What a blend does
A blend joins up to five data sources on one or more join keys — shared dimensions present in both.
The classic use is combining ad spend from one platform with conversions from another, matched on date and campaign.
Configuring it
Select two components and choose Blend data, or Resource → Manage blends.
For each source you pick the dimensions, metrics and a date range. Then set the join condition: the key fields, and a join type — left outer, inner, full outer, cross.
Left outer from your primary source is the usual choice, keeping every row from the main table.
The join key must actually match
If one source stores campaign names as Summer_Sale and the other as summer sale, nothing matches and every blended metric comes back null.
Create a calculated field on one side to normalise the key before blending — lowercase, trimmed, punctuation removed.
Date granularity matters too. Blending daily on one side against monthly on the other produces mostly nulls.
Limitations
- Maximum five sources
- Join keys must be dimensions, not metrics
- Blends can be slow, since each source is queried separately
- Some connector-specific features are unavailable on blended data
If you find yourself building an elaborate blend, consider whether the join belongs upstream — in BigQuery, or in the Sheet — where it will be faster and easier to maintain.
Blend versus one source
Blending is for data that genuinely lives in different places. Two tabs of the same spreadsheet should be joined in the spreadsheet, not blended in Looker Studio.