A business website backup should cover its database, application and themes, uploaded files, and the configuration needed for recovery. Test the complete set in an isolated environment. Agree how much recent data the business can reconstruct and how long essential service can be unavailable, then choose a schedule, retention policy and responsible people. Acceptance should cover product pages, both languages and enquiry records.
This checklist is for business websites that maintain product specifications, manuals and bilingual enquiries. Use the self-hosted CMS selection guide for deployment requirements and the supplier handover and migration checklist when changing providers. This article covers ongoing backup and incident recovery.
1. Define the business data window and recovery deadline
The business owner should approve a recovery point objective (RPO): the maximum acceptable window of recent data loss. Also agree a recovery time objective (RTO): the maximum time from a defined incident start until essential business service is restored. Record the timing convention, required service scope and approving person, with separate requirements for content and enquiries.
A site publishing weekly news and receiving daily enquiries can use different protection schedules for those data types. Have the technical owner explain the proposed protection and measure it through a drill. Infrastructure availability, accessible pages and sales access to enquiries each need an acceptance result.
A planning row can contain: product and bilingual content | acceptable reconstruction window | enquiry tolerance | essential-page recovery time | notification recovery time | owner. Fill in business-approved values and retain supplier commitments separately from measurements.
2. Inventory a complete recoverable backup set
- Database: sites, categories, models, fields, content and translations, settings, accounts and permissions, and the enquiry records within the agreed retention scope.
- Application and themes: the matching source version, dependency lock files, custom changes, themes, and required build outputs or workable build instructions.
- Files: uploaded images, manuals, certificates and other referenced material. Include external object storage according to its actual location.
- Controlled configuration: environment instructions, required settings and keys, DNS records, access arrangements for storage and email services, and scheduled tasks or queues.
For each set, record site coverage, versions, start and completion times, a file manifest, checksums and recovery instructions. For multisite deployments, list every site and shared resource covered by the database and document the restoration impact. Database translations and text or language variants inside uploaded files need their own checks.
The official WordPress backup guidance identifies database and file backups and recommends treating them as a corresponding set. Use this as a scope reference; the project maintainer must confirm the actual CatchCMS directories, database and external storage.
3. Establish consistency and account for ongoing changes
Database exports and file copies have time windows. An attachment replaced during that window can produce a restored record pointing to the wrong file version. Have the technical owner implement a tested consistency method, such as briefly pausing editing and file replacement, or using database and storage mechanisms that support consistent recovery.
Record the pause scope, start and end times, and resumption of writes. When enquiries continue during the window, preserve those incremental records separately. Use a common time zone and record each source. Check the backup job result, file integrity and restored business behaviour as separate outcomes.
Choose frequency and retention around actual business needs, including delayed discovery of accidental deletion, failed updates and server loss. Keep a recoverable version before system or theme changes and bulk edits, and verify that older required recovery points remain retrievable.
4. Store an independent copy and test retrieval and decryption
Prepare a backup destination independent of the production server. Confirm retrieval arrangements during a cloud account incident, permission revocation or supplier change. Use controlled transfer, encrypted storage and least-privilege access. Keep access keys separately from backup copies and name a primary recovery contact and a substitute.
Manage backups containing enquiries, accounts or configuration within the approved access scope. Put only controlled evidence locations in ordinary handover records. Have an authorised recovery operator download, verify and decrypt a copy, recording the time required. Keep secrets and personal information in the agreed secure channel.
For hosting packages, confirm both backup scope and restoration permissions. The official cPanel backup documentation describes full and partial backups separately; automatic restoration of a full backup requires WHM or the hosting provider. For that hosting arrangement, include provider contacts, access and waiting time in the recovery plan.
5. Perform a complete restoration in isolation
- Prepare a separate database, file location and test hostname, restricting access to the acceptance team. Configure test email and external integrations and verify every connection target. Follow the staging isolation checklist.
- Have the receiving operator deploy the matching application and theme, import data, and restore files and controlled configuration using the written instructions. Record step timings and failures.
- Reconcile counts by site, category and language. Inspect representative specifications, body text, images and language-specific attachments. Use a routine editor account to edit and read back a sample.
- Open both language versions of the homepage, product list and detail pages. Check HTTP status, title, description, canonical, language links and applicable structured data. Read specifications, open downloads and reach the contact path on a phone.
- Submit a clearly labelled test enquiry. Verify its stored record, test email and receiving person using the enquiry and notification checklist.
Record recovery completion after all essential checks pass. Assign failed items and repeat the relevant checks after correction. Database import, environment setup, file retrieval and business acceptance all contribute to the measured restoration record and help refine the target and plan.
6. Preserve enquiries from the incident window before restoring
First preserve incident evidence and the latest data still available, then approve the recovery point. The sales owner should reconcile enquiries, follow-up changes and other required records created after that point. Where the current system remains readable, export the authorised scope securely. Email and sales records can help cross-check sources; mark missing fields for confirmation.
Use an incremental ledger: source system or site | original record ID | submission time and time zone | source | processing status | destination record | reviewer. Check matching IDs against source and time. The technical owner implements and verifies the merge, while sales confirms duplicates and outstanding items. The responsible person authorises real notifications based on the reconciliation.
Before returning to production, retain a snapshot of the recovered environment and merge evidence. Confirm write destinations, notifications and external integrations. A shared-database restoration affecting several sites requires preservation of each site's incremental data and approval of the full impact.
7. Fictional example: measuring a four-hour recovery target
All times and results in this example are fictional. A bilingual equipment website agrees a two-hour RPO and four-hour RTO. Its latest consistent recoverable backup point is 08:00, and the agreed incident clock starts at 09:30. The backup data window is one hour and thirty minutes, with content and enquiries in that window requiring reconciliation. Reconstruction depends on the records actually available.
The operator starts retrieval and deployment at 09:45, finishes data restoration at 11:10, and completes bilingual pages, attachments, editing and test enquiry checks, incremental data merging and business approval by 11:50. The product owner reconciles one specification change and sales reconciles three new enquiries, with both signing the merge results. All agreed essential business service is restored at 11:50, two hours and twenty minutes after the agreed incident start.
If retrieval, a missing key or provider waiting time pushes recovery beyond four hours, record that the drill missed its target. Adjust storage, permissions or service arrangements and repeat it. These figures illustrate timing and reconciliation; project performance must be measured.
8. Record acceptance, cost and responsibility together
For each drill, record: site | backup point and version | copy location | retrieval and decryption result | incident clock start | recovery completion | data gap and incremental ledger | bilingual and enquiry checks | operator | business approver | issues and retest date. Schedule recurring drills and repeat relevant checks after changes to the application, storage or supplier.
CatchCMS provides multisite, multilingual content, content models and form records that can be checked against the actual project configuration after restoration. The language content checklist helps identify translated text and attachments. Full-site backup scheduling, encrypted storage, failure alerts, database restoration and incremental reconciliation are implemented and tested within the project's operations plan. Theme version switching manages template assets.
Budget for storage, transfer, recovery resources, drills, incident response and work beyond included service hours. Use the long-term cost checklist. Name the people who detect failures, retrieve copies, restore service and accept business operation. Record service hours and response arrangements from the actual quotation.
How often should a business website be backed up?
Start with the acceptable reconstruction window for content and enquiries. Check completion times, failure retries and copy availability, then compare the actual recoverable data point demonstrated in a drill with the agreed objective.
What needs checking after the database is restored?
Verify the matching application, themes, uploaded files, configuration and external services. Test both language versions, editing and enquiries. Retain count checks, sample results and the treatment of records created in the recovery window.
How can a small business start a recovery drill?
Appoint business and technical owners, restore one complete set in an isolated environment, and measure the time from retrieval to business acceptance. Address the gaps before agreeing the ongoing schedule and maintenance budget.
Prepare your site and language list, database and attachment locations, maintenance contacts and acceptable data window. Use project consultation to confirm the CatchCMS backup and recovery implementation scope and arrange a restoration with reviewable evidence.