← All posts

Why a Good Backup Strategy Matters (3-2-1 and Beyond)

Everybody agrees backups are important. Almost nobody has a good one. I've seen people lose years of photos to a single dead drive, and I've watched a "we'll set up backups next sprint" project lose a database on a Friday afternoon. The pattern is always the same: it felt like a problem for future-you, right up until future-you showed up.

So let's talk about what a decent backup strategy actually looks like, and how to build one whether you've got zero budget or a bit of cash to spend.

Why You Actually Need One

Drives die. That's the boring, predictable reason. But hardware failure is only one of the ways you lose data. There's also:

  • Accidentally deleting the wrong folder (we've all done it)
  • Ransomware encrypting everything it can reach
  • A bad update or a buggy script that quietly corrupts files
  • Theft, fire, or a flooded basement
  • A cloud account getting locked or hacked

Notice that most of those have nothing to do with a drive failing. That matters, because it means "I have a second disk" isn't a strategy.

The 3-2-1 Rule

This is the classic, and it's classic for a reason. It's simple enough to remember and it covers most disasters:

  • 3 copies of your data: the original plus two backups
  • 2 different types of storage: say, your laptop's SSD and an external drive, or a NAS and the cloud
  • 1 copy offsite: somewhere that won't burn down or get stolen with the rest

The logic is that no single event should be able to take out everything. A dead drive kills one copy. A house fire kills the local copies. Having all three scenarios covered means you'd need seriously bad luck to lose it all.

Things That Look Like Backups (But Aren't)

This is where most people get burned, so it's worth being blunt.

RAID is not a backup. RAID keeps a machine running when a disk dies, which is great for uptime. But if you delete a file or ransomware encrypts it, RAID happily replicates that change to every disk. Same goes for corruption.

Sync is not a backup. Dropbox, OneDrive and friends mirror your folder. Delete something and it vanishes everywhere. Get encrypted by ransomware and the encrypted files sync right along. Some of these services do offer version history, which helps, but treat that as a bonus and not your plan.

Snapshots on the same machine aren't enough either. They're lovely for "oops, I broke my config," but if the machine or the storage pool goes, so do the snapshots.

A real backup is a separate copy, with history, that a mistake or an attacker on your main system can't easily touch.

Leveling Up: 3-2-1-1-0

Ransomware changed the game a bit. Modern attacks go looking for your backups and encrypt or delete them first, so a backup drive that's permanently plugged in is a sitting duck. That's why you'll see the rule extended to 3-2-1-1-0:

  • 1 copy that's offline or immutable: unplugged, air-gapped, or stored somewhere that doesn't allow deletes for a set period (object lock on cloud storage, for example)
  • 0 errors: you've actually verified that the backups restore

That last one is my favorite, because it's the step everyone skips. A backup you've never restored is a hope, not a backup. Silent failures are real: the job has been "running" for eight months and has been failing for seven of them.

Building It for Your Budget

Here's how I'd approach it at different spending levels.

Practically free. Use what you already have. Keep your data on your main machine, back it up to an external drive with something like restic, Borg, or even plain rsync, and rotate a second drive that lives at a friend's or family member's place (or your desk at work). Swap them every month or so. It's not glamorous, but it hits 3-2-1 and the offline copy comes for free, since a drive sitting in a drawer can't be encrypted.

A few bucks a month. Add encrypted cloud backup. Object storage from providers like Backblaze B2 costs only a handful of dollars per terabyte per month, and tools like restic encrypt your data before it leaves your machine, so you're not trusting the provider with your files. Keep a local copy for fast restores and the cloud copy for disasters. Do check current prices and restore (download) fees before committing, since those vary and change.

Serious but still reasonable. Get a NAS for local backups with snapshots, push an encrypted copy to cloud storage, and turn on immutability or retention locks for the cloud side so even a stolen password can't wipe your history. Use separate credentials for the backup system than the ones you use day to day, and turn on MFA.

Running a business or a production site. Everything above, plus automated monitoring and alerts on backup jobs, longer retention (weeks or months, not days), and scheduled restore drills. Time how long a full recovery takes, because that number is what your downtime really looks like.

Make It Boring and Test It

The best backup strategy is the one that runs without you thinking about it. Automate it, set up an alert for when it fails, and put a recurring reminder in your calendar to restore a few files at random. Once a quarter is plenty for personal stuff. Do a full restore at least once, ideally onto a fresh machine, so you know the whole chain actually works.

Don't forget to back up the things you'll want when everything's on fire: your password manager's export, your encryption keys, and your config files. Having a pristine backup you can't decrypt is a special kind of heartbreak.

You don't need a perfect setup. You need two copies you didn't have yesterday, one of them somewhere else, and a restore you've tried at least once. Go plug in that drive.