Connector types
| Type | Examples | Notes |
|---|---|---|
| Sheets, BigQuery, GA4, Ads, Search Console | Free, first-party, well supported | |
| Partner | Facebook Ads, Stripe, HubSpot | Usually paid, third-party built |
| Community | Open source | Free but unmaintained risk |
| File upload | CSV | Static; no refresh |
Start with Google connectors where you can. Partner connectors carry a monthly cost per source and can break when the upstream API changes.
Google Sheets
The most common starting point. Two things cause most problems:
Looker Studio expects one header row at the top with data immediately beneath. Title rows, merged cells or blank rows above the headers produce columns named Column 1, Column 2.
Use a named range where the sheet has anything else on it, rather than pointing at the whole tab.
Also tick Use first row as headers, and be aware that adding columns to the sheet later requires refreshing the fields in the data source before they appear.
BigQuery
The right choice at scale. You can point at a table, or write custom SQL in the connector.
Custom SQL is powerful but note that every chart on the page runs against it, so an expensive query multiplies. Where that matters, materialise a summary table in BigQuery and connect to that instead.
Billing is per query, so a report on a large table refreshing frequently has a real cost. Check the estimate before publishing widely.
GA4
The GA4 connector exposes GA's own dimensions and metrics rather than raw rows. Two things to expect:
- Sampling on large date ranges, which makes numbers approximate
- Data freshness of 24–48 hours for some metrics
Neither is a bug, but both surprise people comparing a Looker Studio figure to the GA interface.
Reusable versus embedded
A data source can be reusable — usable across many reports and editable independently — or embedded in one report.
Reusable is almost always right. Calculated fields and formatting defined once are then shared, rather than rebuilt per report.