How to test your business backup before you need it: a 30-minute restore check
Short answer
Restore 3 random items to a separate folder and time each one: a file, a shared folder and an email. If any is missing, stale, unreadable or needs a password nobody has, you've found a gap. CISA advises keeping offline, encrypted backups and testing them regularly, so repeat the check every 3 months.

Key takeaways
- A "backup completed" message proves a job ran, not that you can restore.
- Test three random items to a scratch folder and time each one.
- Check that your admin login can't also delete the backup.
- Repeat the small test every quarter and a big restore once a year.
Most offices find out whether their backup works on the worst day of the year. That's a bad time to learn the last good copy is 8 months old, or that nobody has the password to the backup console. This guide is a 30-minute check you can run on a quiet afternoon, with or without an IT person. It tells you what to fix before you need the backup. If an outside team handles your IT requests, here's what an outsourced helpdesk should and shouldn't cover.
What does a backup test actually prove?
A green "backup completed" email proves one thing: the software finished a job. It says nothing about whether you can get a file, a mailbox or a whole server back, how many minutes that takes, or who on staff knows how. A restore test answers all four.
CISA's #StopRansomware Guide makes the same point. Its first prevention practice is to maintain offline, encrypted backups of critical data and regularly test their availability and integrity in a disaster recovery scenario. The guide also notes that many ransomware variants hunt for backups they can reach and delete or encrypt them, which is why a copy sitting on the same network as your files doesn't count for much.
What should you know before you start?
Write down five answers first. If you can't answer one, that's your first finding, and you haven't spent a minute testing yet.
- What is backed up: file server, laptops, Microsoft 365 mail and OneDrive, line-of-business databases, phone system settings?
- Where do the copies live, and can the same login that runs your office also delete them?
- How often does it run, and what is the newest copy you could restore from right now?
- Who can log in to the backup console, and where is that password kept?
- How long can the business stay down before it hurts: a morning, a day, a week?
How do you run a 30-minute restore test?
Use a test restore to a separate location. Never restore over live files just to see what happens.
- 01
Pick three things to restore
One ordinary file from last week, one folder from a shared drive, and one email from a mailbox. Choose them at random, not the ones you know are fine.
- 02
Start a clock
Note when you ask for the restore and when the file opens. That gap is your real recovery time for small items.
- 03
Restore to a scratch folder
Send the data to a new location, not the original. Then open each item and confirm it's complete and readable.
- 04
Try one bigger restore
Pick a whole folder of several gigabytes, or one workstation if your backup covers them. Large restores are where slow links and size limits show up.
- 05
Write down what broke
Missing items, a slow step, a password nobody had, an old copy. Each one is a fix, with a name next to it.
What results should worry you?
| What you see | What it usually means | What we'd do |
|---|---|---|
| The newest copy is days or weeks old | A job is failing or was never scheduled for that system | Fix the schedule and set an alert on failed or missed jobs |
| A file restores but won't open | The copy was taken while the file was in use, or the data is damaged | Check the backup method for that app and test again |
| Only the person who set it up can log in | One person holds all the access | Store credentials in the company password manager and name a second admin |
| Restore works but takes far longer than you can be down | The data is large or the connection is slow | Back up the critical systems more often and plan their restore order |
| The backup can be deleted from a normal admin login | Ransomware that gets that login can erase your copies too | Add a copy that can't be changed or deleted, kept apart from the main network |
Why a backup on the same network isn't enough
If your backup drive is mapped as a letter on the file server, then ransomware that encrypts the server can usually reach the backup too. This is the reason CISA asks for offline copies and why we ask one blunt question on every first call: if the person who runs your IT lost their password tomorrow, could a stranger delete your backups?
The idea behind the old 3-2-1 habit still helps here: more than one copy, on more than one kind of storage, with at least one copy away from the office. The newer addition is that one of those copies can't be altered or deleted for a set period, even by an admin. You don't need to memorize the numbers. You need an honest answer to "what survives if our main login is stolen?"
How often should you test?
Run the small test above every 3 months, and the bigger restore at least once a year. Run one right away after any big change: a new server, a move to Microsoft 365, a new office, or a change of IT provider. A test after an incident doesn't count as a plan.
Fifteen minutes with 3 random files catches most silent failures, such as a drive that filled up or a job that stopped when someone changed a password. Once a year, do the big one, because only a big restore answers questions about speed and size. Moving project files to a new platform is one of those moments, so test a restore before you start. Our comparison of Egnyte and SharePoint for construction project files covers what to check in that move.
What if the restore fails?
Don't panic, and don't start changing settings at random. A failed test on a quiet Tuesday is the cheapest way to find a problem. Write down the exact error, the time, and which of the 3 items failed, then work through the list below in order.
- Check the job history for that system. Look for failures in the last 30 days, not just last night's run.
- Confirm the backup account hasn't expired or had its password changed. Plenty of silent failures start this way.
- Check free space at the destination. A full disk stops new copies while old ones sit there looking fine.
- Ask the vendor or your IT provider for the last successful restore point for that system, and write the date down.
- Fix the cause, run the same test again that day, and keep both results in one file.
Who should run the test?
Not the person who set the backup up. They already know where everything is, so they'll find every file on the first try. Ask an office manager or a second admin to follow the steps above with the written instructions only. If they get stuck, so would you in a real emergency at 2 a.m. with the owner on the phone.
We follow the same rule for our own clients. Our on-call technician handles urgent issues after hours, and the restore steps are written down so that any of our engineers can run them. We run 24/7 threat detection and response for computers and Microsoft 365 email, powered by Huntress, and recovery is the other half of that job. Detection tells you something is wrong. A backup you've tested gets you back to work. If you're also tightening sign-in security, read does MFA stop phishing in Microsoft 365.
Where does our team fit in?
We run backup and recovery for New York businesses from our offices in Astoria and Williston Park, and we've been doing managed IT in New York since 2001. If you'd rather not run the test yourself, we'll do it with you and hand back a one-page list of what to fix. The cloud backup service page lists what we back up and how, and the business continuity and disaster recovery page covers planning the restore order. If a phished account is what worries you more, read what to do in the first hour after a Microsoft 365 account gets phished.
Questions people ask
Is a cloud backup the same as a backup that survives ransomware?
No. A cloud copy that your normal admin login can delete is still exposed. Ask whether at least one copy is kept apart or can't be changed, and test that you can restore from it.
Does Microsoft 365 back up my email and OneDrive for me?
Microsoft runs the service, but its services agreement asks you to keep your own regular backups of your content. Treat deleted or encrypted data as your responsibility to recover.
How long should a restore take?
It depends on the size of the data and the connection. A single file should take minutes. A whole server can take hours or longer. The test gives you your own number, which is the only one that matters.
Can we test without disrupting staff?
Yes. Restore to a separate folder or test machine, so live files are never overwritten. Most of the small test can run while everyone works.
Sources



