Analytics & Attribution
GA4 Cleaned Up the Source Names. Your Trend Line Still Needs QA.
Updated: October 26, 2026
GA4’s cleaner Source Group and Source Platform reporting can reduce messy naming, but it can also create a visible trend break if your dashboards, filters and historical logic still use raw source values. Build a mapping table, compare totals across the cutover and document which dimension each KPI now uses.
What GA4 changed
Google’s 2026 release notes describe standardized reporting for third-party platforms. Disparate values such as variations of Instagram can be bundled into a cleaner Source Group, while Source Platform can identify a broader platform such as Meta Ads. This helps analysts compare channels without maintaining endless text filters.
Normalization is not the same as historical continuity. A cleaner label may change grouping behavior even when traffic volume does not change.
Separate four layers
| Layer | Question | Example |
|---|---|---|
| Raw collection | What arrived with the event? | UTM source and medium |
| Processed source | How did GA4 record the traffic source? | Session source / medium |
| Normalized group | How did Google consolidate variants? | Instagram Source Group |
| Platform | Which advertising ecosystem owns it? | Meta Ads Source Platform |
A dashboard that switches from raw source to a normalized group is changing its definition, not merely its label.
Run a cutover comparison
- Export the top raw source and medium combinations.
- Add Source Group and Source Platform beside them.
- Compare sessions, users, key events and revenue.
- Find rows that merge, split or remain unclassified.
- Review paid, organic and referral edge cases.
- Record the date the dashboard definition changed.
Use enough time to include weekly and campaign-cycle variation. A one-day check can miss sources that appear only during a specific promotion.
Watch dimension scope
GA4 exposes first-user and session-scoped traffic-source dimensions for different questions. Do not compare a first-user Source Group with a session-level campaign KPI and assume they describe the same acquisition event. Google’s documentation explains that the User acquisition report uses first-user dimensions while Traffic acquisition uses session-scoped dimensions.
This is exactly why MTC recommends a KPI definition record before polishing dashboards.
Keep raw UTMs governable
Normalization should not become permission to ignore tagging standards. Clean campaign IDs, sources and media remain essential for joins, exports, cost data and exception analysis. Keep a governed UTM dictionary and monitor new raw values that map unexpectedly or stay outside the intended group.
Update dependent assets together
- Looker Studio calculated fields and filters.
- BigQuery transformation logic.
- Executive scorecards and annotations.
- Paid-media reconciliations.
- Channel targets and forecasts.
- Alert thresholds based on historical baselines.
If only one layer changes, two teams can report different channel totals while both believe they use “Meta” or “Instagram.”
Frequently asked questions
Should Source Group replace source / medium?
Not universally. Use the normalized dimension for cleaner rollups and retain raw or processed source / medium for diagnosis, tagging QA and detailed reconciliation.
Will historical reports change?
Validate rather than assume. Compare the same dates and metrics across old and new definitions and document any difference.
What belongs in the reporting note?
The dimension name, scope, cutover date, mapping exceptions, dashboard owner and whether targets were rebaselined.