The 3-2-1 Backup Rule Explained (and How to Test Your Backups)
What the 3-2-1 backup rule means, how a small business or manufacturer can set it up affordably, and how to test backups so a restore works when you need it.
It's 7:40 on a Monday morning and the front office can't open the shared drive. File names have turned into strings of gibberish, and there's a text file on every desktop with instructions for paying in cryptocurrency. Someone says, "It's fine, we have backups." Then the IT manager logs into the backup server and finds it encrypted too, because it was joined to the same domain and used the same admin password as everything else.
I've been doing network and infrastructure work for 15 years, and that scene is the reason I bring up the 3-2-1 backup rule with every small business we talk to. Most companies do have backups. Far fewer have backups that would survive the kind of attack that's most likely to hit them, and fewer still have actually tried restoring from them.
This guide explains the 3-2-1 rule in plain terms, shows how a small business or manufacturer can set it up without a huge budget, and walks through how to test your backups so you know they'll work before you need them.
In this guide
- What the 3-2-1 backup rule actually means
- Why the 3-2-1 rule matters more for small businesses now
- What to back up (it's more than the file server)
- A practical 3-2-1 backup setup for a small business
- How to test your backups
- Common backup mistakes we find
- Frequently asked questions
- Not sure your backups would hold up?
What the 3-2-1 backup rule actually means
The rule is simple enough to write on a sticky note:
- 3 copies of your data: the live production copy plus two backups.
- 2 different types of storage, so a single failure (a bad disk array, a failed NAS) can't take out every copy.
- 1 copy kept offsite, away from your building.
The logic is that each layer protects against a different kind of disaster. Multiple copies protect against a deleted file or a corrupted database. Different media protect against hardware failure. The offsite copy protects against fire, flood, theft, or a burst pipe over the server closet.
The modern version: 3-2-1-1-0
The original rule was written before ransomware was a business model. Many backup vendors and security teams now talk about 3-2-1-1-0, which adds two things:
- 1 copy that is offline, air-gapped or immutable, meaning it can't be changed or deleted for a set period, even by an administrator account.
- 0 errors when you verify and test restores.
That extra "1" is the part most small businesses are missing. That "0" is the part almost everyone is missing.
Why the 3-2-1 backup rule matters more for small businesses now
Ransomware groups learned years ago that a company with good backups doesn't pay. So a common playbook is to get in, quietly find the backup system, and delete or encrypt it before launching the main attack. If your backup server sits on the same network, uses domain credentials and has its storage mapped as a regular share, an attacker with admin rights can usually reach it.
Small and mid-sized companies are attractive targets because they tend to have flat networks, shared admin passwords and nobody watching logs at 2 a.m. For a manufacturer, the cost isn't only lost files. It's idle machines, missed ship dates and customers who start looking at other suppliers.
Good backups also show up in other places:
- Cyber insurance. Insurers commonly ask whether you keep offline or immutable backups, how often you test them, and whether backup access is protected by MFA. Weak answers can affect your premium or coverage.
- Compliance. Frameworks like NIST 800-171 and HIPAA include expectations around protecting and recovering data. Your assessor or attorney can tell you exactly what applies to you.
- Everyday mistakes. Most restores I've done weren't for ransomware. They were for an overwritten spreadsheet, a deleted folder, or a server update that went sideways.
If you're working through the basics, backups are one item on a longer list. Our post on 10 steps to protect your company covers the rest of the foundation.
What to back up (it's more than the file server)
When we ask "what do you back up?", the answer is usually "the file server." That's a start. Here's a fuller checklist to walk through with your team.
Servers and core systems
- File servers and NAS devices
- ERP, MRP and accounting databases (back up the database properly, not just the folder it lives in)
- Domain controllers and Active Directory
- Line-of-business application servers
- Virtual machines, ideally as full image backups so you can restore the whole system, not just its files
Cloud and SaaS data
- Microsoft 365 or Google Workspace: email, OneDrive, SharePoint, Teams
- Any cloud apps that hold customer or financial data
This one surprises people. Microsoft and Google keep their platforms running, but protecting your data inside them is largely your responsibility. Their built-in retention and recycle bins only cover a limited window, and they won't help much if an attacker or a departing employee deletes data and nobody notices for months. A third-party backup for your cloud tenant is usually worth the modest per-user cost.
Shop-floor and OT systems
For manufacturers, this is the list that often gets forgotten:
- PLC programs and HMI projects
- CNC programs and machine parameters
- Configuration files for SCADA or monitoring software
- Recipes, setups and quality data stored on local PCs
Losing a PLC program can mean calling the integrator who wrote it years ago and hoping they still have a copy. Exporting these on a schedule and storing them with your other backups is cheap insurance. We touch on this in our guide to securing shop-floor networks without stopping production.
Network and security devices
- Firewall, switch and wireless controller configurations
- Endpoint security and MDM policy exports
Rebuilding a firewall from memory after a hardware failure takes hours. Restoring a saved config takes minutes.
A practical 3-2-1 backup setup for a small business
You don't need an enterprise budget to follow the rule. Here's a layout we commonly recommend for a company with somewhere between 20 and 200 employees. Your exact choices will depend on how much data you have and how fast you need to recover.
Copy 1: Production data
This is your live data on servers, PCs and cloud apps. It isn't a backup, but it counts as one of the three copies.
Copy 2: A local backup for fast restores
A dedicated backup appliance or a separate server with its own storage, kept on site. This is what you restore from day to day because it's fast.
Set it up so it's hard for an attacker to reach:
- Don't join it to your main Windows domain.
- Give it unique credentials that aren't used anywhere else, and protect the console with MFA.
- Put it on its own network segment and only allow the traffic it needs.
- Turn on immutability or retention locks if your backup software supports it.
Copy 3: An offsite, immutable copy
Replicate backups to a cloud storage target or a vendor's cloud with object lock or immutability turned on. That means even if someone gets the admin password, they can't delete those backups until the retention period ends.
Some businesses still rotate removable drives or tapes offsite. That can work, but it depends on a person remembering to swap the media every week, which is exactly the kind of task that quietly stops happening. If you go that route, put it on someone's calendar and check it.
Decide your recovery targets first
Before buying anything, answer two questions for each important system:
- How much data can we afford to lose? If the answer is "an hour of orders," nightly backups aren't enough. This is your recovery point objective (RPO).
- How long can we be down? If the ERP being offline for two days stops shipping, you need a plan that restores it in hours. This is your recovery time objective (RTO).
These answers decide how often you back up and what kind of restore capability you pay for. Without them, it's easy to overspend on systems that don't matter and underspend on the one that does.
How to test your backups
A backup you've never restored from is a hope, not a plan. I've seen backup jobs report "success" for months while quietly skipping a database that was locked, or backing up an empty folder after a server move. Testing is how you find those problems on a quiet Tuesday instead of during an outage.
Check the job reports every day
Someone should look at backup alerts daily, or at least get an email when a job fails or is skipped. A failed job that sits unnoticed for three weeks means three weeks of exposure.
Do a file-level restore every month
Pick a few random files from different shares and restore them to an alternate location. Open them. Make sure they aren't corrupted and are the version you expected.
Do a full system restore every quarter
This is the test that matters most. Pick one critical server, such as your ERP or file server, and restore the whole thing into an isolated test environment.
- Restore the server image to a sandbox network that can't talk to production.
- Boot it and confirm the operating system comes up cleanly.
- Log in to the application and check that recent data is there.
- Time the whole process from start to working system.
- Compare that time to the RTO you agreed on. If it took eight hours and the business needs four, you've found a gap.
Run a tabletop exercise once a year
Gather the IT lead, an owner or executive, and someone from operations. Walk through a scenario: "Ransomware hit overnight and the main server is encrypted. What do we do?" Talk through who decides what, who calls the insurer, where the backup credentials are stored, and which systems come back first. You'll usually find at least one assumption nobody had checked.
Write down what you found
Keep a simple log: date, what you restored, how long it took, whether it worked and what you fixed. That record is useful for your own planning, and it's the kind of evidence an insurer or assessor may ask to see.
Common backup mistakes we find
My partner Mukesh has spent 25 years in cybersecurity, and between the two of us, these are the problems we run into most often:
- Backups on the same domain and credentials as everything else. If one admin password unlocks both the servers and the backups, an attacker gets both.
- No offline or immutable copy. Everything is reachable over the network, which means everything can be encrypted over the network.
- Cloud data not backed up at all. Microsoft 365 is treated as if it backs itself up.
- Retention too short. If you only keep 14 days of backups and an attacker sat in your network for a month, every restore point may already include their foothold.
- Never tested. The jobs say "success," so everyone assumes the restore will work.
- One person knows how it works. If your backup knowledge lives in one person's head, write it down and store a copy where you can reach it during an outage.
- Shop-floor systems left out. PLC and CNC programs live only on the machines or on a laptop in the maintenance office.
If several of these sound familiar, you're not alone. Many smaller companies don't have time to keep up with backup testing in addition to everything else. That's one of the reasons some move to a managed provider. We compare the options honestly in MSP vs. hiring a full-time IT person.
Frequently asked questions
Is the 3-2-1 backup rule still relevant with cloud backups?
Yes. Cloud storage can be your offsite copy, but it shouldn't be your only backup. A local copy restores much faster, and having copies on different platforms protects you if one provider has an outage or your account is compromised.
Does Microsoft 365 back up my data?
Microsoft keeps its service running and offers some retention features, but it doesn't provide the kind of point-in-time backup most businesses assume. Deleted items are only recoverable for a limited window. A third-party Microsoft 365 backup gives you longer retention and an independent copy.
How often should a small business test its backups?
A good baseline is a quick file restore every month and a full system restore of at least one critical server every quarter. Check backup job alerts daily. If you change servers or applications, test again soon after the change.
What is an immutable backup?
An immutable backup is one that can't be modified or deleted for a set retention period, even by someone with admin access. It's one of the most effective defenses against ransomware that targets backups, because the attacker can't destroy it with stolen credentials.
How long should we keep backups?
It depends on your industry, contracts and any regulations that apply to you, so confirm the specifics with your attorney or compliance advisor. From a security standpoint, many businesses keep at least 30 to 90 days of restore points, plus monthly copies kept longer, so they can go back to a point before an attacker got in.
Not sure your backups would hold up?
If you read the checklist above and weren't sure how your company would answer some of it, that's normal. Most businesses only find out how good their backups are when they need them. It's a lot less stressful to find out now.
Our free cybersecurity assessment includes a look at your backup setup: what's covered, what isn't, whether you have an offline or immutable copy, and whether a restore is realistic within the time your business can afford to be down. You'll get a plain-language summary of what we found and what we'd fix first. Reach out here to set it up.
Related reading
- Cybersecurity for Roanoke Small Businesses: 10 Steps to Protect Your Company
- How to Secure Shop-Floor (OT) Networks Without Stopping Production
- Managed IT Services in Roanoke, VA: What Local Businesses Should Expect
Harsha is a co-founder of SoftwarePac Technologies in Roanoke, VA, and a network engineer with 15 years of experience.




