Because there is always one report somebody needs in a spreadsheet, and it is always produced by a person who has other work. It arrives late, or with last week's filter still applied.
A scheduled export runs a saved query on a timer and delivers the result as CSV to a destination, or makes it available through the API on the same schedule.
The schedule owns the query, so the definition cannot drift. Each run is logged with its row count, and a run that fails is visible rather than silently absent.
A commission export ran every Monday at 07:00 for eleven months. The two occasions its row count dropped sharply were both caught the same morning, because the run log showed it.
They are not an integration. If two systems need to agree continuously, the API and webhooks are the right tool; a nightly CSV is a report, and it is a day out of date by design.
Custom dashboards share the same saved queries, Books consumes financial extracts, and the API surface exposes identical data for anything that would rather pull than receive.
Email, a storage bucket, or fetched from the API. Destinations are configured per schedule.
It is logged and flagged. Silence is the failure mode this exists to remove.
Yes, any saved query, subject to the permissions of whoever owns the schedule.
Fourteen days, every module, no card. Or half an hour with someone who will run it on your own records and tell you where it does not help.