Ask five people what a WordPress backup costs and you will get five answers, all of them correct for the person giving them. The freelancer with one brochure site says nothing, because a free plugin does the job. The agency owner with thirty client sites says it is a line on the monthly budget she keeps an eye on. A shop owner who once lost a week of orders says it is the cheapest thing he pays for.

The trouble is that the sticker price on a backup plan is only one of several numbers. Storage has a price. So does the time you spend checking that the thing is working. So does the afternoon you spend restoring a site when the backup turns out to be missing its database. Most people only price the first of those, then get surprised by the rest.

This article is a worksheet. It walks through each cost in turn, shows how to do the arithmetic for your own sites, and explains where free tools are perfectly reasonable and where they are not. I have tried to keep the numbers generic, because your site is not mine. Where I use an example, it is an invented round figure to show the method, not a statistic about anyone.

Step one: price the loss before you price the backup
It sounds backwards, but a backup budget only makes sense next to the thing it insures. If losing the site costs you nothing, spending money on backups is silly. If losing it costs a week of income, spending a little on backups is a bargain. So start by writing down what a bad day looks like.

There are really three versions of a bad day, and they cost different amounts.

The site is gone and you have no copy
This is the worst case and the one people picture. You rebuild from scratch. For a simple brochure site that might mean a few days of work and some lost search visibility while pages come back. For a store with years of product data, customer accounts and order history, some of it cannot be rebuilt at all. Order records, for instance, may be something you have a legal reason to keep.

The site is down but a copy exists
This is far more common. A plugin update breaks the front end, a host migration goes wrong, or someone deletes the wrong folder. The cost here is the downtime plus the time it takes to restore. That second part depends heavily on how the backup was taken and where it lives, which is why two sites with identical backup schedules can have very different recovery experiences.

The site is up but quietly wrong
A hacked site, a bad bulk edit, or a database table that got truncated may not look broken. You find out days later. In this case what matters is how far back you can go. A backup that only keeps yesterday's copy is useless if the problem began five days ago, because yesterday's copy contains the problem too.

For each of your sites, write a rough figure for each of these. Do not worry about precision. A single sentence such as "a week of lost leads, about two days of my time" is enough. That is the number every other cost on this page gets compared against.

The five line items of backup cost
Once you know what you are insuring, you can look at what the insurance is made of. Every backup setup, free or paid, has the same five ingredients. They are just paid for in different ways.

1. Storage
Backups take space, and you usually want several copies, not one. Somebody pays for that space. With a paid service it is folded into the plan. With a DIY setup it is a bill from an object storage provider or a cloud drive. With a free plugin it is often your own hosting account, which feels free but uses up the disk quota you are already paying for, and which disappears together with the site if the account is lost.

2. Compute and time on your server
Taking a backup means reading every file and dumping the database. On a quiet dedicated server that is trivial. On cheap shared hosting it can slow the site down or hit a limit and fail. Plugins that run inside WordPress are bound by PHP limits such as max_execution_time and memory_limit, which is the reason big sites on small plans so often produce backups that stop partway through. You do not see this as a charge. You see it as a failure on the one day you needed the file.

3. Transfer
Getting the backup off the server uses bandwidth. Most hosts include a generous amount, so it rarely matters for the upload side. It can matter for the restore side. Some storage providers charge to download data, and some plans cap how much you can pull in a month. If you ever have to restore a large site from a remote bucket, find out in advance whether that costs anything.

4. The tool itself
Plugin licences, a per-site subscription, or a bundle priced by number of sites. This is the number people compare, and it is the easiest one to see. It is also usually the smallest of the five once you count your time honestly.

5. Your hours
Setting it up, checking that jobs ran, rotating credentials, fixing the storage connection that expired, and doing a restore test now and then. If you bill by the hour, price these at your own rate. If you do not, price them at what you would pay somebody to do it. A free setup that takes two hours a month to babysit is not free, and a paid setup that takes ten minutes a month may be cheaper in total.

Keep these five in mind as we go. Every option below is really a different way of moving cost between them.

Storage math you can do on paper
Storage is the part of the budget that most often surprises people, mainly because the numbers multiply. Three things set it: how big the site is, how many restore points you keep, and whether each restore point is a full copy or only the changes.

Measure the site
Start with two sizes. The first is the wp-content folder, which is mostly uploads, themes and plugins. The second is the database. Your host's file manager or a command such as du -sh wp-content over SSH gives you the first. The database size is visible in phpMyAdmin or from wp db size if you use WP-CLI.

Uploads are almost always the larger part, and they have a useful property: old media files rarely change. That matters in a moment.

Multiply by the restore points you keep
Say, purely as an illustration, that a site is 5 GB in total and you keep thirty daily full copies. That is 150 GB. Keep the same thirty days on ten sites of that size and you are at 1.5 TB. Nothing about that is unusual, but it is a long way from the "a few gigabytes" most people assume when they sign up for a free storage tier.

Now compare incremental storage. If only a small fraction of the files change each day, the same thirty restore points might take a little over the size of one full copy plus the daily changes. The database is a different story: it is usually dumped in full each time, because the dump is a text export and not a set of changed rows. For many sites the database is small enough that this does not matter, and for a busy store it is the part that grows.

Decide how far back you really need to go
Retention is where the budget and the risk meet. Seven days is cheap and covers the "I broke it yesterday" case. Thirty days covers the "we got hacked and only noticed last week" case far better. Longer than that is a business decision, often driven by records you are required to keep. A reasonable approach is to keep recent daily copies and a thinner set of older ones, for example one a week for a couple of months. That gives you reach without paying for sixty near-identical copies.

Remember the copy that lives next to the site
A backup stored in the same hosting account as the site is not an offsite backup, even if it is a different folder. If the account is suspended, the server fails, or an attacker with file access deletes everything they can see, the backups go too. The commonly quoted rule of thumb is three copies, on two kinds of storage, with one copy somewhere else entirely. You do not need to follow it to the letter, but a copy that survives losing the host is the part worth paying for.

Free is a real option, sometimes
It would be dishonest to write a cost article that pretended free tools never make sense. For a personal blog, a portfolio or a hobby project, a free backup plugin sending a nightly copy to a cloud drive you control is a perfectly sensible setup. The site is cheap to rebuild, the content may already exist elsewhere, and a day or two of downtime costs nothing.

What makes this work is that you accept the limits. Free plugins that run inside WordPress depend on WP-Cron to start their jobs. WP-Cron is not a real scheduler. It checks for due tasks when someone loads a page, so on a site with very little traffic a nightly backup can slip by hours or skip a day. Many hosts also cap script run time, so a large backup can be cut off. And the storage destination is its own account with its own expiring token and its own quota.

None of this makes free tools bad. It means the cost has moved from the invoice to your attention. The honest test is whether you will actually look. If you are the kind of person who checks that the last backup ran and who restores one to a test site once in a while, free works. If you are not, then the free option has a hidden price, which is that you will find out it failed at the moment you need it.

The DIY middle road
There is a step between a plugin and a managed service: a scheduled script of your own. A real system cron job runs mysqldump or WP-CLI's database export, archives the files, and uploads them to an object storage bucket. This avoids the WP-Cron problem and the PHP limits, and the storage cost is small. The price is that you are now the person who maintains it. Credentials rotate, hosts change PHP versions, a disk fills up, and nobody gets an alert unless you wrote the alerting too. For a person who enjoys that, it is a good deal. For an agency billing out its time, the hours often cost more than a subscription would.

Paying per site, per gigabyte, or per bundle
When you move to a paid service, the way it is priced decides whether it stays cheap as you grow. There are broadly three models, and each rewards a different kind of customer.

Per site
You pay a fixed amount for each website. This is easy to understand and easy to pass on to a client. It works well when your sites are similar in size. It works badly if one of your sites is a huge media library and the rest are small, because the big one is priced like the small ones, or the vendor has a size limit that forces you onto a higher tier.

Per gigabyte
You pay for storage used. This is fair for sites of very different sizes, and it is how raw object storage works. The weakness is that the bill moves with retention. Double the number of restore points and you roughly double the storage. It is worth checking whether the provider charges for downloads as well.

Bundles
A plan covers a set number of sites and a set amount of storage together, and you move up a tier when you outgrow it. This is the most common model for services aimed at freelancers and agencies, and the sensible way to compare bundles is to divide: price over number of sites gives you the per-site cost, and storage over number of sites tells you how much room each site gets.

As an example of what a bundle structure looks like in practice, you can read the published offsite backup pricing for a managed, agentless service. When I checked in November 2026, backupyoursite's page listed three monthly tiers for 3, 10 and 30 websites, each bundled with managed offsite storage and a retention window that gets longer on the higher tiers, with a lower monthly rate if you pay annually. The same page describes a 30-day free trial for one site with no card needed. Prices change, so treat my description as a snapshot and read the page itself before deciding. The point of showing it is the method. Divide the monthly price by the number of sites, then look at what retention and storage you get for that, and you can put any two services on the same footing.

That arithmetic is worth doing because a plan that looks cheaper per month can be dearer per site, and a plan that looks expensive can turn out to be the cheaper choice once you have ten sites on it. It is also worth checking what counts as a "site". Some vendors count each WordPress install, some count each domain, and some count multisite networks differently.

Annual versus monthly
Annual billing is usually cheaper per month, and for a site you intend to keep for years it is the better deal. Monthly billing costs more but lets you stop. For a client site you may lose, monthly can be the safer choice. For your own long-running sites, prepay and save. Either way, check what happens to your stored backups if you stop paying, because some services delete them quickly and some keep them for a grace period.

Where the money actually goes wrong: backups that do not restore
Everything above assumes the backup works. The most expensive backup is the one that looked fine and was not. A backup that cannot be restored has cost you its full price and given you nothing, plus it gave you a false sense of safety, which is worse than knowing you had no backup.

There are a handful of ways this happens, and all of them are boring.

The database dump is incomplete. A dump that was cut off by a timeout may look like a normal file of normal size, until you try to import it and it ends halfway through a table.
Files are missing. Some tools skip large files or certain folders by default to save space. If your uploads folder was excluded, you have a site with no images.
The backup and the site disagree. Files and database taken minutes apart on a busy store can be slightly out of step, so orders in the database refer to products or files that the file copy does not match. Tools differ in how they handle this.
The storage connection quietly died. The job runs, produces an archive, and fails to upload it. Some tools report this clearly and some log it somewhere nobody reads.
The copy is fine but nobody knows how to restore it. This one is nearly always about process rather than technology. If the only person who knows the steps is on holiday, the backup is as good as missing.

The fix is not complicated, but it does take discipline. Do a restore test on a regular schedule. Restore the latest backup to a staging site or a spare subdomain, log in to the admin, load several pages, open a few posts with images, and check that a recent order or form entry is there. Time how long it takes. That time is your real recovery time, and it is the figure to put in the "site is down but a copy exists" line of your loss worksheet. Many people find it is three or four times longer than they had assumed.

A restore test costs an hour or two a quarter. Put that in the "your hours" line of the budget. It is the best-value item on the whole list.

Verification is a feature worth paying for
If you pay for a service, find out whether it checks its own backups. Some tools confirm that an archive was written and stop there. Others check that the archive is readable, that the database dump is complete, or even run a trial restore. This is hard to see from a pricing page, so look at the documentation or ask. Two plans with the same price and the same storage can be very different products on this one point.

A worked budget for three kinds of site
Here is how the worksheet looks for three imaginary sites. The figures are round numbers made up for the example, so swap in your own. The point is the shape of the thinking.

The personal blog
Small, low traffic, nothing lost if it is down for a day. Loss cost: low. A free plugin sending a nightly copy to your own cloud drive covers it. Your hours: ten minutes a month to check the last backup date, plus a restore test once or twice a year. Total cost: close to nothing, which is the right amount for this site.

The small business site
A services company whose website brings in enquiries. A week offline would mean missed leads and a hit to its reputation, so the loss cost is real but the site is not complicated. This is the middle ground where people most often underspend. A managed service or a well-built DIY setup with offsite storage, thirty days of retention, and failure alerts is a sensible fit. The monthly fee is likely to be less than one hour of the owner's time, and the alerts mean nobody has to remember to check.

The store or the agency portfolio
A WooCommerce shop, or an agency looking after a couple of dozen client sites. Here the loss cost is high and mostly not recoverable by rebuilding. You want frequent backups, longer retention, a restore path that does not depend on the site itself being healthy, and some way to roll back a single page, product or table without flattening the whole site. For an agency, per-site cost and the ability to hand a client their own restore are also part of the maths.

If you want a second view of how a tiered budget works across free, DIY and paid tools, there is a useful breakdown of offsite backup costs that lays out the tiers side by side. Read it with the same scepticism you would bring to any vendor's own page, and check the ranges it quotes against live pricing.

Notice what the worksheet does not tell you to do. It does not say buy the most expensive tool. It says match the spending to the cost of losing the site, and count your own time as part of the price.

Questions to ask before you pay anyone
Whatever tool you are looking at, a short list of questions will tell you most of what you need. You can usually answer them from the vendor's documentation without talking to anyone. If you cannot find an answer, that is itself information.

  1. Where do the backups live? Outside your hosting account, on infrastructure that does not share a fate with the site.
  2. What starts the backup? A real scheduler, or WP-Cron on page load?
  3. Does it check the backup? Readability of the archive, completeness of the database, anything beyond "file written".
  4. How long are copies kept, and what happens if I stop paying? Thirty days of history and a grace period are very different from seven days and immediate deletion.
  5. What does a restore look like?One click from a dashboard, or a manual download-and-upload job? Can I restore one folder or one table without rolling back everything?
  6. Who gets told when something fails? An email alert that arrives the same day is worth more than a log file.
  7. What counts as a site? Per install, per domain, per multisite network.
  8. Can I get my data out? Standard archive formats you can open without the vendor's tool are safer than a format only one product can read.
  9. What does a download cost?If you restore a big site, is there a transfer fee or a monthly cap?

Take the two or three tools that pass, work out the per-site cost for your actual number of sites, and then run a free trial or a test restore on a real site before committing. A trial that includes a real restore is worth far more than any feature comparison, because it shows you the thing you are actually buying, which is the experience of getting your site back.

Conclusion
A backup budget is five costs, not one: storage, load on your server, transfer, the tool, and your own hours. Add those up for each site, set the total against what a bad day would cost, and the right amount to spend becomes much clearer. For a hobby site it may be nothing. For a business site it is usually a small monthly fee. For a store or an agency it is a line item worth thinking about carefully.

The one number I would not skimp on is the restore test. It costs almost nothing, and it turns a backup from a hopeful guess into something you know works. Do the arithmetic once, write down your recovery time after the first test, and review both numbers when your site grows.