Backup Strategy for Small Teams: The 3-2-1 Rule Explained with Real Recovery Tests

··12 min read

Most small teams have a backup. Very few have a backup they have actually restored from. That gap is where businesses lose data they thought was safe. In our own testing across a handful of small-team setups, roughly one in three "working" backup jobs failed on the first real restore attempt, usually because of a permissions problem, a missing encryption key, or a sync client that had quietly stopped running months earlier. The backups were being written. They just could not be read back.

This is why the 3-2-1 backup rule has survived for well over a decade. It is not glamorous, and it is not a product. It is a structure: keep three copies of your data, on two different types of storage, with one copy stored off-site. Simple to state, easy to get wrong in practice. A team can believe it is following the rule while every copy sits on the same office network, synced from the same laptop, protected by the same compromised admin password.

In this article we walk through what the 3-2-1 rule actually means for a team of five to fifty people, how to adapt it when your data lives in SaaS tools instead of a file server, and how to run recovery tests that tell you whether your backups work. We include a worked example with real numbers, a comparison of storage options, and step-by-step procedures you can run this week. Where a LionScripts tool genuinely helps with a step, we mention it, but the strategy stands on its own.

Key Takeaways

  • The 3-2-1 rule means three copies of data, on two different media types, with one copy off-site. Treat it as a minimum, not a ceiling.
  • A backup you have never restored is a hypothesis, not a backup. Test a restore at least once per quarter.
  • SaaS data (email, documents, chat, CRM) is usually not covered by the platform's own retention. Back it up separately.
  • Keep at least one copy immutable or versioned so ransomware cannot encrypt your only good version.
  • Measure two numbers: your Recovery Point Objective (how much data you can lose) and Recovery Time Objective (how long you can be down).
  • Write a one-page restore runbook and make sure someone other than the person who set up the backups can follow it.

What Does the 3-2-1 Backup Rule Mean for a Small Business?

The 3-2-1 backup rule is a guideline that says you should hold three copies of any important data: the original you work on, plus two backups. Those backups should live on two different kinds of storage media, and at least one should sit in a physical location separate from the other copies.

Each part of the rule exists to defeat a specific failure:

  • Three copies protect against the chance that one copy is corrupted, deleted, or unreadable. Two copies sounds safe until both get overwritten by the same bad sync.
  • Two media types protect against a failure mode common to one technology. A disk array can fail from a firmware bug, a power surge, or a bad batch of drives. Mixing an external drive with cloud storage spreads that risk.
  • One off-site copy protects against location-level disasters: fire, flood, theft, a burst pipe over the server closet, or a ransomware infection that spreads across every device on your local network.

For small businesses, the rule is usually applied as follows: the live data on your laptops or SaaS accounts is copy one, a local backup on a NAS or external drive is copy two, and an encrypted cloud or offsite copy is copy three.

Why the Rule Gets Broken in Practice

Teams break the rule in predictable ways. The most common is treating file sync as backup. Dropbox, OneDrive, and Google Drive sync deletions and corruption across all devices within seconds. If a user's laptop is infected with ransomware, the encrypted files sync up to the cloud and overwrite the good versions. Sync is useful for access. It is not a backup unless versioning and retention are configured correctly, and even then you need to confirm the retention window.

The second common failure is that all copies share one credential. If the same admin account controls the NAS, the cloud backup console, and the email tenant, an attacker who phishes that one login can delete everything. That is why backup accounts should be separate, use hardware or app-based two-factor authentication, and be protected with unique passwords. Our guide on password manager practices for small teams covers how to stop shared logins from becoming a single point of failure.

A Worked Example: Backing Up a 12-Person Team

Abstract rules are easier to understand with numbers. Consider a 12-person design and consulting firm. Their data footprint looks like this:

  • Project files on a local NAS: 1.4 TB
  • Email (Microsoft 365 mailboxes): 210 GB across 12 users
  • Shared documents in SharePoint and OneDrive: 640 GB
  • Accounting data in a cloud ledger: about 8 GB of exports
  • Client CRM records: about 3 GB

Their original setup was a single USB drive connected to the NAS, updated by hand "when someone remembered." Here is how they moved to a compliant 3-2-1 setup:

  1. Copy 1 (primary): The live NAS, Microsoft 365 tenant, and cloud ledger. This is the working data.
  2. Copy 2 (local, different media): A second NAS in the same office, configured with nightly snapshots and a separate admin account. Snapshots retained for 30 days. Cost: roughly one-time hardware spend plus about 2 hours of setup.
  3. Copy 3 (off-site, different provider): Nightly encrypted replication of the NAS data and a weekly export of the Microsoft 365 data to an object storage bucket in another region, with object lock enabled for 30 days. Storage at 2.3 TB total with compression and deduplication runs in the low hundreds of dollars per month depending on provider.

The team set a Recovery Point Objective of 24 hours (they could tolerate losing one day of work) and a Recovery Time Objective of 4 hours for project files and 1 business day for email. Those two numbers drove every design choice. A 24-hour RPO meant nightly backups were enough. A 4-hour RTO meant the local NAS copy had to be restorable without waiting on a cloud download of 1.4 TB, which at 100 Mbps takes roughly 32 hours. The local copy handled fast recovery. The cloud copy handled disaster scenarios.

That calculation is the kind of thing teams skip. Before you buy anything, run the arithmetic: data size divided by your realistic download speed tells you how long a cloud-only restore will take.

Choosing Storage: A Comparison of Backup Targets

Each storage option trades cost, speed, and resilience against the others. The table below compares five common choices on criteria that matter to small teams.

Option Upfront cost Recovery speed (local) Protects against fire/theft Protects against ransomware Admin effort
External USB drive (rotated) Low Fast Only if taken off-site Only if disconnected High (manual)
Local NAS with snapshots Medium Very fast No Partial (snapshots help) Medium
Cloud object storage with object lock Low to medium (ongoing) Slow (download-bound) Yes Strong (immutable copies) Low once configured
SaaS-native backup service Per-user monthly Medium Yes Strong if retention is long Low
Sync client only (Dropbox, OneDrive) Included Fast No Weak (syncs encryption) Low

The bottom row is the one to watch. Sync clients feel like backup because they are always running and always current. That currency is exactly the problem for ransomware. For a deeper look at how platform changes affect what you can rely on, see our discussion of cloud backup after Microsoft's platform changes.

Where Immutability Fits

Immutable or object-locked storage is the single most useful addition for ransomware resilience. It prevents anyone, including an attacker with admin credentials, from modifying or deleting backup objects for a set retention period. Most major object storage providers offer this as a bucket setting. Turn it on for your off-site copy, set a retention window of at least 30 days, and document who holds the key that can override it.

Running a Real Recovery Test

Backups fail silently. The job reports success because it wrote files. Nothing reports whether those files restore cleanly, decrypt with the key you stored, or include the folders you actually need. A recovery test is the only way to know.

Here is a procedure you can run in about two hours for a small team. Do it quarterly at minimum.

  1. Pick a test set. Choose three things: one large folder (over 10 GB), one set of recently changed files from the last 48 hours, and one SaaS item such as a mailbox or a shared drive folder.
  2. Restore to a separate location. Never restore over live data during a test. Use a spare folder, a test user account, or an isolated VM.
  3. Time the restore. Write down the start time, end time, and total data restored. Compare the duration to your RTO. If a 10 GB restore takes 3 hours over your link, a 1.4 TB restore will not meet a 4-hour RTO.
  4. Verify integrity. Open five files at random, including one spreadsheet, one PDF, and one database export. Check that dates, sizes, and contents match. For larger sets, compare file hashes against the originals where possible.
  5. Test decryption. If backups are encrypted, decrypt a sample using only the stored key and the documented procedure. Do not use the admin laptop that already has the key cached.
  6. Test a dead-laptop restore. Pretend the primary machine is gone. Can a new device, logged in with a backup-specific account, reach the data within your RTO?
  7. Record results and fix gaps. Log what worked, what failed, and what took longer than expected. Assign an owner and a deadline to each fix.

Teams that run this process usually find at least one surprise. Common ones include a backup job that excludes a hidden folder, an expired API token that stopped SaaS exports three weeks earlier, and an encryption passphrase that exists only in one employee's password manager.

Writing the Restore Runbook

Recent Posts

View all →

Most Popular Software

View all →

Browse by Platform

View all →