Every small IT team "has backups." Almost none has proof of a restore.
An untested backup is a theory. The quarterly-restore-test idea is right, but most teams skip it because "quarterly restore test" sounds like a project. Make it a 15-minute habit instead.
The 15-minute version
- Pick one random file from your oldest client backup set (2 min).
- Restore it to a scratch folder on a machine that is NOT the source server (8 min).
- Open it. Confirm it's readable and current. (2 min)
- Log three numbers: date, restore time in minutes, pass/fail. (3 min)
That's the whole test. Run it monthly, rotating clients. In a year you've proven every client at least once.
Why the log matters more than the test
When a client asks "are our backups good?", you open a one-page log with twelve dated entries. That one page has closed more renewals for small MSPs than any slide deck — it converts "we think so" into "here's the evidence."
The failure modes the test catches (and which full DR drills often miss):
- Backup job succeeded, but the archive is corrupt
- The restore needs a credential nobody has anymore
- The restore works but takes 6 hours (surprise: your RTO is 6 hours)
- Storage filled up last week and nobody noticed
Done-for-you version — printable restore-test log + quarterly drill cards:
- Free incident quick-start: https://hive80lab.gumroad.com/l/ops-starter-kit-free
- Ops Starter Kit (runbooks + restore-test log): https://hive80lab.gumroad.com/l/ops-starter-kit
- Full ops notes library (free): https://hive80-lab.github.io/ops-notes/
Hive80 Lab — practical ops, automation and security resources for small teams.
