A website backup strategy should make recovery predictable, not merely create a folder of old files. This guide explains what a small business website should back up, how often to capture each component, how to retain copies safely, and how to test a restore before an outage or failed update turns into lost business.
Overview
A website is usually more than its visible pages. It may include application files, a database, uploaded media, DNS records, email settings, SSL certificates, scheduled tasks, environment variables, and hosting configuration. A useful backup plan accounts for each dependency and identifies which components must be restored together.
Start by defining two targets:
- Recovery point objective (RPO): the maximum amount of recent work or data the business can afford to lose. If the RPO is 24 hours, a daily backup may be acceptable; a busy store or membership site may need more frequent database backups.
- Recovery time objective (RTO): the target time for restoring a working site. A brochure site may tolerate a longer manual recovery, while a site that receives orders or leads may require a documented, faster process.
These targets determine the schedule, storage, and level of automation. Hosting snapshots can be useful, but they should not automatically be treated as the entire backup plan. A backup stored only on the same hosting account may be unavailable if the account, server, credentials, or provider access is affected.
For WordPress sites, the essential recovery set normally includes the WordPress files, the database, themes, plugins, uploads, and configuration values. Before a major update, use a staging environment where practical; the staging environment guide explains how to separate testing from the production site.
Checklist by scenario
Use this scope table as a starting point, then adjust it to match the site’s platform and business processes.
| Component | What to capture | Suggested frequency | Restore notes |
|---|---|---|---|
| Website files | Application code, themes, plugins, templates, and public assets | Daily or after meaningful changes | Confirm file permissions, ownership, and required runtime versions |
| Database | Content, users, orders, form submissions, settings, and metadata | Daily at minimum for active sites; more often when the RPO requires it | Test importing a copy and check character encoding, tables, and application connections |
| Media | Images, documents, videos, and generated assets | With file backups; more frequently during publishing periods | Verify paths, thumbnails, permissions, and links from restored content |
| DNS | Nameservers, A and AAAA records, CNAMEs, MX records, TXT records, and verification entries | After every DNS change and during scheduled configuration exports | Keep a readable record outside the DNS provider so records can be recreated |
| Mailbox data, aliases, forwarding rules, filters, and relevant provider settings | According to the mail provider’s export and retention capabilities | Website backups usually do not include hosted email; confirm this separately | |
| Certificates | Certificate information, renewal method, and any required private-key material | Document after issuance or renewal | Protect private keys carefully; in many setups, reissuing is safer than storing a key |
| Configuration | Environment variables, cron jobs, deployment settings, firewall rules, and service dependencies | After configuration changes | Record secrets securely rather than placing them in an ordinary backup folder |
For a brochure or informational site
Capture the site files, database if the CMS uses one, media library, DNS records, and configuration. A daily file and database backup may be more than necessary for a rarely changed site, but weekly backups combined with a backup before edits provide a practical baseline. Keep a current copy of the website launch checklist, administrator contacts, and hosting access instructions.
For a lead-generation or appointment site
Prioritize form submissions, appointment data, and other records that may not exist in the website files. Back up the database at a frequency that matches the acceptable loss window. Confirm whether submissions are also delivered to business email or a separate service; do not assume one copy protects against every failure.
For an online store, membership site, or publishing site
Use a shorter database backup interval when new orders, accounts, comments, or paid content are created frequently. Combine automated backups with transaction or platform exports where available. Before plugin, theme, application, or infrastructure changes, create a known-good backup and record the exact change being made.
For a developer-managed application
Version-control application code and infrastructure definitions, but do not assume a repository contains production data, uploaded files, secrets, or provider-side settings. Export those items separately. Developers can also document the deployment command, runtime version, database engine, external services, and required environment variables. For containerized workloads, pair image or source recovery with persistent-volume and database recovery; see the Docker application deployment checklist for related operational considerations.
What to double-check
Use more than one storage location
Keep at least one backup copy separate from the production hosting account. Options may include an object-storage bucket, a different cloud provider, or encrypted offline storage. The right choice depends on access controls, retention needs, recovery speed, and the sensitivity of the data. Test that the backup can actually be downloaded by an authorized person without relying on the production server.
Separate backup frequency from retention
Frequency answers how often a copy is created. Retention answers how long those copies remain available. A practical example might be keeping recent daily copies for a short period, weekly copies for several weeks, and a monthly copy for longer-term reference. Adjust this to the business’s legal, operational, and content needs. Retention is not a substitute for testing: a large collection of unusable backups still provides little protection.
Protect access and sensitive data
Use separate backup credentials where possible, limit access to the people and systems that need it, and enable multi-factor authentication on storage and hosting accounts. Encrypt backups in transit and at rest when the platform supports it. Treat database exports, customer records, mailbox archives, configuration files, and private keys as sensitive. Store encryption keys or recovery credentials in a secure system that is accessible during an incident but not bundled unprotected with the backup.
Verify the backup, not just the job
A successful scheduled task may only mean that a process ran. Check that the expected files exist, the database export has a plausible size and timestamp, and the storage destination is reachable. Set alerts for failed jobs, low storage capacity, expired credentials, and missing recent backups.
Document dependencies
Write down the domain registrar, DNS provider, hosting account, database details, email provider, SSL renewal method, third-party integrations, and emergency contacts. If a domain expires or DNS points to the wrong service, a restored server may still be unreachable. Review the guide to pointing a domain to hosting when documenting DNS dependencies, and keep renewal access separate from the website server.
Common mistakes
- Backing up only the public files: CMS content and business records often live in the database.
- Keeping every copy on the same server: a server problem, account lockout, or destructive event can affect production and backups together.
- Ignoring email and DNS: these services may be managed outside the hosting account and may not be included in a hosting backup.
- Saving secrets in plain text: configuration exports can contain credentials, tokens, or private keys.
- Testing only after a disaster: the first restore should not happen under pressure.
- Restoring over production immediately: restore into a temporary location first, compare the result, and preserve current data before replacing anything.
- Assuming a hosting feature is a complete policy: confirm scope, schedule, retention, storage location, and restore access rather than relying on a label such as “automatic backups.”
When to revisit
Review the backup plan before seasonal planning cycles, major campaigns, platform updates, hosting changes, and migrations. Revisit it whenever the site adds a store, membership feature, new form workflow, business email, external integration, or a larger media library. A domain transfer or DNS change is also a useful prompt to verify that domain, DNS, hosting, and email recovery information is documented together. The domain transfer checklist can help identify related dependencies.
Run a restore test monthly for an active or business-critical site, or quarterly for a low-change informational site. Perform an additional test before significant updates and after changing hosting providers. Use this repeatable procedure:
- Record the current RPO and RTO targets and select a recent backup.
- Confirm the backup timestamp, contents, integrity check, and storage location.
- Restore files and the database into a staging site or isolated environment, not directly over production.
- Apply the documented runtime, configuration, permissions, and environment variables.
- Temporarily connect the restored site to a test domain or protected URL.
- Check the homepage, key landing pages, forms, logins, search, media, checkout or appointment flows, redirects, and email notifications as applicable.
- Confirm that DNS, SSL, scheduled tasks, third-party integrations, and database connections behave as expected. Use the SSL setup guide when HTTPS is part of the test.
- Measure the elapsed restore time and note any manual steps, missing files, broken links, or inaccessible credentials.
- Update the runbook, correct the backup process, and record the test date and result.
A backup strategy is complete only when another authorized person can follow the documentation and recover a working site. Put the next restore test on the calendar, and update the plan whenever the website, hosting environment, or business risk changes.