Back to Resources
Template

The SME Backup Health Checklist

A healthy backup practice answers "yes" to twenty questions across five areas: coverage, schedule & retention, verification, security, and recovery.

0 of 20 checked

A healthy backup practice answers "yes" to twenty questions across five areas: coverage, schedule & retention, verification, security, and recovery. Score below fifteen and your business is one bad day away from a loss it can't recover from. The full checklist is below, written so a business owner can answer it without asking IT to translate.

How to use this checklist

Twenty yes/no questions, about fifteen minutes. Answer honestly: "I think so" counts as no, because the whole point of backup health is that thinking so isn't knowing so. Count your yes answers and check the scoring guide at the end. Items marked auto-checkable are things monitoring software can answer continuously so a human doesn't have to remember to.

Coverage: is everything that matters actually backed up?

  • 1. Do you have a written list of every system your business depends on: databases, shared files, email, accounting, HR/payroll?
  • 2. Is every system on that list included in a backup job today (not "was, when we set it up")?
  • 3. When a new system or server was added in the last year, did someone add it to the backups within a week?
  • 4. Are cloud services (email, shared drives, SaaS data) covered by a backup you control, not just the provider's own recycle bin?

Schedule & retention: often enough, kept long enough?

  • 5. If your newest backup were lost work since yesterday, would the business accept that? (If not, is the backup running more often than daily?)
  • 6. Do you keep older copies (a week ago, a month ago), not just the latest one that may already contain the mistake or the malware?
  • 7. Is your retention period a decision someone made (written down), rather than whatever the software defaulted to?
  • 8. Would you notice within one day if a scheduled backup stopped running? auto-checkable

Verification: would you know if it silently failed?

  • 9. Does anything check that the backup file actually exists and is non-empty after each run, beyond the job saying "completed"? auto-checkable
  • 10. Is backup integrity verified (checksums, readable format), so silent corruption would be caught? auto-checkable
  • 11. Would a backup that suddenly shrank to half its normal size trigger an alert? (Shrinking backups are a classic ransomware tell.) auto-checkable
  • 12. Do backup failures reach a human (an alert someone owns) rather than a log nobody reads?

Security: can one bad event take backup and original together?

  • 13. Is at least one copy stored offsite or in cloud storage, somewhere a fire, theft or flood at your office can't reach?
  • 14. Is at least one copy protected from being changed or deleted (immutability / object lock), so ransomware with admin credentials still can't destroy it? auto-checkable
  • 15. Are backup credentials different from everyday admin credentials, and limited to the people who need them?
  • 16. Are backups containing personal or sensitive data encrypted at rest?

Recovery: has anyone ever actually restored?

  • 17. In the last six months, has anyone performed a real restore test (database mounted, VM booted, or files opened) and written down the result? auto-checkable
  • 18. Do you know how long a full restore of your most important system takes, as a measured number rather than a guess?
  • 19. Could someone other than "the one person who knows" perform the restore, following written steps?
  • 20. After the last restore test, did anyone check the restored data was complete, not just that the restore finished? auto-checkable

Scoring

What your count means

Yes answersWhere you standDo first
18–20Genuinely healthy (rare). Your remaining risk is drift: staying at 18+ next quarter.Automate the checks so the score can't quietly decay.
15–17Sound foundation with real gaps, usually in verification or restore testing.Schedule one real restore test this month (Q17) and fix alerting (Q8, Q12).
10–14You have backups, but you could not prove they work, and one of the classic silent failures may already be present.Q13 and Q14 first (offsite + immutable copy), then Q9–Q11.
Below 10One bad day (ransomware, fire, or a single dead disk) is an existential risk right now.Today: confirm one restorable copy of your most critical data exists somewhere offsite. Everything else is second.

Passing the checklist once is a snapshot, not proof; every answer decays. The job that runs today stops next month; the new database misses the backup list. The eight auto-checkable items are the ones Backup Governor re-answers continuously, so the score you show your boss stays true. For what "proof" looks like, read How to Prove Your Backups Actually Work.

Want to see these best practices in action?

Start for free and experience proper backup monitoring firsthand.

No credit card required
Cancel anytime