SOP: Spreadsheet vs database for operational data

SOP: Spreadsheet vs database for operational data

Run this when a spreadsheet starts running a real process.

Stage 1 — Stress-test the sheet (owner: data owner)
☐ Row count, and how many people edit it weekly.
☐ Does anything downstream read it programmatically?
☐ Has anyone broken a formula or overwritten a column this month?

Stage 2 — Choose
☐ Spreadsheet: instant, flexible, everyone knows it. Use for one-off analysis, drafts, and data under ~1k rows with one editor. Cost: no validation, no audit trail, fragile to a stray paste.
☐ Database (or a structured app on top): enforced types, access control, history. Use when multiple people write, other systems read, or a bad edit is costly. Cost: setup and a maintainer.

Stage 3 — If migrating (owner: data owner)
☐ Freeze the sheet, snapshot it, define the schema first.
☐ Keep the sheet read-only as the historical record.
☐ Done-when: writes happen in one place with validation.

Trigger: a sheet became infrastructure. Done-when: home and validation decided.

Save this. Run it every time.