Almost every business we talk to says they have backups. Far fewer can say when they last restored something from them, how long a full recovery would take, or whether the backups would survive a ransomware attack.
That gap matters, because backups don't usually fail loudly. They fail quietly, for months, until the day you need them. Then you find out a critical folder was never included, the backup drive filled up last spring, or the attacker deleted your backups before encrypting everything else.
The only backup that counts is one you've successfully restored. A "job completed" email tells you software ran. It doesn't tell you the data is complete, uncorrupted, or recoverable in a reasonable time.
Seven ways backups fail
01Something important was never included
Backups are configured once, then the business changes. A new shared folder, a line-of-business database, or a laptop where someone keeps "the real" spreadsheet never gets added. Nobody notices until it's gone.
02Cloud data isn't actually backed up
Many businesses assume Microsoft 365 or Google Workspace backs up their email and files. These services keep things running and offer some short-term recovery, like a recycle bin and version history, but that isn't the same as an independent, long-term backup. Deleted items eventually age out, and a compromised account can delete or encrypt synced files.
03Jobs have been failing silently
The backup drive is full, a password changed, a service account expired, or alert emails go to someone who left the company. The job has been failing or skipping files for weeks.
04Ransomware got to the backups first
Modern ransomware attackers look for backups before they launch the attack. If your backups sit on a network drive or use the same admin credentials as everything else, they can be deleted or encrypted along with your production data.
05The data is there, but corrupted
Files can be backed up in a damaged state, especially open databases or virtual machines captured mid-write. Without verification, you won't know until you try to open them.
06The restore takes far too long
Downloading several terabytes from cloud storage over an office internet connection can take days. The data may be safe, but your business can't wait that long.
07Nobody knows how to restore
The person who set up the backups left, the encryption key or admin password isn't documented, or the restore steps exist only in someone's head. In a crisis, that's a serious problem.
Know your RPO and RTO
Before you can judge whether your backups are good enough, decide what "good enough" means. Two numbers define it:
- Recovery Point Objective (RPO): how much data you can afford to lose, measured in time. If you back up nightly, a failure late in the day could mean losing a full day's work. Is that acceptable?
- Recovery Time Objective (RTO): how long you can afford to be down while systems are restored. Four hours? One business day? A week?
These are business decisions, not IT decisions. Set them per system. Your accounting system and email may need tighter targets than an archive of old project files. Then test against them.
Follow the 3-2-1-1-0 rule
The classic 3-2-1 backup rule has been updated for the ransomware era:
| Rule | Why it matters | |
|---|---|---|
| 3 | Keep three copies of your data | Your live data plus two backups |
| 2 | On two different types of storage | One failure type can't take out everything |
| 1 | One copy off-site | Survives fire, flood or theft at the office |
| 1 | One copy immutable or offline | Can't be deleted or encrypted by attackers |
| 0 | Zero errors after verification | Proven by testing, not assumed |
"Immutable" means the backup can't be changed or deleted for a set period, even by someone with admin credentials. Many backup products and cloud storage services now offer it, and it's one of the most important protections against ransomware.
A practical restore-testing schedule
Testing doesn't have to be a huge project. Build it up in layers:
01Daily: review backup reports
Make sure someone actually reads the success and failure alerts, and that they go to a shared mailbox or ticketing system rather than one person's inbox. Investigate every failure and every "completed with warnings."
02Monthly: restore a few files
Pick a handful of files at random from different locations, including email and cloud storage, and restore them to a separate location. Open them to confirm they're intact. It takes about 15 minutes.
03Quarterly: restore a full system
Restore a complete server or virtual machine into an isolated test environment and confirm it boots and the application works. Time how long it takes and compare it to your RTO.
04Annually: run a disaster recovery exercise
Walk through a realistic scenario ("ransomware encrypted the main server and its local backups on Monday morning") with the people who'd be involved. Restore from your off-site or immutable copy. Confirm you can find every password, key and phone number you'd need, and update the plan with what you learn.
Also test after big changes. New servers, new software, office moves and changes to your backup product are all good reasons to run an extra test.
Keep a simple test log
Write down every test. A log shows trends (are restores getting slower?), proves due diligence to auditors and cyber insurers, and gives the next person a record of what works. A few columns are enough:
| Date | What was restored | From | Result | Time taken | Notes |
|---|---|---|---|---|---|
| Monthly | 5 files from shared drive and 2 emails | Cloud backup | Pass | 12 min | All files opened |
| Quarterly | Accounting server (VM) | Immutable copy | Pass | 3 hr 40 min | Within 8-hour RTO |
| Quarterly | Design team share | Local appliance | Fail | n/a | Folder not in backup set; added |
Example entries for illustration. A failed test that finds a problem is a success: you found it before it mattered.
Five-question self-check
If you can't answer "yes" to all five with confidence, it's time to test:
Can you say yes?
0 of 5The bottom line
Backups are insurance you hope you'll never use. But unlike most insurance, you can check whether the policy will actually pay out, and you should, regularly. A little time spent on restore testing now is the difference between a bad afternoon and a business-threatening outage later.
Not sure your backups would hold up? We can run a restore test with you, review your setup against the 3-2-1-1-0 rule, and document a recovery plan your whole team can follow.