All insights

Backups need a recovery plan

A backup notification is only part of the picture. Know what is covered, how restoration works, and what your team will do during an outage.

A violet strap lifts a spare metal fitting from its case.

The practical takeawaySchedule a controlled restore test and record exactly what it takes to bring your most important website journeys back.

  1. Choose a backup
  2. Restore a test copy
  3. Verify the site

A reassuring backup notification tells you that a process ran. It does not tell you whether the right information was included, whether someone can access it during an incident, or whether the restored website will connect correctly to the rest of your business.

Recovery is the practical outcome you want. If an update breaks the site or important information is lost, your team needs a clear way to get back to work. You can make that outcome more dependable by treating backups as one part of a recovery plan, with coverage, ownership, and tests that reflect how the website is actually used.

Define what needs to be saved

Start by listing the pieces of your site: application files, uploaded media, databases, configuration, and any separate systems that hold customer or donation records. Ask the provider to show which backup covers each item and what is explicitly excluded.

In a typical WordPress installation, files and the database are separate parts of the backup. WordPress’s official backup documentation explains why both are needed for a full restoration. A copy of the theme does not contain the posts and settings stored in the database.

External services need their own consideration. Restoring the website will not necessarily restore a CRM, an email platform, or a payment service. Document the boundary between systems so the team knows which provider to contact and which information must be reconciled separately.

Choose recovery expectations with the business

Two questions make the discussion concrete: how much recent information could you reasonably reconstruct, and how long could the important website functions be unavailable? Your answers should guide the backup schedule and recovery arrangements.

An illustrative brochure site that changes monthly has different needs from a nonprofit receiving frequent registrations during a campaign. A nightly backup might preserve the brochure content adequately while leaving the second team with a significant reconciliation problem. The point is to choose a schedule around the work, rather than accept a default without discussion.

Put realistic expectations in writing. A backup interval is not a guarantee of zero data loss, and a restore button does not promise a particular recovery time. Account access, the size of the site, the incident itself, and connected services can all affect the work involved.

Test restoration before an emergency

Arrange a controlled restore to an isolated test environment with the person responsible for the website. The test should avoid sending real customer messages, triggering payments, or writing test information into live integrations. Decide those boundaries before starting.

Then follow a short, representative checklist. Can the website load? Is the expected content present? Can authorized staff sign in? Do images and downloads work? Can the team complete the important customer journeys using test data?

Measure the process and record the missing pieces. NIST’s small-business recovery guidance emphasizes defined responsibilities and checking the integrity of restored assets. A rehearsal gives those responsibilities and checks a concrete form for your website.

Keep the recovery route available

Think through what happens if the normal hosting account is inaccessible. Who can contact the provider? Where is the recovery information held? Does the organization have an authorized way to access the backup without depending on the same account that has failed?

Protect backup access with appropriate permissions and account security. Keep copies and retention arrangements suited to the risks you are addressing. Deletion permissions deserve particular attention: a backup is less useful if an incident can remove both the live site and every recovery copy through the same unchecked access.

If you suspect a security incident, avoid treating an arbitrary recent backup as automatically trustworthy. Have the responsible technical team investigate the likely timeline and recovery approach. Restoring a compromised version or leaving the original issue unresolved can repeat the problem.

Write a runbook someone can follow

Your runbook can be short enough to use under pressure. Include the website and provider names, current emergency contacts, where approved recovery credentials are managed, the recovery steps, and the order in which important functions should return.

  • Who decides that recovery should begin?
  • Who performs the technical work and verifies the result?
  • How will staff and affected customers receive updates?
  • Which records or integrations need reconciliation afterward?
  • What evidence confirms that normal service has resumed?

Make your next maintenance conversation a recovery conversation. Ask for the date of the last successful test, the functions checked, and any open issues. If the answer is unclear, schedule a rehearsal. TripleSkinny can help map the dependencies, coordinate the test, and keep the plan current as your website and business change.

Sources & further reading