pbPassingBI
/
Formatting & distribution intermediate 5 min

Report performance

Why a report is slow, and the changes that actually help.

What you'll be able to do
  • Diagnose a slow report
  • Reduce query load
  • Use extracts and aggregation appropriately

Why reports get slow

Every component runs its own query. A page with fifteen charts is fifteen queries, and the page is only as fast as the slowest one.

The usual causes, in rough order:

  • Too many components on one page
  • A large or unaggregated source
  • Blends, which query each source separately
  • Custom SQL doing heavy work per chart
  • Community connectors with slow APIs

Extract data

Add → Extract data creates a static, aggregated snapshot of a source, up to 100 MB, on a refresh schedule.

This is often the single biggest improvement available. Instead of querying a large BigQuery table live, you query a small pre-aggregated extract. The cost is that data is only as fresh as the last extract.

Aggregate upstream

The general principle: do the heavy work before Looker Studio sees the data.

A summary table in BigQuery, or a Sheet holding aggregated rather than raw rows, makes every chart faster at once. A report over a million raw rows will never be fast, however it is built.

Practical reductions

  1. 1Split dense pages — five focused pages beat one page of twenty charts
  2. 2Remove charts nobody looks at
  3. 3Replace blends with an upstream join where possible
  4. 4Lengthen data freshness so results are cached longer
  5. 5Limit default date ranges — 28 days rather than all time

Check the real cause

Before optimising, load the report and watch which components resolve last. It is usually one or two, and they are usually a blend or a community connector — not the thing you assumed.

Key points
  • Every component is its own query — dense pages are slow by construction
  • Extract data is often the single biggest speed improvement
  • Aggregate upstream; a report over a million raw rows will never be fast
Check yourself