Filtering on Average Position in Data Studio is one of those things that looks straightforward but consistently catches people out. Apply a standard filter expecting to see only queries ranking in positions 1–3, and you'll either get unexpected results or a "Re-aggregating metrics is not supported" error. Here's why — and how to fix it.
Why Normal Filters Don't Work for Average Position
Average Position is an aggregated metric. It represents the mean ranking of your website for a given query across multiple impressions over the selected time period.
Standard filters in Data Studio operate at the row level, before aggregation happens. This means the filter evaluates individual impression rows, not the aggregated average you see in the chart. It can't properly isolate the rows contributing to the average position values you want to filter on — hence the errors and wrong data.
The Solution: Data Blending
Data blending lets you re-aggregate metrics within Data Studio, bypassing the limitations of pre-aggregated fields. It's a workaround, but it's the right one for this problem.
Here's how to set it up:
Step 1: Open Manage Blends
In your Data Studio report, go to Resource → Manage Blends and create a new blended data source.
Step 2: Configure Table 1
- Data source: Google Search Console
- Dimension: Query
- Metrics: Average Position, Clicks, Impressions, CTR (add whatever you need)
Step 3: Add Table 2
- Data source: Google Search Console (same source)
- Dimension: Query
Step 4: Set the Join
Configure the join as Left outer, joining on Query.
Step 5: Apply the Filter
Add a filter to your blend — for example: Average Position is less than 4 to show only top-3 rankings.
Step 6: Save and Apply
Save the blended data source and replace the original Search Console source in your chart with the blended version. Your filter will now work correctly.
Dealing with Search Console data at scale? Our Search Console SEO Template handles this kind of filtering out of the box — 15 pages and 300+ charts with position filtering, branded vs non-branded segmentation, and period comparisons all pre-configured.