Sync keeps your files identical across the machines you've connected. It does that whether the change was a good one or not — an overwrite, a bad find-and-replace, a folder dragged into the wrong place at the end of a long day. The thing that undoes any of it is version history, and "version history" on a feature page is a checkbox, not a spec.
This is for anyone deciding where their working files should live, or wondering whether the drive they already use would actually get them out of trouble. Four details decide that, and you can check all four yourself.
The verdict
A drive whose history can't restore a folder you deleted last month is storage, not safety — and you still need a second copy either way.
History inside a drive protects you from your own edits. It sits in the same account, behind the same password, subject to the same billing status as the files it's protecting. Useful, but not a backup.
The four checks
Retention window — how far back history goes
Why it decides anything. You often notice damage long after you did it
How to test it. Look up the window for your plan, not the top plan
Whether history covers deletions
Why it decides anything. Trash and version history are separate systems
How to test it. Delete a throwaway file, empty the trash, try to get it back
Per-file vs. whole-folder restore
Why it decides anything. One file is a click; a folder of them is a project
How to test it. Restore a test folder to how it looked yesterday, in one action
Offline edits on two machines
Why it decides anything. Both machines think they're right
How to test it. Disconnect both, edit the same file, reconnect
Run these on a throwaway folder with junk files in it. Never test recovery on work you can't lose.
1. Retention window
The retention window is how long old versions are kept before they're discarded. This is the number to look up first, and the plan described on a marketing page may not be the plan you're on. Check the window attached to your own tier, and check whether it applies to deleted items too — those are frequently governed by a different clock.
The window matters because of how damage gets noticed. A truncated file or a botched replace-all can sit unnoticed until the next time you open that document, which may not be this week. If your window is shorter than your noticing habits, history will be technically present and practically useless.
2. Does history cover deletions, or only edits?
This is the check the phrase "version history" doesn't answer, and it's the one that catches people.
Versions hang off a file. Delete the file and, depending on the product, there may be nothing left for the versions to attach to — recovery then goes through the trash instead, which has its own retention rules and its own emptying behavior. A drive can have generous edit history and a short, shallow answer for deletions.
The test is blunt and takes minutes. Put a few junk files in a folder. Delete the folder. Empty the trash the way an impatient person would. Now try to bring it back. If you can't, you've learned that your drive protects your typing but not your dragging — and the case it doesn't cover is the harder one to catch in time: a folder that goes missing without anyone remembering when.
The same logic applies to sync-propagated damage. If files get encrypted or scrambled on one machine, the drive does its job and distributes the result. Recovery then depends entirely on how far back you can roll and whether you can roll back everything at once, which is the next check.
3. Per-file restore vs. point-in-time restore
Per-file restore means opening a file's history and picking an older version. Point-in-time restore — some products call it rewind or folder restore — means telling the drive to make a folder look the way it did at a chosen moment, across everything inside it.
For one damaged spreadsheet, per-file is fine. For a folder where a sync went wrong across many files at once, per-file restore turns recovery into repetitive manual work, and manual work at 6pm is where people give up and accept the loss. If the drive holds a project directory rather than a handful of documents, folder-level restore is the feature worth paying for.
Test it directly: change several files in a test folder, then try to roll the whole folder back in a single action. If the only path is file by file, you now know what a bad day costs you.
4. Two machines, both offline
Take a laptop and a desktop off the network. Edit the same file on both. Reconnect.
What you're looking for is which of these your drive does:
- Keeps both, with one saved alongside the other as a conflicted copy — a duplicate file marked to show the drive couldn't decide
- Keeps one as current and pushes the other into version history
- Keeps whichever arrived last and discards the other
The third behavior is the one to know about before it happens, not after. None of these is wrong in the abstract; they're trade-offs between clutter and silence. But a drive that resolves quietly can lose an afternoon's work without producing anything that looks like an error.
If you work on one machine and never edit offline, this check matters much less to you. If you move between a laptop and a desktop, or work on planes and trains, it's the one that will bite.
Who can skip this
- You already run a separate versioned backup, on a different service or a local drive with its own history. Then your cloud drive's history is a convenience, not your safety net, and these checks are informational.
- Your work lives inside an app's own cloud — a notes tool, a design tool, a code host — rather than in files on a disk. Those have their own history models, and this checklist doesn't map onto them cleanly.
- The drive holds only copies of things that exist elsewhere. If nothing in it is the sole original, losing it costs time, not work.
Everyone else: the checks are worth an afternoon, once.
The next action
Run check two today. Make a folder of junk files, delete it, empty the trash, and try to get it back. Of the four, it's the one you can run fastest: it won't tell you how the other three turn out, but it tells you in minutes whether deletions are covered.
Whatever the answer, keep a second copy somewhere the drive can't reach. A commonly repeated shorthand for this is the 3-2-1 rule — three copies, on two kinds of media, one of them off-site — a rule of thumb, not something this checklist tests. Version history is what you reach for when you broke a file. A second copy is what you reach for when something breaks the drive.