How to combine GA4 properties into one report
By the Analistable team · Updated · 6 min read
GA4 roll-up properties combine properties natively but are available only with Google Analytics 360. With standard GA4, export the same report from each property, add a Property column, and stack the exports into one table — or blend them in Looker Studio. Sessions, conversions and revenue can be added up across properties; users can't, because one person may visit several sites.
Part of our guide: How to join spreadsheets on a common column
Options
| Option | Requires | Good for |
|---|---|---|
| Roll-up property | Google Analytics 360 | Ongoing combined reporting inside GA4 |
| Looker Studio | Free | Dashboards across properties |
| Stack exports in a sheet | Free | One-off or monthly analysis |
| BigQuery export | Free export, BigQuery costs | Large-scale, event-level analysis |
Stacking exports in a spreadsheet
- In each property, open the same report (for example Reports → Acquisition → Traffic acquisition) with the same date range, dimensions and metrics.
- Download each as CSV.
- Add a column called Property to each file, or keep the file names meaningful.
- Stack them: Power Query From Folder,
VSTACK, or the CSV combiner which adds a source column for you. - Pivot by Property and channel.
| Channel | shop.example | blog.example | app.example | Total |
|---|---|---|---|---|
| Organic Search | 18400 | 42100 | 2300 | 62800 |
| Paid Search | 9200 | 0 | 1800 | 11000 |
| Direct | 6100 | 3900 | 7400 | 17400 |
Which metrics can be added?
- Additive: sessions, events, key events (conversions), revenue, page views.
- Not additive: total users, active users, new users — the same person can be counted in several properties. Report them per property, or label any total as an upper bound.
- Ratios (engagement rate, conversion rate): recalculate from the summed components rather than averaging the percentages.
GA4's CSV downloads start with comment lines (beginning with #) above the header. Remove them, or skip them in Power Query with Remove Top Rows, before stacking.
Roll-up properties (Analytics 360)
A roll-up property collects data from up to 200 source properties, all of which must also be 360 properties. If your organisation has 360, it's the most convenient option because combined reports, audiences and explorations work as normal.
For questions that span properties and other data — for example GA4 traffic next to Search Console clicks — see combining Google Analytics and Search Console data.
Recalculating ratios correctly
Don't average the conversion rates of three properties. Add up the parts and divide again:
Combined conversion rate = SUM(Key events) / SUM(Sessions)If one property has 40,000 sessions at 1% and another 2,000 at 10%, the simple average is 5.5%, but the combined rate is (400 + 200) / 42,000 ≈ 1.4%.
Stacking with formulas in Microsoft 365 or Google Sheets
Paste each property's export onto its own tab with identical columns (Channel, Sessions, Key events, Revenue). In Excel 365, one formula stacks them and adds the property name:
=VSTACK(
HSTACK(IF(SEQUENCE(ROWS(Shop!A2:A20)), "shop.example"), Shop!A2:D20),
HSTACK(IF(SEQUENCE(ROWS(Blog!A2:A20)), "blog.example"), Blog!A2:D20)
)IF(SEQUENCE(n), "name") simply repeats the label n times so it lines up with the rows. In Google Sheets, stack ranges with curly braces and semicolons, ={Shop!A2:D20; Blog!A2:D20}, and add a Property column on each tab first. Blank rows at the bottom of the ranges come through as zeros or blanks, so filter them out before pivoting. More stacking methods are in combining data from multiple sheets.
Power Query: one folder, one file per property
- Save each property's CSV in one folder, named after the property (for example shop.example.csv).
- Data → Get Data → From File → From Folder, then Transform Data.
- Keep the Name column: it becomes the Property column after you remove the .csv extension with Replace Values.
- In the sample-file query, remove GA4's comment lines with Remove Top Rows and promote headers, so every file is cleaned the same way.
- Load, then build a pivot table with Property as columns and Channel as rows.
Next month, overwrite the files with fresh exports and click Refresh All. The full technique is in Power Query: combine files from a folder.
Why users don't add up: an example
| Property | Total users |
|---|---|
| shop.example | 21000 |
| blog.example | 48000 |
| Sum of properties | 69000 |
If 9,000 people read the blog and then bought from the shop, they're counted in both rows. The real number of different people is lower than 69,000, but neither GA4 export tells you by how much: user identifiers aren't shared between properties. Label the sum as an upper bound, or answer the question at event level in BigQuery, where you can count distinct user_pseudo_id values across both exports — those IDs only match if both sites share the same cookie, which separate domains usually don't.
Checking the combined report
- For each property, compare its column total in the pivot with the total row of its own GA4 export.
- Check the grand total equals the sum of the property totals for additive metrics only.
- Make sure every property used the same date range and the same time zone setting; a property set to another time zone shifts days at the edges.
- Recalculate any rate from the summed columns and confirm no percentage was averaged.
BigQuery: union the event tables
With the free BigQuery export enabled on each property, every property writes its own events_* tables. A UNION ALL combines them for any metric you can compute from events:
SELECT 'shop' AS property, event_date, COUNT(*) AS purchases
FROM `my-project.analytics_111111111.events_*`
WHERE event_name = 'purchase' AND _TABLE_SUFFIX BETWEEN '20260901' AND '20260930'
GROUP BY 1, 2
UNION ALL
SELECT 'blog', event_date, COUNT(*)
FROM `my-project.analytics_222222222.events_*`
WHERE event_name = 'purchase' AND _TABLE_SUFFIX BETWEEN '20260901' AND '20260930'
GROUP BY 1, 2;Replace the project and dataset names with your own. Event-level counts from BigQuery may differ slightly from the GA4 interface, which can apply thresholds and modelled data.
Troubleshooting
| Symptom | Cause | Fix |
|---|---|---|
| Columns don't line up after stacking | Reports exported with different dimensions or metric order | Save the report configuration once and reuse it in every property |
| Revenue totals look wrong | Properties report in different currencies | Convert to one currency before summing |
| Channel names differ | Custom channel groups in one property | Use the default channel group, or map names first |
| Numbers read as text | Thousands separators or the # comment lines in the CSV | Remove comment lines and set the data type in Power Query |
Frequently asked questions
- Can I combine two GA4 properties without Analytics 360?
- Not inside GA4. Use Looker Studio to blend them, or export the same report from each and stack the files in a spreadsheet.
- Can I add up users from different GA4 properties?
- No. Users aren't deduplicated across properties, so a sum overcounts people who visit more than one site.
- How many properties can a roll-up property include?
- Up to 200 source properties, all of which must be Analytics 360 properties.
- Can Looker Studio combine GA4 properties?
- Yes. Add each property as a data source and either show them side by side in one chart per property, or blend them on a shared dimension such as Date or Channel. Check blended totals against each property, because blends join rather than stack.
- How often should I refresh a stacked report?
- Monthly for most teams: export each property for the closed month on the same day, replace the files and refresh. Avoid including the current, incomplete month.
- Why don't my combined sessions match the sum shown in each property?
- Usually different date ranges, filters or report types between exports, or rows lost when stacking. Compare each property's total in the stacked table with its own export.