The unbilled hours
Small unbilled client fixes add up to real money every month, and the fix is removing the possibility, not charging for it or training harder.
Every agency has a number it doesn't track: the hours spent fixing things on sites that were signed off and paid for months ago. Nobody logs them because each one is too small to log.
It's worth actually working out.
The arithmetic
Take a fix that costs ten minutes. It never costs ten minutes. There's the client's message, the context switch out of whatever you were doing, opening the site, working out what changed, fixing it, replying, and getting back to where you were. Twenty-five minutes is closer, and the two context switches are the expensive part.
Say that happens once a month per live client site. An agency with thirty live sites is at roughly twelve hours a month. At a £75/hour blended rate that's £900 a month, or near enough £11,000 a year, spent on work nobody invoices and nobody scheduled.
Run your own numbers. The point isn't our figure, it's that the figure exists and you've never seen it.
Why it doesn't feel like a problem
Because each instance is trivially small and socially awkward to charge for. The client is apologetic. It's ten minutes. You're not going to raise an invoice for ten minutes, and if you did, you'd damage a relationship worth thousands.
So the cost hides. It shows up as evenings, as a vague sense that the team is busier than the project list suggests, and as the thing that makes Friday afternoons unpredictable.
It's worse than the hours
The unbilled hours are the measurable part. Three other costs sit underneath them.
It scales with success. Every site you ship adds to the maintenance load permanently. Your best year produces your worst following year.
It's unschedulable. You can't plan around it, which means it eats buffer — the time that was supposed to absorb overruns on billable work.
It changes the relationship. A client who breaks their site and gets it fixed for free learns that you're free. The next ask is slightly bigger.
What actually causes it
Mostly not the client. Look at what the fixes are, and the pattern is usually: they edited content, and the edit changed the design. Layout broke, image is the wrong shape, heading wraps badly, mobile is wrong.
That happens because they were given a tool that lets them do both. They wanted to add a new location; the tool also offered them the layout.
Three ways out
Charge for it. A maintenance retainer converts the cost into revenue. It works, it's what most agencies do, and it's a real answer — but you're now selling a service whose value proposition is that the site needs fixing.
Train harder. Handover docs, a Loom, a written guide. This works for about a month.
Remove the possibility. Build sites where the content is fully editable and the design isn't — not as a permission, but structurally, so there's no location grid to widen. The client gets what they wanted, which was to add the location. You get a site that can't be broken by someone adding a location.
Start by measuring it
Whatever you do next, spend a month logging it. Every message, every small fix, with a rough time. Most agency owners are surprised by the total, and it's difficult to make a decision about a cost you haven't seen.