The Agency's Own KPIs and Kill Switches

This lesson is deliberately about the agency's own operating health, not the client-facing marketing metrics (ROAS, MER, CAC) a media buyer reports to a client — those are covered implicitly throughout Modules 2–3 as the language of the work itself. What follows is the dashboard a founder should be watching about their own business.

---

#The core operating metrics, gathered in one place

Every figure below was introduced with its source in Modules 3–4; this lesson exists to put them together as a single running dashboard rather than leaving them scattered across the modules that derived them.

All confidence tags are as stated where each figure was first introduced in Modules 3–4; treat this table as a consolidated reference, not a new independent claim.

#Client concentration — the risk that kills otherwise-healthy agencies

This is worth its own section because it's the single most common structural cause of an agency's collapse that has nothing to do with the quality of its work. Published concentration-risk benchmarks:

• Healthy: top client under 15% of revenue; top 5 clients under 40%; top 10 clients under 60%.
• Moderate risk: top client 15–25%; top 5 clients 40–55%.
• Material risk: top client 25–40% — this level of concentration is severe enough that it triggers a 1x–2x EBITDA valuation discount in a business sale or acquisition context, because a buyer is explicitly pricing in the risk that the relationship doesn't survive the sale.
• Structural risk / act now: any client consistently above 20% of revenue for three or more consecutive months; if the top 5 clients exceed 50% of total revenue, a single departure can trigger a genuine cash-flow crisis, not just a bad quarter.

[Directional] across this whole scale — consistent across agency-valuation and financial-benchmarking sources, though the exact percentage breakpoints vary slightly by source.

The kill-switch version of this: track top-client-as-percentage-of-revenue monthly, not annually — concentration creeps up gradually as one account grows faster than the rest of the book, and by the time it shows up in an annual review it's often already at the material-risk threshold. The action at each threshold isn't "fire the client" — it's actively grow the rest of the book faster, and treat any single new large account with real caution about how fast you let it become the majority of your revenue.

#The utilization kill-switch, in both directions

Utilization below 75% means the agency is carrying labor cost the AGI isn't covering — a genuine margin problem, and the first place to look when margin data (Module 4) is running below the healthy band. Utilization materially above 85% is a different, less obvious problem: it means there's no slack in the system, so a single sick day, a single lost account's transition period, or a single new-business pitch consuming senior staff time creates immediate delivery risk across every other account. The kill-switch isn't just "hire when utilization crosses 85%" — it's recognizing that sustained high utilization is itself a leading indicator of the delivery-quality problems that Module 6 showed drive 48% of client churn, before that churn shows up in the retention numbers.

#Payback-period logic applied to the agency's own client-acquisition cost

The client-facing world talks about CAC-to-LTV ratios and payback periods for the client's own customers. The same logic applies one level up, to the agency's own cost of acquiring a client — and it's worth computing explicitly rather than assuming new business is free once the pipeline is running. If landing a real AOR-tier client costs real founder or sales time over the 2–4 month cycle from Module 5, plus any pitch-specific costs (case-study production, travel, a working-session deliverable), that cost should be weighed against the account's own margin contribution (Module 4's worked example) over a realistic retention horizon (Module 4's churn figures) — an account likely to churn within a year, at 18–33% typical annual churn, has to earn back its acquisition cost faster than a policy of assuming every account stays for years would suggest.

#The honest summary kill-switch list

• Top client above 20% of revenue for 3+ consecutive months → actively diversify, don't just monitor.
• Utilization sustained below 75% for a full quarter → a real staffing or pipeline problem, not noise.
• Utilization sustained above 85% for a full quarter → a real delivery-risk problem, address before it shows up as churn.
• AGI margin on a specific account below roughly 10% (Module 4's worked $50,000/month example landed here) → renegotiate the rate, restructure the scope, or treat the account as a loss-leader deliberately rather than by accident.
• Client churn running meaningfully above the 18–33% band for your contract type → the problem is very likely delivery quality (the 48% figure from Module 6), not bad luck — audit account health directly rather than assuming it will regress to the mean.