After working with dozens of digital agencies — from scrappy teams of five to multi-million pound operations preparing for exit — a pattern keeps repeating itself. Revenue grows. Headcount grows. But EBITDA doesn't keep pace. The founders are working harder than ever and still wondering where the margin went.
The culprit is rarely the obvious stuff. It's not your pricing (though that might need work too). It's not one big problem you can fix in a strategy day. It's a collection of structural leaks — each one individually tolerable, but together quietly undermining the business you're trying to build.
Here are the three we see most often, and why they're so easy to miss.
Scope creep that never gets billed
Every agency has a version of this story: a client asks for something small — "just a quick amend," "a couple of extra rounds," "can you also...?" — and the team delivers it because the relationship matters and the ask feels minor. Multiply that across a 20-person agency with 15 active clients and you're looking at hundreds of hours a year that simply disappear.
The problem isn't that your team is too generous. It's that the systems for tracking, flagging, and converting scope creep into commercial conversations either don't exist or aren't consistently used. Project managers shouldn't have to decide in the moment whether something is billable — that decision should already be made by the process.
A well-run agency with a clear change control process can typically recover 5–10% of revenue it's currently giving away. At a £2m agency, that's £100k–£200k back on the table — before you've changed a single client.
Untracked utilisation and the hidden cost of your senior team
Most agency owners know their headline utilisation figure. Fewer know the utilisation of their senior people specifically — and that distinction matters enormously.
Senior team members carry a disproportionate share of non-billable work: new business, line management, client escalations, internal meetings, pitching. In many agencies we audit, founders and senior leads are running at 40–50% billable utilisation while junior and mid-weight staff sit at 70–80%. The problem is that senior people are also the most expensive — so every hour they spend off-client is costing you significantly more than the same hour from someone else.
This isn't an argument for burning out your senior team. It's an argument for being intentional. Which activities genuinely need a senior person? What can be delegated or systematised? Is new business activity producing a return, or is it speculative time that could be reduced with better lead generation?
Agencies that map utilisation by role grade — rather than just headcount averages — almost always find significant recoverable margin hiding in plain sight.
Retainer clients that have quietly become unprofitable
Retainers feel safe. Predictable monthly revenue, long-term relationships, no constant re-pitching. But retainer profitability erodes in slow motion — and most agencies don't notice until the damage is done.
It typically happens in one of two ways. Either the scope of the retainer has gradually expanded (see Leak 01) without a corresponding fee increase, or the cost to service the client has risen — through salary inflation, tool costs, or the client simply becoming more demanding — while the monthly fee has stayed flat. Clients who've been with you for three or four years are often the worst offenders: they feel like anchors of stability but are frequently your least profitable work.
The fix is straightforward but requires discipline: a quarterly profitability review by client, with a clear methodology for costing hours against retainer value. Most agencies we work with have never done this. When they do it for the first time, they're usually surprised to find that one or two of their "best" clients are running at near-zero or negative margin once all costs are properly allocated.
The commercial conversation that follows — a fee increase, a scope reset, or occasionally a managed exit from a client — is uncomfortable. But it's far less uncomfortable than discovering the problem during a due diligence process.
.webp)
.webp)
