Why Brands Fire Agencies That Are Winning - and How We Priced Around the 'Agency Tax'

The best way to get fired by a brand is to do such a good job that they stop being able to feel what you do for them.

That sounds backwards, so let me say it plainly. Agencies don't usually get cut for failing. The failing ones get cut quickly and nobody agonises over it. The painful firings, the ones that surprise people, happen to agencies that are quietly winning. The work is good, the numbers are up, and one day the brand looks at the invoice, looks at how smoothly everything is running, and thinks "do we even still need these guys?"

I've watched it happen to good operators, and I've thought hard about why, because we run an agency and I'd rather understand the thing than pretend it doesn't exist. Here's my honest take on the whole outsource-insource cycle, when in-housing actually makes sense for a seven-figure brand, and the one structural change that stops the relationship dying by accident.

The 'agency tax' is a feeling, not a number

Every brand runs an outsource-insource loop whether they name it or not. Early on, outsourcing is obvious. You're doing $80k a month, you can't justify a full-time creative strategist and a media buyer and an editor on payroll, so you hire an agency that already has all three. Completely the right call.

Then you grow. And somewhere on the way up, the fee stops feeling like an investment and starts feeling like a tax.

Nothing actually changed about the work. What changed is the comparison happening in the founder's head. When an agency is bringing a brand to life and finding the messaging that connects, the fee feels cheap. When the account is humming along and the founder can suddenly imagine doing it in-house, the same fee feels like a toll they're paying out of habit. That's the tax feeling. It's a story about value, not a line on a spreadsheet.

And here's the uncomfortable bit. The better the agency is at making things look effortless, the easier it is for the brand to underrate the effort. Smooth is dangerous. If a founder can't see the work, they assume there isn't much of it.

Why "they're crushing it" becomes a reason to cut

There's a version of this that's almost funny if it weren't so common. A brand keeps an agency around partly out of loyalty, sees the agency doing well, posting about wins, and concludes: they clearly don't need us, so we can walk away without it hurting them.

It's a bit daft when you write it down, but the logic is real. People find it far easier to cut you when they think it won't be painful for you. So an agency loudly flexing how well it's doing is, in effect, advertising that it's a safe one to drop.

The flip side matters too, and I want to be fair here. If an agency is genuinely crushing it and clearly bringing the value, "they're making good money off us" is not a reason to fire them. Two hundred hours a year that generates a serious return is a bargain, and good on them. The toxic combination is the one to watch for: a partner you don't feel is adding much, who's also telling everyone how well they're doing. That's the recipe to get cut, and fairly so.

When in-housing genuinely makes sense

So let's get to the real question a seven-figure founder is actually asking. Not "are agencies good or bad," which is a silly debate, but "should this function be inside my walls now?"

My honest view: there's no such thing as a good agency or a bad agency in the abstract. There are only good people, and good people exist on both sides of the line. So the decision isn't agency versus in-house as a tribe. It's a few specific questions.

• Is this need now constant, or still spiky? Agencies are brilliant at consistency and at things you need a lot of, predictably. The moment your needs start changing shape every quarter, a new angle here, a new content style there, you'll feel yourself asking the same partner to do different things and getting frustrated. That friction is usually the signal that the work wants to live in-house, where saying something once is enough.
• Have you got enough volume to keep a specialist busy and growing? A full-time media buyer or editor only makes sense if there's a full-time job's worth of work and a path for them. Under that, you're paying for idle time and they're getting bored.
• Do you actually have the management capacity? This is the one founders skip. In-house isn't free just because there's no agency fee. You're now hiring, training, retaining and managing a person, with all the messiness that comes with humans. An agency you can hold accountable from the outside. An employee you have to lead.
• Are you in-housing to save money, or to get something better? Saving money alone is a weak reason and usually a false economy once you count your own time. Getting closer to the work, faster iteration, deeper brand knowledge that compounds, those are good reasons.

Here's where I'll be straight, even though it's against my own interest. For a lot of brands, the smart end state for a few core functions is in-house. Bring the day-to-day that you need constantly and can manage well inside. The right move is often to keep an outside partner for the things you need occasionally, for fresh perspective, and for the specialist work that doesn't justify a hire.

That last point is the one people miss. An outside team that doesn't live in your business every day walks in with a perspective you simply cannot have from the inside. You collect biases over time. You stop being able to see your own brand. A good external partner breaks that, especially on creative. So the question usually isn't "agency or in-house," it's "which parts of this should sit where."

How we priced and report around the tax

If the tax is a feeling that builds when value goes invisible, the fix is structural: never let the value go invisible. Don't let the relationship coast on inertia until one day it snaps. Make the brand re-decide on value, out loud, on a schedule.

So a few things we do deliberately.

We say the quiet part at the start. Every relationship has an arc, and it will end one day, and that's fine. I'd rather tell a client up front that there may come a point where in-housing a function is the right move for them, and that when we get there I'll say so. Naming it kills the weird loyalty dynamic where neither side wants to raise the obvious. It also means nobody feels trapped, which, oddly, is what makes people stay.

We re-evaluate the value every quarter, not by vibes but on paper. Once a quarter we put the relationship back on the table: here's what we did, here's what it drove, here's what it cost, and here's whether the maths still says outsource. If a function has matured to the point where in-housing it would genuinely serve the brand better, we'll be the ones to flag it. Losing a slice of scope while keeping trust beats clinging to scope and losing the client.

And we report so the work stays visible. The whole tax problem comes from effort disappearing behind smooth results. So our reporting is built to keep showing the engine, not just the dashboard number, what was tested, what was learned, where the next gains sit, what we'd do with more budget and what we'd cut. A founder should never have to wonder what they're paying for, because it's in front of them every month.

None of that is a retention trick. It's just refusing to let a good relationship die of inertia, which is how most of them actually die.

The question worth asking your own setup

So if you're a founder staring at an agency invoice and feeling that first flicker of "is this still worth it," I'd resist the urge to make it a yes-or-no, keep-them-or-cut-them call. Break it down instead.

Pull the last two quarters. For each function your agency runs, ask the three honest questions: is the need constant or spiky now, have I got the volume and the management capacity to do it better in-house, and am I trying to save money or get something genuinely better. Then look at whether you can actually see the value, or whether it's just gone quiet because everything's running smoothly.

Run that and the answer usually sorts itself, function by function, instead of in one emotional swing. And if you want to see what re-deciding on value every quarter actually looks like on paper, the way we lay it out so the maths is always in the open, reply and I'll walk you through the reporting model we use. It's the same view we give clients, no strings.