Most agencies fail because the people running them never built an accurate mental model of what kind of business they are actually in. The creative is a craft. The agency is a machine, and like any machine, it obeys laws whether or not you understand them.
What follows are the mental models that experienced operators carry around in their heads. None of them are about making better ads or prettier decks. They are about the underlying economics, incentives, and constraints that decide whether a talented shop becomes a durable business or a treadmill that exhausts everyone on it. Think of them less as tips and more as physics: forces you can work with or get crushed by, but not ignore.
The engine: you are selling time, and that's the problem
Strip away the brand strategy and the showreel and a traditional agency is a remarkably simple machine. It buys people's time at one price and sells it at a higher one. Everything else, the awards, the culture, the case studies, is in service of widening that spread.
Once you see the business this way, the master metric becomes obvious: utilisation, the share of your team's available hours that you can actually bill. The gap between a team running at 60% and one running at 80% is the gap between a struggling agency and a healthy one, and it is almost entirely invisible day to day. Nobody walks into the office and announces that utilisation slipped four points last month. It just quietly shows up in the bank account a quarter later.
Sitting on top of utilisation is the model David Maister made famous in Managing the Professional Service Firm: the leverage pyramid, or "finders, minders, grinders". Profit doesn't come from senior people doing brilliant work. It comes from senior people winning and overseeing work that more junior, less expensive people execute. An agency where the founders and creative directors are still in the weeds producing deliverables is, structurally, a low-margin business no matter how good the output is. You can be the most talented shop in your city and still be poor, because you've built a pyramid with no base.
This leads directly to the trap that catches almost everyone: the capacity trap. Because revenue is welded to hours, the only obvious way to grow is to add headcount. But headcount growth is linear, fragile, and expensive. Every new person is a fixed cost you have to feed with billable work, and the moment the pipeline hiccups, your beautiful growth becomes your beautiful liability. Agencies plateau not because they run out of demand but because they run out of a way to grow that isn't "hire more humans and hope".
Pricing: the hourly model is quietly working against you
Here is the most uncomfortable idea in the entire business, and the one most agencies refuse to sit with: billing by the hour inverts your incentives. You make more money by being slow, inexperienced, and inefficient. You make less money every time you get faster, smarter, or better at your craft. An agency that has spent fifteen years getting genuinely excellent at something is penalised by the hourly model for that excellence, because the work that used to take a week now takes a day.
The escape, argued most forcefully by Blair Enns, is value-based pricing: price the outcome to the client rather than the cost to you. The client does not actually want forty hours of design. They want a rebrand that lets them raise prices, or a campaign that moves a number. Price to that, and your speed and skill become assets instead of liabilities.
Whatever pricing model you use, you have to understand the difference between your rate card and your realisation. The rate card is the fiction you put in proposals. Realisation is what you actually collect after discounts, write-offs, over-servicing, and the hours nobody dared to bill. Plenty of agencies that look profitable on a spreadsheet are silently leaking 20 to 30 percent of their revenue in the space between those two numbers.
The biggest single leak is scope creep: the slow, polite expansion of work beyond what the statement of work covered, which somehow never gets re-priced. It rarely arrives as one big ask. It arrives as "just one more round", "can you also look at the social versions", "quick favour before Friday". Each one is too small to fight about. Together they are the most common silent killer of agency margin there is. The discipline of re-contracting when scope changes is unglamorous and worth more than almost any new-business win.
New business: the feast, the famine, and the pitch
Agencies live inside a cruel rhythm. When you are busy delivering great work, you stop selling. The pipeline dries up. The delivery ends, and suddenly there is a gap with nothing behind it, so you scramble, sell hard, win, get busy, and stop selling again. This feast-or-famine cycle is not a sign of a badly run agency; it is the default state of the business. Smoothing it is a permanent act of management, not a problem you solve once.
The cycle is made worse by bad pitch economics. The competitive pitch (speculative, unpaid, low win-rate, and expensive to mount) is one of the worst customer-acquisition mechanisms in any industry. An agency that grows primarily by pitching is running a far worse business than one that grows by reputation and referral, because the latter arrives pre-sold, with the price-shopping already over. The goal is to be chosen, not to compete.
The shape of your revenue matters as much as the amount. Retainers versus projects is a foundational distinction: a retainer gives you predictable recurring revenue and lets you plan capacity with confidence, while project work is lumpy, unpredictable, and forces you to resell yourself constantly. The healthiest agencies tilt deliberately towards retained relationships and towards land-and-expand: growing a small first engagement into a large account over time. Expanding an existing, trusting client is dramatically cheaper than winning a new logo, and yet most agencies pour their energy into the new logo because it's more exciting.
Positioning and the client relationship
The most counterintuitive model in the business is that specialisation beats generalisation, what David C. Baker calls building a position of expertise. Every instinct screams the opposite: narrowing your focus feels like turning away revenue. But a narrow, credible specialisation makes you the obvious choice rather than one of nine interchangeable options, lets you charge a premium for genuine expertise, and shortens the sales cycle because you no longer have to convince anyone you understand their world. The generalist competes on price and relationships. The specialist competes on being the right answer.
Underneath every engagement is a truth that's easy to forget when you're in love with your own creative: clients are buying a reduction in their own risk. A marketing director who hires you is, on some level, protecting their own position, betting their credibility on you not making them look foolish. This is why the relationship with the account lead so often matters more than the work itself, and why trust, responsiveness, and the feeling of being in safe hands routinely beat raw talent in the buying decision. You are not selling creativity. You are selling confidence.
Two structural constraints shape the entire game and are worth naming explicitly. The first is conflict of interest: you generally can't serve direct competitors, which means within any given category your addressable market is capped, since winning the biggest player in a sector often locks you out of all the others. The second is the procurement squeeze. As clients professionalise how they buy, creative and marketing services get pushed towards commoditisation, lined up in spreadsheets, and beaten down on price. The only durable defence is to have differentiated yourself so clearly that you're no longer in the comparison at all, which loops right back to specialisation.
The escape hatch: productise, or run the treadmill forever
If you hold all of these models at once, a single conclusion starts to assemble itself. The agency business, in its traditional form, is a machine that converts human hours into money, and almost every force acting on it (the capacity trap, the hourly incentive inversion, the feast-or-famine pipeline, the procurement squeeze) pushes margins down and stress up.
The way out is productisation: turning bespoke, hours-based service into repeatable, fixed-scope offerings, and eventually into intellectual property or software that earns money without a person burning an hour for every rand it makes. A productised offering can be sold the same way twice, priced on value, delivered with leverage, and marketed on a clear specialisation. It breaks the iron link between revenue and headcount that traps everyone else. It is the difference between owning a business and owning a job that happens to employ other people.
None of this makes the craft matter less. Great work is still the reason anyone calls you in the first place. But great work is the cost of entry, not the source of durability. The agencies that last are the ones whose founders understood that they were running two businesses at once, a creative practice and an economic machine, and refused to let their love of the first blind them to the physics of the second.
