If your Drupal site is maintained by an agency, you're probably on a retainer. Ten hours a month, maybe twenty. A monthly fee somewhere between a few hundred and a couple of thousand pounds. It feels like proper cover — a team on hand, updates handled, nothing to worry about.
It's worth understanding what's actually happening to those hours.
How the hours get used
A straightforward Drupal security update — the kind that lands a few times a month — takes a competent developer around one to two hours from start to finish. That includes pulling a fresh copy of the database, running the update locally, checking nothing has broken, pushing to staging, and deploying to the live site.
Agencies typically bill three to four hours for the same job.
The extra time isn't fabricated exactly — it's just where the money goes. Someone has to read the initial advisory, create a ticket, assign it, update the ticket when it's done, write a brief summary for the monthly report, and have a project manager check it off. None of those people touched the site. All of them appear on your invoice.
At £65–£95 an hour on a retainer rate, a single security update can quietly consume £200–£380 of your monthly allocation. If there are two or three advisories that month — which is not unusual — the hours are gone before anything else gets looked at.
What happens to the hours that aren't used
Most retainer agreements have a clause about unused hours. They roll over (up to a cap), lapse at the end of the month, or get redirected to other work at the agency's discretion.
Organisations without a developer in-house rarely consume their full allocation. There's no internal pressure to push work through. Nobody is tracking whether the hours are being used well. The site works. Everything seems fine.
Unused hours that lapse are, straightforwardly, money the agency keeps for work it didn't do. Hours that roll over create a nominal credit that rarely gets spent. Hours redirected to "feature work" might be useful, or might just be the agency moving resource around to suit itself.
The retainer model works well for organisations that actively push work through it. For everyone else, it functions more like a standing order than a service.
Charities are particularly exposed
A charity with a Drupal site is often in an uncomfortable position. The site was built by an agency, probably grant-funded, and handed over when the project closed. Someone signed a maintenance retainer at the same time, because it seemed sensible. There's no technical person internally who knows what that retainer covers, or whether anything is actually being done.
What makes this worse is that many charity websites aren't just informational. They have donation forms, event booking, membership systems — payment modules that handle real financial transactions. A Drupal site processing donations is handling payment data. If that site is running on unpatched software, the exposure isn't just reputational. It's a potential data breach, with everything that implies for the people who trusted the charity with their card details.
The retainer they're paying for should be the thing that prevents this. Too often it's the thing that makes them feel like it's being prevented, which is not the same thing.
The question worth asking your agency
It's a simple one: how many hours did last month's security updates actually take, and how were the remaining hours used?
A good agency will answer this without hesitation. The report will be specific — update X took Y hours, here's what was tested, here's what else was done with the remaining time.
If the answer is vague, or if the monthly report is a list of updates applied with no hour breakdown, that tells you something.
Why the model persists
Retainers aren't inherently dishonest. For organisations with ongoing development work — new features, integrations, content support — a block of hours can make sense. The agency has resource allocated, the client gets a predictable cost.
The problem is that "maintenance retainer" and "development retainer" have become the same product. An organisation that just needs its security updates handled gets sold the same contract as a business actively building new features. The hours look the same on paper. The value is very different.
What maintenance actually requires
A Drupal security update — the core kind, the ones that patch vulnerabilities in Drupal's official security advisories — follows a consistent process. Run the update in a local environment isolated from the live site. Run the database update scripts. Export the configuration changes. Test manually and with any automated tests. Push to staging. Deploy to production.
That process doesn't require a project manager, a ticket system, or a monthly account review. It requires a developer who knows Drupal and a workflow that doesn't cut corners.
The overhead that inflates the hours isn't protecting your site. It's covering the agency's cost of having a team.
There's also the matter of contrib modules — the add-ons your site uses on top of Drupal core. These need security updates too, and they're often where vulnerabilities actually live. A maintenance service that only watches core isn't watching everything.
A different way to think about it
Maintenance doesn't need to be a block of hours you may or may not consume. It can be a fixed monthly cost for a defined scope: security updates applied, tested, and deployed. No unused hours. No mystery about where the time went. No retainer balance to track.
If you're not sure what your current arrangement actually covers, a free site check will tell you where things stand — what version Drupal is on, which modules have outstanding updates, and whether the security picture matches what you're paying for.