Hosting backups vs offsite WordPress backups: what each covers
“The host handles backups” can describe a useful recovery service or an assumption nobody has checked. The same is true of “we have offsite backups.” Location matters, but so do access, retention and the ability to recover without the original account.
Ask what failure you want the backup to survive. An accidental plugin deletion, a compromised hosting login and an unavailable hosting provider are different situations. One recovery route may not cover all of them.
Start with access during the failure
For the hosting backup, find out who can request a restore and whether they need the production control panel to do it. Check if the client owns that account or if it sits under an agency login. Record the support route for a locked or unavailable account.
For the offsite copy, ask the same questions. An archive in a different bucket is not independent if the only credentials are stored on the failed server. Keep authorized recovery access in a protected place that the responsible operator can reach.
The WordPress backup handbook discusses recovering files and database data. Whatever service produces the copy, confirm which parts it includes and how you would restore them into another installation.
Compare recovery points, not storage totals
A large backup allowance can still be unsuitable if it retains only the wrong recovery points. Ask how frequently backups run, how long older versions remain available and whether failed jobs are visible. Find out what happens when the storage limit is reached.
For a frequently updated site, distinguish the database schedule from the file schedule. The client’s acceptable data loss should inform the recovery policy. Do not promise that a daily backup preserves every order placed during the day.
Check deletion behavior too. If removing a site or closing the hosting account immediately removes every backup, document the export step that must happen first. Retention should be a written rule, not something you discover while cancelling a service.
Use this decision worksheet
- Who owns the account and who can recover it?
- Can recovery proceed if the production server is gone?
- Which files, databases and configuration are included?
- How old could the latest usable recovery point be?
- Can you retrieve a copy before the account is closed?
- Who performs the restore, and what is outside their support scope?
- When was the recovery route last tested?
Answer these for both services. If they cover different failure scenarios, keeping both may be reasonable. If they duplicate the same dependency and neither has been tested, the second invoice may add less protection than expected.
Account for the work of restoring
A downloadable archive and a managed restore service are not the same offer. With the archive, your team may need a replacement host, database access, URL changes and verification time. With a managed service, check whether the provider restores only to its own infrastructure.
Include transfer time and access delays in a rehearsal. Avoid comparing services solely by a marketing claim about restore speed; your site’s size and recovery destination matter.
IrmaWP’s managed storage allowances and overage terms are one part of this decision, not a replacement for a retention policy. Whichever system you use, run a restore test from the destination you would depend on during an incident. If a question in the worksheet has no answer, send it to the account owner before the next renewal. That unresolved dependency matters more than another copy of the same archive.
Have a correction or a workflow we should cover? Tell the IrmaWP team.