
Business Continuity & Backup Guide
A backup strategy is the plan for getting the business working again after ransomware, theft, hardware failure or a simple mistake, fast enough that the outage never becomes a crisis. This guide covers the 3-2-1 rule, the gap in Microsoft 365 and Google Workspace, recovery objectives with worked examples, and a restore-test schedule you can adopt as is.
1. What is the 3-2-1 backup rule?
The 3-2-1 rule means keeping three copies of your data, on two different types of storage, with one copy off site. Each number removes a different way of losing everything at once.
Example: an office running its accounting database on a local server keeps the live database on the server (copy 1), hourly backups on a NAS in the comms cupboard (copy 2), and a nightly immutable copy in cloud storage in another city (copy 3). Many providers now extend the rule to 3-2-1-1-0: one copy offline or immutable, and zero errors in the last restore test.
2. Does cloud sync count as a backup?
No. OneDrive, Google Drive and Dropbox sync every change to every device within seconds, including deletions and files encrypted by ransomware.
Sync keeps files available. A backup keeps them recoverable, by holding copies from before the problem started, somewhere the problem cannot reach. Version history helps with single files, but it is limited in time and an attacker with admin access can delete it.
3. Does Microsoft back up my Microsoft 365 data?
Not in the way most people assume: Microsoft keeps the service running, but recovering deleted or damaged data is your responsibility.
- Microsoft ensures the infrastructure works (uptime).
- You are responsible for the data (retention and recovery).
4. What are RPO and RTO?
RPO (recovery point objective) is how much recent data you can afford to lose. RTO (recovery time objective) is how long a system can be down before the damage is serious.
Set both per system, not once for the whole business, because the right target depends on what each system costs you per hour. Tighter targets cost more: a 15-minute RPO needs replication, while a 24-hour RPO needs only a nightly backup job.
| System | Example RPO | Example RTO | What that requires |
|---|---|---|---|
| Accounting database | 1 hour | 4 hours | Hourly backups to a local NAS, a nightly off-site copy and a written restore runbook |
| 24 hours | 2 hours for one mailbox | Daily cloud-to-cloud backup with single-item restore | |
| Shared files | 24 hours | Minutes for a file, 1 day for a whole library | Daily backup plus version history |
| On-premises server | 4 hours | 8 hours | Image backups that can boot as a virtual machine while the server is repaired |
| Booking or point-of-sale system | 15 minutes | 1 hour | Replication or continuous data protection to a standby system |
Worked example: the accounting server fails at 15:00 and the last backup ran at 14:00, inside its one-hour RPO, so the finance team re-captures one hour of entries. The restore takes three hours, inside its four-hour RTO, so they are working again by 18:00. If either number had been missed, the backup design, not the staff, needs fixing.
5. Why is a restore test the only proof a backup works?
A restore test is the only proof because backup software reports success when it copies data, not when that data can actually be restored.
Corrupted backup files, expired credentials and missing encryption keys all show up at restore time. Testing on a schedule moves that discovery to a quiet Tuesday instead of the day you need it.
| How often | Test | Pass when |
|---|---|---|
| Daily | Read the backup job report and rerun any failed job | Every job succeeded or was rerun the same day |
| Monthly | Restore a random file and a random email | Both open intact, and the time taken is recorded |
| Quarterly | Restore a whole folder or mailbox, and boot one server backup as a virtual machine | Done within the RTO, with data inside the RPO |
| Yearly | Full recovery exercise of a critical system from the off-site copy | The system owner signs off and the runbook is updated |
| After any big change | Restore test of the changed or new system | The new system is restorable before it is relied on |
6. What does POPIA expect from backups?
POPIA requires reasonable technical measures to prevent the loss of personal information, and a tested backup is one of the plainest ways to show you took them.
Backups hold the same personal information as the live systems, so they need the same protection: encrypt them, restrict who can restore them, and delete old copies when the retention period ends.
Need this designed and run for you? See backup and disaster recovery, or keep the office online through power cuts with the load shedding IT continuity guide.