By Craig S. Cody, CPA, Certified Tax Coach. Published August 21, 2026.
Here's the answer up front. There are three numbers that tell you whether a marketing agency is healthy, and they aren't revenue and they aren't headcount. They're people costs, overhead and net profit, each measured as a share of what your agency actually keeps. And here's the part almost nobody says out loud: those three have to add up to 100. So profit isn't a fourth thing you go chase. It's arithmetic. It's whatever the first two didn't already spend.
I've owned a CPA firm for 23 years, and we work with more than 70 agencies every month, many of them Agency Management Institute members. I can usually tell you where an agency stands in about the time it takes to run three divisions. Not because I'm clever. Because the three numbers constrain each other, and once you see two of them you already know the third.
That constraint is the whole reason this check works. It's also the reason most owners never fix their profit line: they treat it as a goal, and a goal is something you can want harder. A residual isn't.
They're the 55/25/20 split, and the denominator is AGI.
Quick vocabulary note, because the term collides with something else and the collision matters. AGI here means agency gross income: your billings minus the pass-through money that was never yours, which is media, freelancers, production, printing, client software. It's not the Adjusted Gross Income line on your 1040. If you've sat in a tax conversation with me, hold those two apart. AGI is the money that's actually yours to run the business on.
Against AGI, the target split is:
One thing to be straight about: that model isn't mine. It comes from Drew McLellan and the Agency Management Institute, and it's published. What I can tell you is that it holds up across the P&Ls that cross my desk every month. The average U.S. agency nets under 10%. The ones running to these three screens land nearer 17% to 20%. That gap isn't talent. It's whether anybody's watching the split.
Because it changes what you're allowed to do about a bad profit number.
Say you're a $1.2 million AGI agency. On target, that's $660,000 to people, $300,000 to overhead, and $240,000 of profit.
Now let payroll drift to 58% and overhead to 28%. Nothing dramatic happened. You added a coordinator you needed, and the software stack grew the way software stacks grow. But 58 plus 28 is 86, so your profit line is 14%, or $168,000.
You just lost $72,000 of profit and every single decision that caused it was defensible.
This is the part I want you to sit with. You can't go get that $72,000 back by focusing on profitability, because profitability was never the input. It was the output. The only two levers that exist are the 55 and the 25.
Because six points of drift doesn't arrive in one piece. It arrives at about half a point a month.
That's not theory. We run these three screens on our clients' numbers every month, not once a year at tax time.
Spread across a year, that $72,000 shows up as roughly $500 a month of slippage against a $100,000 monthly AGI. Nobody catches $500. There's no month where the P&L looks alarming. There's no meeting where someone says "our overhead ratio moved." You find it in March of the following year, when your accountant hands you a return and the profit number is smaller than you remember agreeing to.
Overhead is the worst offender here, and it's worth naming why. Payroll changes are events. You interview, you make an offer, you feel it. Overhead changes are subscriptions. A seat here, a tier upgrade there, an office you signed for more humans than you have now. Nothing ever announces itself, and nothing ever gets cancelled.
A monthly check doesn't have to be sophisticated. Three divisions, three ratios, written down where you can see last month next to this month. What you're looking for isn't the level. It's the direction.
You find out which of the two inputs moved, and you fix that one.
If people costs crossed 55%, the question is almost never "who do we cut." It's whether the work coming in supports the team you built, or whether you're carrying capacity for revenue that left. That's a pricing and a utilization conversation before it's a headcount conversation.
If overhead crossed 25%, go line by line and be honest about what's still earning its place. This is boring work and it's the highest-return hour in your month.
And if both look fine but profit still doesn't, your denominator is probably wrong. That's a real and common problem, and I wrote about it separately: pass-through spend and early billing both distort the base you're dividing by, which makes every ratio on this page lie to you. Fix the base first, then run the screens.
If you want the wider set of numbers you should be able to state cold in a finance conversation, the five numbers you have to be able to defend covers that. This piece is deliberately narrower: three numbers, one equation, once a month.
If you're pre-revenue or under a few hundred thousand of AGI, these ratios will look terrible and that's fine. At that size the owner is the delivery team and the split is meaningless. Watch cash instead.
If you're a solo operator with no payroll, the 55 doesn't apply to you in any useful way. And if your books are on a cash basis and you only look at them at tax time, don't start here. Start with getting a monthly close you trust. Running ratios on numbers you don't believe just gives you confident wrong answers.
Above 20% of AGI is the benchmark from Drew McLellan and AMI. Agencies that genuinely run to the metrics tend to land between 17% and 20%. The U.S. average is under 10%.
No. AGI is billings minus pass-through costs like paid media, freelancers and production. For a lot of agencies that's a difference of 20% to 40% of the top line, which is why benchmarking against revenue gives you a flattering, useless answer.
People costs are humans delivering client work: salaries, payroll taxes, benefits, and contractors filling roles you'd otherwise hire. Overhead is everything else that keeps the doors open. Draw the line once and keep it consistent, because a moving definition is worse than a slightly imperfect one.
Yes, at a reasonable market rate for the work you actually do. Leaving yourself out makes the payroll ratio look healthy and the profit look real when it's really just your unpaid labor.
Ten minutes with a P&L you already have, once a month.
Your profit line isn't a goal you can want your way into. It's the space your payroll and your overhead leave behind, and the only way to protect it is to watch those two every month while the drift is still half a point.
That's how agency owners I work with keep more of what they make, and it's how they flourish through the years when revenue is flat.
If you want more of this kind of thinking on paper, you can request a free copy of my book at the link below.
Request a free copy: https://www.craigcodyandcompany.com/free-book
And if you'd rather just talk it through, let's talk.