
Key takeaways
- Follow the 3-2-1 rule: 3 copies of your site, on 2 different kinds of storage, with 1 copy offsite.
- Decide 2 numbers before you need them: how much data you can afford to lose (RPO) and how long you can be offline (RTO).
- A backup needs 4 parts to be useful: files, database, uploads, and configuration. Missing any one means a broken restore.
- An online store may need backups every hour. A brochure site that changes monthly may be fine with monthly ones.
- A backup you’ve never restored is a guess. Test a full restore at least a few times a year.
Almost every site has some kind of website backup. Far fewer have one that’s been tested, stored somewhere an attacker can’t reach, and paired with a plan for actually getting the site back. The gap shows up on the worst day possible, when the restore fails or the only copy turns out to be on the server that just died.
What is the 3-2-1 backup rule?
The 3-2-1 rule says to keep three copies of your data, on two different types of storage, with one copy in a different location. For a website, that usually means the live site, a backup on separate storage, and a third copy with a different provider or region.
Each copy covers a different failure. A second copy on the same server helps when someone deletes the wrong page. Only the offsite copy helps when the server fails or the hosting account gets compromised.
A common mistake is treating the host’s built-in snapshots as the whole plan. They often live under the same account as the site, so if that account is compromised, the backups can go with it.
Offsite and immutable copies
An immutable backup can’t be changed or deleted for a set period, even by someone with your login. That matters because some attacks go after backups first. If an intruder can reach your backup storage with the same credentials they used on your site, they can wipe it before you notice. Keep at least one copy offsite, on separate credentials, with deletion locked or delayed. Many cloud storage services offer this as an object lock or retention setting.
What do RPO and RTO mean for a website?
RPO (recovery point objective) is how much data you can afford to lose, measured in time. RTO (recovery time objective) is how long the site can be down before it really hurts. Together they tell you how often to back up and how fast your restore process has to be.
In plain English:
- RPO answers “if we restore from the last backup, how much work is gone?” If you back up every night and the site fails at 6 p.m., you lose up to a day of orders, form entries, and edits. If that’s unacceptable, your backups need to run more often.
- RTO answers “how long until we’re back?” If the business can live with a day offline, a manual restore is fine. If every hour costs sales, you need a tested, fast process and probably a standby environment.
Set these with the business side, not just whoever runs the server. “We can lose four hours of data and be down for two” is a decision, and once it’s made, the backup schedule mostly writes itself.
What should a website backup include?
A complete website backup includes four things: the code files, the database, the uploads folder, and the configuration. Backing up only some of them is how teams end up with a site that technically restores but is missing its images, its settings, or its last month of orders.
- Files. The theme, plugins, and any custom code. If your code lives in a Git repository, that’s a backup in its own right, but keep it in the plan.
- Database. Every page, post, product, order, user, and setting. On most WordPress sites this is the part that changes most often.
- Uploads. Images, PDFs, and other media. Often the largest part of the backup, and the easiest to forget when it’s stored separately.
- Configuration. Files like wp-config.php, server settings, redirect rules, environment variables, and DNS records. Keep secrets in this part encrypted.
Also write down anything outside the server the site depends on: payment gateway settings, email delivery setup, API keys for integrations. You don’t back those up the same way, but you need them listed to rebuild quickly.
How often should you back up a website?
Back up as often as your content and data change. A site that takes orders or submissions all day needs frequent database backups. A site that changes a few times a month can back up much less often. The table below is a starting point, not a rule.
| Site type | Database | Files and uploads | Why |
|---|---|---|---|
| Brochure site, rarely edited | Monthly, plus before any change | Monthly, plus before any change | Very little changes between backups |
| Marketing site edited weekly | Daily | Weekly | Content edits add up fast |
| Lead generation site with forms | Daily or more | Weekly | Form entries may only exist in the database |
| Online store | Hourly or continuous | Daily | Every lost hour is lost orders |
| SaaS app or membership site | Continuous or near it | Daily | User data changes all the time |
Whatever the schedule, take a manual backup right before any major update, migration, or redesign. That’s the moment you’re most likely to need one.
For reference, the Basic plan in our ongoing support and retainers includes monthly backups, which fits the first row of that table. Sites that change more often need a schedule to match, and that’s worth raising when you scope a plan.
Not sure your backups would actually work?
Free 30-minute call. No pitch, no pressure.
How do you test a website backup?
You test a backup by restoring it somewhere that isn’t production and checking that the site works. A green checkmark in a backup dashboard tells you a file was written. It doesn’t tell you the file is complete, the database imports cleanly, or the site runs on it.
A useful test restore goes like this. Restore the latest backup to a staging server. Load the homepage and a few deep pages. Log in to the admin. Submit a form. If it’s a store, check that recent orders are there. Time the whole thing, because that time is your real RTO, whatever the plan says. Do this a few times a year, and every time the backup setup changes.
What should a website disaster recovery plan include?
A website disaster recovery plan is a short, written list of who does what, in what order, when the site goes down or gets compromised. It should fit on one page and be stored somewhere you can reach when the site and its server are both unavailable.
Here’s a simple runbook to adapt:
- Confirm the problem. Check whether the site is down, defaced, or slow, and whether it’s the site, the server, or DNS. Note the time it started.
- Alert the right people. Contact whoever owns hosting and maintenance, plus one decision-maker on the business side. Use a contact list stored outside the site.
- Contain it. If it’s a security incident, take the site offline or into maintenance mode, and change admin, hosting, and database passwords from a clean device.
- Pick the restore point. Choose the most recent backup from before the problem started. For a hack, that may be older than the last backup.
- Restore to staging first. Bring the backup up somewhere safe, check that it works, and patch the hole that caused the incident before going live.
- Go live and verify. Point the live site at the restored version. Test forms, checkout, and logins, and watch error logs for a few hours.
- Recover lost data. Recreate any orders, entries, or edits made between the restore point and the incident, using payment processor records, email notifications, or form tool logs.
- Write it up. Record what happened, how long recovery took against your RTO, and what you’re changing so it doesn’t repeat.
Step 5 is the one people skip under pressure, and it’s the one that matters most for a hack. Restoring an old copy of a site without closing the hole just resets the clock until the attacker comes back in the same way.
Backups are also one layer of a bigger website security setup. They’re what you fall back on when patching, access control, and monitoring don’t stop something. If you’re building a new site, plan for this from day one. Clean code, a short plugin list, and a documented setup make restores much faster, which is how we approach WordPress development.
The bottom line
A reliable website backup strategy keeps three copies in two kinds of storage with one offsite and locked, runs on a schedule that matches how fast your data changes, and gets tested with real restores. Pair it with a one-page recovery runbook so the bad day is about following steps, not deciding them.
If you want backups, hosting, and security handled under one ongoing plan, see our support and retainer plans, or book a free call to check your current setup. Agencies that need backups and recovery covered for client sites, under the agency’s name, can book a call to talk through white-label support.
Common questions
How often should I back up my website?
Match the schedule to how often your data changes. A brochure site edited monthly can back up monthly, plus before any update. A marketing site edited weekly should back up its database daily. An online store usually needs hourly or continuous database backups, because every gap is lost orders.
Is my web host’s backup enough?
Usually not on its own. Host backups are convenient and quick to restore, but they often sit with the same provider and account as your site. If the account is compromised or the provider has an outage, you can lose both. Keep at least one copy offsite, under separate credentials.
How do you back up a WordPress website?
Back up the files and the database together: the wp-content folder with themes, plugins, and uploads, the wp-config.php file, and a full database export. A backup plugin or your host’s tool can automate this on a schedule. Keep one copy offsite under separate credentials, and test a restore on staging.
What is the difference between backup and disaster recovery?
A backup is a copy of your site’s files, database, uploads, and configuration. Disaster recovery is the plan for using those copies: who acts, in what order, how fast the site must be back, and how you confirm it works. You need backups for disaster recovery, but backups alone aren’t a plan.