Blog

Backups That Survive Ransomware: Offline, Unchangeable, and Tested

Every cyber-insurance application asks whether you back up your data. In our review of 13 public carrier applications (November 2026), all 13 asked. It is the easiest box on the form to tick, and it's also the one that matters most if the worst happens.

Ransomware is software that scrambles (encrypts) your files and demands payment for the key. When it hits, the backup is the difference between a bad week and a business-threatening month. That's why the carriers don't stop at "do you have backups?" The follow-up questions are where quotes are won and lost.

The follow-up questions carriers ask

In the same review:

  • 10 of 13 asked whether you keep a backup that is offline, air-gapped or otherwise isolated, meaning a copy that can't be reached from your everyday office network.
  • 9 of 13 asked whether you test restoring from backup.
  • 6 of 13 used the word "immutable," which just means the backup can't be changed or deleted during the period you've chosen to keep it.
  • 5 of 13 asked whether getting into the backup system itself requires MFA (a second step beyond a password).

Put those together and the carrier is really asking three things. Is there a copy the attacker can't reach? Can that copy be changed or wiped? And when did someone last prove it works?

Why "we back up" isn't enough anymore

Modern ransomware crews don't just encrypt the files in front of them. They look for the backups first. If the backup system uses the same logins as the office network, or sits on a drive that's always connected, it can be encrypted or deleted along with everything else. CISA's ransomware guidance makes the same point and recommends offline, encrypted backups that are tested regularly. You can read it at cisa.gov/stopransomware.

When I ran IT at a company of about 200 people, the hardest part of backups was never the software. It was the assumptions. People assumed the nightly job worked because nobody had complained. People assumed the cloud copy was separate because it was "in the cloud." And everyone assumed a restore would be quick, because nobody had timed one.

The sync-folder trap

The most common misunderstanding I see is treating a sync folder as a backup. OneDrive, Google Drive, SharePoint and Dropbox are excellent tools for sharing and working from anywhere. But a sync folder copies changes, and that includes bad changes. If ransomware encrypts the files on a laptop, the sync service can faithfully copy the encrypted versions up to the cloud.

These services do keep version history, and that can help. But version history isn't designed or managed as your recovery plan, and more than one carrier application asks about sync tools by name for this reason. A real backup is a separate product with its own retention rules and its own protected logins, set up so that a problem on the office side can't reach back and damage it.

What "isolated" and "immutable" look like in practice

You don't need to know the technology. You need to recognize the answers.

Isolated means the backup lives somewhere the everyday network and everyday logins can't touch. That might be a dedicated backup cloud service with separate credentials, a backup appliance on its own network segment, or, at the simplest, a drive that's disconnected and stored away. The test question: if someone stole the office administrator's password, could they delete the backups? If yes, it isn't isolated.

Immutable means the backup system itself refuses to change or delete a copy until its retention period runs out, even when an administrator asks. Many current backup products offer this as a setting. The test question: is it turned on, and for how many days?

MFA on the backup console means that even the people allowed to manage backups need a second factor to log in. That protects the keys to the last line of defense.

Testing: the date carriers want

A backup that's never been restored is a hope, not a plan. Nine of the 13 applications we read ask about restore testing, and more of them are asking how long a full recovery would take.

A restore test doesn't have to be dramatic. Pick something that matters, such as one shared drive, the accounting database, or a single server, and restore it to a safe place. Check that the files open. Write down how long it took. Do that on a schedule, and keep the dates.

That timing matters to the business as much as to the carrier. If restoring the accounting system takes three days, the CFO should know that now, not during an incident.

What to ask about cloud apps

Your company data doesn't all live on your own computers anymore. Email, files, CRM, accounting and payroll may all live in cloud services. Some of those services keep their own copies, and some don't keep what you'd need. Microsoft and Google, for example, protect their own infrastructure, but recovering from your own users or an attacker deleting data is largely your responsibility. Ask your IT provider which cloud services are backed up by a separate product, and which ones rely on the vendor's built-in retention.

What to do Monday morning

  1. Write down where the official backup lives, with the product name and whether it's a true backup or a sync folder.
  2. Ask the isolation question: could someone with the everyday office administrator password delete or change the backups? Get the answer in writing.
  3. Ask whether immutability is turned on, for how many days, and whether logging into the backup console requires MFA.
  4. Find the date of the last successful restore test. If the answer is "we think last year," schedule a restore of one important shared drive this week and record how long it took.
  5. List your cloud services (email, files, accounting, CRM) and mark which ones have a separate backup.

Shorehelm's readiness assessment includes these backup questions, worded the way the carriers ask them.

All essays