survives a board meeting — is what this tutorial does.
Looker Studio
Data blending is the killer Looker Studio feature — combining GA4 + Google Ads in one chart, joining a Sheet to GA4, building cross-source ROAS calculations. It's also where most DIY users hit a wall. Here's the working model.
Google Analytics 4
GA4 isn't hard to install — it's hard to install *correctly* so the data is actually usable six months from now. This is the walkthrough that prevents the rebuild most owners do at month four.