The Real Unit Economics

Figures below are industry benchmarks synthesized from agency-consultancy and financial-modeling sources, not this course's own audited data — every specific range carries its confidence tag, and none of them substitute for building your own model against your own real cost structure once you have one.

---

#The one distinction that makes every other number legible: AGI vs. gross revenue

Gross revenue for a managed-media agency includes the client's media spend passing through the agency's books, even when that spend earns the agency nothing beyond a thin markup. Adjusted Gross Income (AGI) strips out media spend, production costs, and any subcontractor cost passed through at or near zero markup — leaving only the money the agency actually earns for its own labor. [Directional]

This distinction matters because gross revenue badly flatters an agency with large passthrough budgets. An agency billing $2 million/year in gross revenue because it passes $1.8 million of client media spend through its books, keeping $200,000 in actual fees, is a $200,000 AGI business, not a $2 million one — and every margin, staffing, and revenue-per-employee benchmark in this lesson is meaningless unless computed against AGI, not the passthrough-inflated gross number. Every figure below is an AGI-basis figure unless stated otherwise.

#Profit margin, by agency size

Two things worth naming rather than averaging away. First, the two industry-wide figures (20.7% vs. 13%) disagree by more than a rounding error, most likely because they're measuring against different bases (gross revenue vs. AGI, or different survey populations) — this is exactly the AGI-vs-revenue trap the previous section warns about, showing up in the source data itself. Second, and more useful than any single average: specialization is worth roughly 10–20 percentage points of margin across every size tier in this table. A generalist agency ("we do everything for everyone") consistently runs 15–20% while a specialist in one vertical or one platform runs 25–40%. This isn't a coincidence — Module 4's failure-modes lesson traces the mechanism: a generalist has to rebuild expertise for every new client's vertical, while a specialist reuses the same playbook, benchmarks, and creative learnings across every account, which is the cross-client-learning mechanism from Module 2 compounding specifically within a niche rather than diffusing across unrelated ones.

#Cost structure

Labor consumes 50–70% of total agency revenue — the single largest cost line by a wide margin, which is exactly what you'd expect from a services business selling expert time. Overhead runs roughly 30% of AGI — office, tools, insurance, admin, sales/marketing cost to acquire new clients. [Directional] for both.

#Worked example: what the $50,000/month account from Module 3 actually keeps

Module 3 worked through a $50,000/month client at 15% commission, netting the agency $7,500/month ($90,000/year) in fees for that one account. Apply the cost-structure benchmarks above to that revenue line, not to the client's total $50,000 spend (only the $7,500 is AGI — the other $42,500 is passthrough media, which the client benefits from but the agency doesn't keep):

• Labor allocated against this account, at the midpoint of the 50–70% range (60%): $90,000 × 0.60 = $54,000/year — this is the loaded cost of whatever fraction of a media buyer's, strategist's, and account manager's time this one account actually consumes, not a full salary; on a single mid-sized account this is realistically a fraction of one or two people's time, not a dedicated hire.
• Overhead allocated at 30% of the $90,000 AGI: $27,000/year — tools, office/admin, a slice of sales cost to have acquired this client in the first place.
• Remaining margin: $90,000 − $54,000 − $27,000 = $9,000/year, or 10% — near the low end of the 15–20% "healthy" band above, which is the honest lesson here: a single mid-sized account, priced at the standard 15% commission rate, is thin on its own. Real agency profitability comes from spreading fixed overhead (the office, the core tools subscription, the owner's own strategic time) across enough accounts that overhead's share of each account's revenue shrinks — the same operating-leverage logic behind every services business, not something unique to advertising. This is also the direct mechanical reason declining commission percentages at higher spend tiers (Module 3) don't actually hurt margin the way they look like they should: a $500,000/month account at 8% still nets $40,000/month, and the labor required to run it doesn't scale anywhere near proportionally with the tenth account's added complexity the way overhead allocation improves per-dollar.

#Revenue (AGI) per employee

$150,000–$200,000 per employee is the healthy band for most service agencies; media-buying specialists specifically should target $250,000+ per employee given the higher-leverage, less production-heavy nature of the work compared to, say, a creative/design agency. [Directional] UK-specific benchmarks run lower in absolute terms (£100,000–£120,000/head as a healthy 2026 band) — a currency and market-size difference, not a contradiction, but worth knowing if you're benchmarking against a UK-published source. [Directional]

Utilization rate — the share of a delivery employee's paid hours actually billed to client work — targets 75–85% for media-buying agencies specifically. Below that band, you're carrying labor cost the AGI isn't covering; materially above it, you're one sick day or one lost account away from missing deadlines across the board, because there's no slack in the system. [Directional]

#Client churn, by contract type — the number that actually predicts agency survival

Retention, not new-client acquisition, is what separates a durable agency from one perpetually replacing lost revenue. Annual churn varies sharply by how the client is contracted, and the pattern is directionally intuitive once you see it: the more a contract structure ties compensation to a provable outcome, the more exposed the relationship is to any single bad quarter.

By agency size: 11–25 employees, $1M–$5M revenue: ~24% annual churn. 51+ employee full-service agencies: 12–15%, the industry's best retention, driven by deeper account relationships and more diversified service lines that survive any single stakeholder's departure. [Directional]

The mechanism behind the size effect connects directly back to Module 2's trust point: a larger, more established agency has more relationship surface area with the client organization (multiple stakeholders, multiple service lines, a longer track record) — so losing one champion inside the client's org doesn't automatically end the relationship the way it can for a two-person shop whose only relationship is with one marketing director who might leave their job.

48% of clients who switch agencies cite dissatisfaction with delivery as the reason, and this research found a claim that nearly 40% of businesses using an agency are already planning to switch within six months — flagged [Speculative], single-source, and worth treating with real skepticism given how easy that kind of statistic is to generate from a small, unrepresentative survey. The 48% delivery-dissatisfaction figure is more consistently corroborated across sources and worth taking more seriously. [Directional]