The 83% threshold is empirical, not theoretical — it's what buyers actually see in diligence: agencies that dip under it tend to stay there until something structural changes, because the operational drag of replacing that many clients every year is almost impossible to outrun through organic growth. The symptoms cluster: retention under 83% for two or more years, new business that only offsets churn rather than growing the book, service capacity consumed by onboarding, and producer pay tilted to new business with no reward for keeping clients.
§ 01 · Why small agencies leakPersonal effort, not infrastructure.
The pattern concentrates under $500K in revenue for a structural reason. At that scale the agency runs lean — one owner-producer, a CSR or two — with no formal renewal program, no retention dashboard, no cross-sell motion built into the service cycle. Retention happens because the owner personally knows the clients, which means it fails the moment the owner is stretched, a long-tenured CSR leaves, or the book outgrows the owner's ability to touch every renewal. A $3M agency retains through infrastructure — dedicated account managers, a service manager watching retention by book, renewal workflows that run without heroics. Small agencies retain through effort, so any disruption to that effort drops the number fast.
The treadmill effect: servicing an existing client costs roughly a third of the effort of onboarding a new one. An agency losing 20% of its book a year converts about 60% of its service capacity into onboarding — which is exactly the capacity it would need to fix the leak. The drag prevents the investment that would stop it.
§ 02 · What it does to valueA haircut, then harder terms.
Buyers don't apply a footnote — they apply a multiple reduction, and it's large. A healthy small, personal-lines-heavy book might trade at 5.5–6.5× EBITDA; the same revenue and margin with sub-83% retention typically trades 0.5–1.0× lower. On a $300K EBITDA book that's $150,000 to $300,000 of consideration gone, purely on the retention number. And the structure tightens below the line.
| Lever | Healthy book | Leaky bucket |
|---|---|---|
| Multiple | 5.5–6.5× EBITDA | 0.5–1.0× lower |
| Earn-out share | 10–20% | 25–40% |
| Earn-out metric | Revenue / EBITDA | Policy retention at 12 & 24 mo |
| Seller transition | 6–12 mo | 18–24 mo |
None of these terms are punitive — they're rational reactions to measured risk. But together they shift value from cash-at-close to contingent consideration and make the seller's outcome depend on what the book does after the owner walks away. The transition plan is where it bites hardest: an absent seller in a leaky-bucket agency is the single biggest post-close risk a buyer faces, so the extended period gets tied to the earn-out — no presence, no payout.
§ 03 · Operational or structuralThe diagnostic that decides everything.
Not every sub-83% agency is the same, and buyers who walk past all of them miss real opportunities. An operational leak looks like an absent renewal process, producer comp that ignores retention, and a book that drifted monoline — all fixable by a buyer with better infrastructure, which makes a 0.5–1.0× discount accretive from day one for a strategic acquirer whose fixes transfer faster than the discount decays. A structural leak looks different: the book concentrated in a re-rating market, primary carriers non-renewing, the relationship-owning producer retiring immediately, or a community reputation problem. Those aren't capacity problems, and no operational tightening fixes them. Ask the diagnostic question early — what is causing this leak? Operational, and the discount is a buying opportunity; structural, and the discount is a warning.
§ 04 · The 18-month fixThree moves that add 3–5 points.
For a seller — or a buyer modeling the upside — the number is more malleable than it feels; three or four points in 12–18 months is realistic, and every point converts to multiple. Install a renewal discipline: a 90-day-out touch on every active client, a carrier-shopping review for any premium up more than 8%, a 30-day confirmation touch — agencies that build this from scratch typically see a 2–3 point lift in a year just from reduced passive churn. Cross-sell to lift policies per client: single-policy clients retain far worse than multi-policy ones, so an umbrella onto a home-and-auto account materially increases stickiness. Give retention an owner: the reason it stays invisible is that nobody's job says "retention" — assign it, report it monthly, tie some comp to it, and attention alone closes most of the gap. None are exotic; they're operational hygiene the treadmill prevents, which is what the 18-month runway is for. For a seller, the expensive move is hiding the leak — the discount a buyer applies after catching a surprise is always worse than the one priced into the LOI.
◆
Terminology on this shelf
- Leaky Bucket Syndrome
- Policy-count retention below 83%, where churn outruns responsible new-business replacement.
- The treadmill effect
- Replacing lost clients consumes the service capacity that would otherwise prevent the leak — a self-reinforcing drain.
- Multiple haircut
- The 0.5–1.0× EBITDA reduction buyers apply to a sub-83% book.
- Retention-based earn-out
- Contingent consideration tied to policy retention at 12 and 24 months post-close.
- Operational vs structural leak
- Operational leaks (process, comp, monoline drift) are fixable and priceable; structural leaks (market, carrier, owner) usually aren't.
- Renewal discipline
- A standing 90-day / 30-day renewal cadence — the single highest-ROI retention fix in a small agency.