2026-08-27

Looker Studio GA4 Numbers Don't Match GA4? Here's the Exact Audit

You pull up your Looker Studio dashboard, then open the GA4 interface to double-check a number for a client call — and the totals don't match. Sessions are off by a few percent, or conversions are wildly different, or one card shows zero when GA4 clearly shows activity. You didn't change anything. The connector just doesn't agree with the source anymore.

Why this happens

These are ranked by how often they're actually the culprit, based on what typically breaks first.

1. Date range mismatch. Your Looker Studio page-level date range control, a chart-level custom date range override, or a comparison date range is set differently than what you're looking at in GA4. This is the single most common cause and the easiest to miss because the date picker looks "right" at a glance.
2. Different default channel grouping or attribution model. GA4's UI reports (especially Acquisition and Advertising snapshot) can use different attribution settings than what the GA4 connector pulls into Looker Studio by default. If GA4's UI is set to "Data-driven" attribution and your Looker Studio report is pulling event-scoped or session-scoped metrics without that model applied, conversion counts will diverge.
3. Sampling or thresholding on the GA4 side. High-cardinality reports or properties with Google Signals enabled can trigger data thresholding (small counts get withheld to protect user privacy) in the GA4 UI but not identically in the API response Looker Studio uses, or vice versa. Look for the thresholding icon in GA4 UI reports as a tell.
4. Blended data joining GA4 with another source incorrectly. If the scorecard or table is built from a Blend (formerly "Blend Data") that joins GA4 with Google Ads, a CRM export, or a second GA4 stream, a mismatched join key or join type (LEFT vs INNER) can silently drop or duplicate rows, inflating or deflating totals.
5. A filter applied in Looker Studio but not in GA4 (or the reverse). Filters live at three levels in Looker Studio — data source, chart, and page/report — and it's easy to forget one is still active from a previous edit. Similarly, a GA4 UI comparison or segment can be applied there but not replicated in the report.
6. Metric compatibility conflicts inside GA4's data model. GA4 metrics belong to different scopes (user, session, event) and some metric/dimension combinations are simply incompatible, causing Looker Studio to silently return null or a partial value instead of throwing a visible error.
7. Stale cache or a paused/disconnected data source. Looker Studio caches data source results. If someone toggled "Owner's Credentials" data access, or the underlying GA4 property/view was swapped, cached numbers can persist until a manual refresh.

Fix, step by step

1. Open both reports side by side and confirm the exact date range in the Looker Studio Date Range control (top right of the report) matches the GA4 UI's date range, including any comparison period.
2. Check every chart for a chart-level "Date Range Dimension" override — right-click the chart, choose Edit, and inspect the Data tab for a custom range that ignores the page control.
3. In GA4, go to Admin > Attribution settings and note the reporting attribution model. Then check whether your Looker Studio table uses the connector's default (event-based) metrics or a custom calculated field — mismatched attribution models explain most conversion-count gaps.
4. In the GA4 UI, look for the thresholding icon (a small triangle/circle) next to any metric — this indicates data was withheld from that report due to low user counts or Google Signals thresholds.
5. If the chart uses a Blend, open Resource > Manage blends, and verify the join key (usually Date, and a dimension like Page path or Campaign) and the join configuration (LEFT OUTER vs INNER) — an INNER join drops any row that doesn't exist in both sources.
6. Click into the data source itself and check for filters under Resource > Manage added data sources > [your GA4 source] > Edit. Then check chart-level filters (Filter section in the chart Data panel) and page-level filters (top of the page in view mode, if enabled).
7. Confirm every dimension/metric pairing on the chart is marked compatible — Looker Studio shows a warning icon on incompatible combinations in the chart editor.
8. Force a refresh: click the three-dot menu on the report and select Refresh data, or edit and re-save the data source connection to clear the cache.

Copy-paste audit checklist

GA4 vs LOOKER STUDIO MISMATCH AUDIT
[ ] Date range in Looker Studio page control === date range in GA4 UI
[ ] No chart-level custom date range overriding the page control
[ ] Comparison period (if any) matches on both sides
[ ] Attribution model in GA4 Admin > Attribution settings noted
[ ] Conversion metric scope (event-scoped vs session-scoped) checked
[ ] Thresholding icon checked in GA4 UI for the same metric/date range
[ ] Google Signals status confirmed (Admin > Data Settings > Data Collection)
[ ] Blend join key confirmed (Resource > Manage blends)
[ ] Blend join type confirmed: LEFT OUTER vs INNER
[ ] Data source-level filters reviewed (Manage added data sources > Edit)
[ ] Chart-level filters reviewed (chart Data panel > Filter)
[ ] Page-level filters reviewed (report view mode, top of page)
[ ] Dimension/metric compatibility warning icons checked on each chart
[ ] Data source cache refreshed (three-dot menu > Refresh data)
[ ] Owner's Credentials / data access permissions unchanged

How to verify it worked

Pick one metric (e.g., Sessions) and one dimension (e.g., Date) with no filters or blends applied. Build a bare-bones table in Looker Studio with just that pairing, set the exact same date range as GA4's Reports > Engagement > Overview, and compare the totals. If they match, add your filters, blends, and calculated fields back one at a time until the number breaks — that's your root cause.

Once your Looker Studio dashboard is trustworthy again, the harder question is what to do with the numbers. Next Action Analyst takes your GA4 or campaign CSV exports and turns them straight into a prioritized list of next actions, so you're not just reconciling dashboards — you're acting on them.