Blend data sources in Looker Studio (up to five tables)

#01How a blend actually works

Google's "How blends work" documentation is required reading. In plain language:

1. You pick up to five tables (each from a data source).
2. For each table you choose dimensions and metrics to include.
3. Data Studio groups each table by its dimensions before joining.
4. A join configuration links tables with an operator and equality conditions.
5. Charts query the blend like a data source, but freshness and credentials still come from the underlying sources.

Because of step 3, if you join Ads and GA4 only on Date and leave Campaign out of one side, multiple campaigns collapse. If you join on Campaign Name with inconsistent naming, rows disappear in an inner join or null-out in an outer join.

#02Join operators for marketers

#03Step-by-step: Ads spend + GA4 key events by date

1. Add a Google Ads table with Date, Cost, Conversions (or your chosen conversion).
2. Add a GA4 table with Date, Key events, Sessions.
3. Join on Date = Date with left outer, Ads on the left if spend completeness matters most.
4. Include only Date as the dimension in each table for a daily scorecard (no Campaign yet).
5. In the chart, SUM Cost and SUM Key events. Create a chart-level Cost per key event.
6. Add a note: "Joined on date only. Not campaign-level attribution."

For campaign-level blends you need a shared key. Campaign name often differs between Ads and GA4 UTMs. Many teams join on a cleaned key in Sheets instead. See lesson 12.

#04Date ranges and filters inside the blend

Each table can have its own date range and filters applied before the join. Use that to exclude brand campaigns on the Ads side while leaving GA4 intact, or to limit GA4 to a country. Freshness for the blend is the minimum of member freshness.

#05Hide repeated join fields

Leave Hide repeated join fields on unless you are debugging mismatches. Otherwise you will see Date (Ads) and Date (GA4) cluttering the field list.

#06Inflated metrics checklist

When a blended ROAS looks "too good":

1. Check join operator (accidental cross join?).
2. Check key cardinality (one GA4 date-campaign to many Ads rows?).
3. Check whether you summed a metric that was already a daily total after a fan-out.
4. Compare each side's totals in separate non-blended charts for the same range.
5. Rebuild with date-only first, then add dimensions one at a time.

#07Worked example: Ledgerline

Ledgerline is a fictional accounts-payable software company. Details are illustrative.

Walkthrough

1. The first leadership blend used campaign name and quietly dropped Performance Max variants.
2. Ops compared blended Cost to a pure Ads scorecard and saw the gap.
3. They switched to date-level efficiency for the exec page and kept campaign joins on a mapped Sheet.
4. Chart-level CPA replaced a data-source field that would not work on the blend.
5. The report note stated the join grain in one sentence.

#08Aggregation-before-join worked example

Suppose Ads has two rows on 1 Nov for Campaign X and Y ($100 and $50). GA4 has one row for 1 Nov with 40 key events.

• If you join only on Date and include Campaign on the Ads side but not GA4, GA4's 40 may fan out across campaigns depending on how you configure dimensions — validate with control scorecards.
• Safer pattern for daily efficiency: both tables at Date grain only on the exec blend; campaign joins live on a separate diagnostic page with a mapping Sheet.

#09Mapping Sheet pattern for campaign joins

Blend Ads → map → GA4 on the clean key. Name mismatches become visible as nulls instead of silent drops in an inner join.

#10When not to blend

• You only need to show two scorecards side by side (no row alignment) — use two charts, not a blend.
• The join key is unreliable this month — fix tracking first.
• You are about to add a sixth table — remodel in BigQuery instead.

#11Common mistakes

• Joining on Display Campaign Name across platforms without a map.
• Cross join by accident.
• Expecting a blend to be reusable like a data source.
• Using data-source calculated fields on blend charts.
• Five-table blends that nobody can debug.

#12Try it yourself

Blend Ads and GA4 on Date only. Compare SUM(Cost) on the blend to SUM(Cost) on a pure Ads chart for last 28 days. If they differ by more than a rounding error, fix the blend before adding dimensions.

#13Quick quiz

1. What is the maximum number of tables in a blend?

Answer: Five.

2. Are join conditions only equality?

Answer: Yes. Data Studio does not support ≠, > or < join conditions.

3. Why might rows collapse before the join?

Answer: Each table is grouped by its selected dimensions; without a unique key, identical dimension sets merge.

#14What should you check before the next lesson?

• One working date-level blend with control scorecards.
• Join grain written on the page.
• Chart-level rate metric, not a broken data-source field.

Previous: Parameters, filters and controls in Looker Studio. Next: Themes, layouts and branding for Looker Studio reports. Or return to the Looker Studio for Marketers track. Related: GA4 Search Console Google Ads links, Google Ads for Marketers, Calculating ROI of GEO.

#15Common questions

How many data sources can I blend? Up to five tables in one blend. Blends are embedded in the report and cannot be reused as standalone data sources across reports, though copying the report copies the blend.

Which join types does Data Studio support? Inner, left outer, right outer, full outer and cross join. Join conditions support equality only.

Why are my blended metrics too high? Usually a join fan-out: one row on the left matched many on the right, or metrics were aggregated incorrectly after the join. Fix join keys and prefer aggregating in each table before joining.