06.1 Making things fast: warehouse and cubes
Normally Analytics asks your database every time a report or dashboard runs. That's always current, but with millions of rows, or many people looking at once, it gets slow and loads your database.
Analytics’s Data Warehouse solves this in two steps:
Your database → warehouse table (a copy) → cube (a summary) → dataset → dashboard
- Warehouse table: Analytics copies a table once into its own fast storage. After that, nothing reads your database. See 06.2.
- Cube: Analytics rolls the copy up to the level your dashboards need, e.g. sales by day, region and product instead of every single order. Often hundreds of times smaller. See 06.3.
- A dataset that reads the cube is then used by reports and dashboards as usual (03.2, choose A warehouse cube).
When to use it
| Use a cube when… | Stay live when… |
|---|---|
| The table is big (millions of rows) | The table is small |
| Many people open the same dashboards | Only a few people use it |
| Numbers from this morning or last hour are fine | You need the last minute’s data |
| You don’t want reports to load your database | The database is built for this load |
A small table doesn’t need a cube. Analytics tells you when a cube wouldn’t be worth building.
Where to find it
Left menu Warehouse (or the Datawarehouse tile). Three tabs:
- Overview: how many cubes and tables, what’s up to date, a Build button per cube
- Cubes: build and manage cubes
- Tables: the copies cubes read from
Cube statuses:
- Up to date
- Older data: its table was reloaded since the last build. It still answers, with the older numbers. Build it again.
- Not built yet: saved but never filled. Click Build.
Who can see it: by default, company admins and super admins.
Next: 06.2 Warehouse tables
Last updated 12 October 2026