Sales and Support: (317) 300-4606

Is Your Backup Real? Three Questions to Ask Your IT Provider

Most backups look fine until the day you need one. Three plain questions to ask your IT provider so you know your company data can actually come back.

Short answer

A backup is only real if someone has restored from it recently and you know how long recovery takes. Ask your IT provider three questions: what exactly is backed up, when was a restore last tested, and how long would it take to get running again. Then ask where the backups live and who can delete them. Clear, specific answers mean you're in good shape. Vague ones mean it's time for a closer look.

Why ask about backups at all?

Because backups fail quietly. A job stops running after a password change, a new shared drive never gets added, or a storage account fills up. Nothing breaks, so nobody notices.

You find out on the worst possible day: a laptop is stolen, a file server dies, or someone deletes a folder the whole team depends on. The point of asking now is to find the gap while it's still a calm conversation.

You don't need to understand the technology to ask good questions. You need answers that are specific enough to check.

Isn't Microsoft 365 or Google Workspace already backed up?

Partly, and this is where many businesses get surprised. Microsoft and Google do an excellent job of keeping the service running. Their systems are built so that a failed server on their end doesn't cost you your email.

Recovering your own data is a different matter. If someone deletes a mailbox, overwrites a shared spreadsheet, or a departing employee's account is removed, the platform's built-in recovery options have time limits and gaps. Ransomware (malicious software that locks your files until you pay) can also sync bad copies right over good ones.

A separate backup of your Microsoft 365 or Google Workspace data closes that gap. It keeps independent copies of email, files and shared drives, on its own schedule, so you can restore something from weeks or months ago without depending on whatever the platform happens to retain.

Question 1: What exactly is backed up?

"Everything" is not an answer. Ask for a list. A good one names each system and how often it's copied.

Things that commonly get missed:

  • Shared drives and Teams or Google Chat files, not just individual mailboxes
  • Accounts for people who have left, which often hold the only copy of client history
  • Line-of-business apps, such as accounting, practice management or CRM systems, and whether the vendor or you are responsible for their backup
  • Files saved only to laptop desktops instead of cloud storage
  • Servers or network storage devices sitting in a closet at the office
  • Website content and DNS records (the settings that point your domain to your email and website)

Once you have the list, compare it with how your team actually works. If your finance lead keeps everything in one desktop folder, that matters more than any diagram.

Question 2: When was a restore last tested?

This is the question that separates a real backup from a hopeful one. A backup job that reports "success" tells you data was copied. It doesn't tell you the copy can be opened, or that the right people know how to bring it back.

Ask when someone last restored a real file, a mailbox and, if you have one, a server, and what happened. A good answer sounds like: "We restored a test mailbox and a shared folder last quarter, it took this long, and here's the note in your documentation."

If the answer is "it's never failed," treat that kindly but as a no. Regular, scheduled restore tests are the only proof that counts.

Question 3: How long would recovery take?

There are two parts to this, and both are worth asking in plain terms.

  • How much work could we lose? If backups run once a night, a failure at 4 PM could mean losing most of a day. Some data deserves more frequent copies than that.
  • How long until we're working again? Restoring one file takes minutes. Restoring every mailbox, or rebuilding a server, can take far longer. Downloading a large amount of data from the cloud is often the slow part.

The right answers depend on your business. A firm that bills by the hour feels a day of downtime differently than one that can work from paper for a while. Decide what's acceptable for your most important systems, then make sure the backup setup actually matches.

Where do the backups live, and who can delete them?

This is the bonus question, and it's become one of the most important.

Attackers who get into a network often go looking for the backups first, because deleting them makes the attack hurt more. A backup that sits on the same network, under the same admin account, can disappear in the same incident.

Things to ask about:

  • Separation. Is at least one copy stored somewhere separate from your main systems, ideally offsite or in a different cloud account?
  • Protection from deletion. Can backups be deleted or changed immediately, or are they locked for a set period? Some services call this immutable storage, meaning copies can't be altered once written.
  • Access. Who holds the admin login to the backup system? Is it protected with multi-factor sign-in (MFA)?
  • Alerts. Who gets notified when a backup fails, and what happens next?

What does a good answer look like?

Here's a quick way to read the responses you get back.

QuestionA reassuring answerWorth a follow-up
What's backed up?A written list by system, with how often each is copied"Everything important"
When was a restore tested?A recent date, what was restored and how long it took"It's never failed" or "We'd test it if needed"
How long would recovery take?An estimate for each key system, agreed with you"It depends" with no follow-through
Where do backups live?At least one separate, protected copy with MFA on accessSame network, shared admin account

None of these "follow-up" answers means your provider is doing a bad job. Often the backups are fine and simply undocumented. The goal is to turn assumptions into something written down.

Do we need a disaster recovery plan too?

Backups are the copies. A disaster recovery plan is the set of steps for using them when something big goes wrong. It answers questions like: who makes the call, which systems come back first, how staff keep working in the meantime, and who tells clients.

For a small or medium-sized business, this doesn't need to be a thick binder. A few pages that name the people, the order of recovery and the phone numbers to call will do more than a long document nobody has read. It also pairs naturally with an incident plan for security events, which we cover in what to do in the first 24 hours after a phishing click.

How does Simple IT think about backups?

We treat backup as part of security, not a separate chore. Our security baseline follows the CIS Controls v8 Implementation Group 1, a practical set of safeguards for smaller organizations that counts data recovery among the essentials. In practice, that means the same habits described above: a written list of what's covered, restores that are actually tried, and copies kept out of an attacker's reach.

If you're adjusting anything about your backup setup, test it on a small set of data before rolling it out to everyone. Changes to retention or storage settings can have side effects that are much easier to catch early.

For the bigger picture on what a small business security setup should include, see the security basics every 25-person company needs, or learn more about how managed IT works with us.

Where to start

Send your IT provider the three questions above and ask for written answers. If you'd like a second set of eyes on what comes back, or you aren't sure who to ask, book a 15-minute call and we'll walk through it with you.

Last reviewed November 17, 2026. This article is general information, not legal, insurance or compliance advice, and reading it doesn't make you a Simple IT client. Have your IT provider confirm any steps for your setup, and talk with your attorney, insurance broker or auditor about how anything here applies to your business. See our Terms of Use.