Does your website actually have backups, and who can restore them

Where the automatic schedule actually saves your files

A settings tab, opened for the first time since launch, shows a backup toggle switched to on. It's been sitting there quietly for months, doing something, though nobody could say exactly what.

Automatic backups run on a set schedule, daily or weekly depending on the plan, and save a copy of the site's files and database. Where that copy is stored matters more than the schedule itself. A copy saved on the same server as the live site disappears in the exact failure that would make you need it: a crash, a bad update, a hacked account.

Schedules vary a lot between hosts. Some back up daily, some weekly, a handful only when you ask. None of that is standard across providers, so don't assume your plan matches a friend's, or even a previous host you've used before.

Ask specifically where backups are stored, not just on what cadence they're made. A copy sitting in separate cloud storage survives a server failure. A copy sitting next to the site it's backing up does not, and that single detail decides whether the backup is worth anything at all.

A backup existing is not the same as a backup working

A schedule can run successfully for a year and still fail the one time it's needed. Corrupted files, partial saves, and jobs that quietly stopped running months ago all look identical to a working backup right up until someone tries to use one.

The only real proof is a restore, done somewhere other than the live site, checked by a person clicking through the result afterward. If that's never happened, the backup is a claim rather than a fact. It's a line item on a settings page, not evidence.

Testing doesn't need to be dramatic. A single restore to a staging copy or a test domain, done every few months, is enough to know the process still works. A backup nobody has opened is worth exactly as much as one nobody has made, until it's proven otherwise.

Doing a manual backup yourself as a stopgap

If nobody can confirm the automatic system works, a manual copy is better than waiting to find out. Nearly every platform has a basic export option somewhere in the admin panel, sometimes just a download of the files and a database dump.

It won't be elegant. It will give you something to fall back on while the real question gets sorted out. Save that file somewhere off the hosting account entirely, a personal drive or a separate cloud folder, not another directory on the same server.

A manual export takes a few minutes the first time and less after that. Keep it as a habit even once a real system is confirmed working. A second copy in a second place never hurts, and it costs you almost nothing to keep one on hand.

How far back a retention window typically reaches

Backup systems don't keep every version forever. They overwrite older copies after a set number of days, or once a fixed number of saved versions piles up, whichever limit hits first.

That window matters because problems aren't always caught the day they happen. A page can sit broken for two weeks before anyone reports it, by which point the clean version you need might already be gone.

A site that changes constantly has more to lose here than one that barely changes at all. A short retention window is riskier than it sounds once you're publishing new content every week. Ask plainly what the retention window actually is, in days, not in vague reassurance about how backups are handled generally.

What a backup does and does not protect against

A backup rolls the site back to a specific point in time. It does nothing to fix whatever caused the damage in the first place, which is the part people forget in the moment of relief after a restore finishes.

If a plugin has a security hole, restoring from backup brings that same hole right back with it, unless someone also closes it separately. A site restored from a clean copy can be compromised again within hours if the real vulnerability was never patched. The backup undid the symptom, not the cause.

Backups and what website security monitoring actually catches solve two different problems. Having one doesn't mean you have the other, and assuming otherwise is how a site gets hit twice in the same month.

Who actually holds the ability to run a restore

Backups existing is one thing. Somebody having the login, the permission level, and the working knowledge to trigger a restore is a separate one, and it's the piece that gets skipped in this whole conversation more than any other.

If that person is a former designer who's stopped answering email, the backup might as well not exist. Write the answer down somewhere you'll actually find again: a name, a login, a phone number if it comes to that. An outage is a bad time to be figuring this out for the first time.

Confirm this now, well before an outage, exactly who on your side or your provider's side actually has the ability to pull that trigger. Ask them directly. Get the answer in a plain sentence you'd trust in an emergency, not a shrug or a guess. If the answer takes more than a sentence to explain, that's worth fixing before you need it, not during the outage itself.

It's one piece of the larger question of who takes care of a website after launch, and if the site does go down, what happens in that first hour depends entirely on whether that person can be reached in time.

More from the blog

All posts