Book quality is the most consequential dimension of customer DD. The headline revenue is the easy number; what produces the revenue — and whether it persists — is the work. Five measurements structure the buyer's diligence. Each catches a different aspect of book health, and together they distinguish a durable revenue line from a fragile one carried by short-term factors.
Two numbers that sellers conflate.
Retention is the most-cited book-quality metric, and the most often misrepresented. Two distinct measurements matter.
Policy retention asks what percentage of policies in force at the start of a period renewed during the period. A book with 1,000 policies in force and 920 renewals has 92% policy retention. The number reflects client-relationship durability — clients who stayed.
Premium retention asks what percentage of in-force premium renewed during the period. A book with $5M in force and $4.85M renewed has 97% premium retention. The number reflects revenue durability — money that stayed.
The two numbers diverge because of premium changes on renewing policies. In a hard market, premium retention often exceeds policy retention by 3–5 percentage points: clients renewed but at higher premium. In a soft market or with shrinking accounts, premium retention often falls below policy retention. Sellers in hard markets emphasize premium retention; the buyer's diligence asks for both.
Two retention numbers, both calculated correctly, both disclosed. Anything less is a diligence gap, not a calculation difference.
Policies per client as a quality signal.
Account density — the average number of policies per client — predicts retention quality. Three patterns produce different post-close outcomes.
- Single-policy clients. Average 1.0–1.2 policies per client. These are typically commodity personal-lines books (auto, home) where each policy is a separate transactional relationship. Retention rates are lower because switching costs are minimal.
- Bundled clients. Average 1.5–2.5 policies per client. Auto-home bundles, commercial-lines combinations, life-and-health add-ons. Retention rates run 3–7 percentage points higher than single-policy books because the switching cost (multiple policies, multiple carriers, multiple coverage analyses) is real.
- Deep accounts. Average 3.0+ policies per client. Commercial-lines specialists with full programs (property, casualty, umbrella, professional liability, cyber, EPL). Retention rates are highest; client lifetime value is highest; producer engagement is highest.
The buyer's diligence calculates account density from the AMS report and segments retention by density tier. A 92% headline retention rate that's 95% on deep accounts and 87% on single-policy clients is a different book than the same 92% headline rate that's 90% on deep accounts and 93% on single-policy clients. Same headline; different post-close retention assumptions.
Concentration that matters.
The 15% rule: no single client should account for more than 15% of agency revenue. Above 15%, single-client departure creates a material EBITDA hit that disproportionately damages the buyer's pro-forma. Multiple haircuts apply: 0.5×–1.0× EBITDA per concentration band above the threshold.
The existential dozen: the top 12 accounts as a percentage of revenue. The 12-account threshold is empirically derived — it's the rough size of a producer's manageable top-account portfolio in a typical commercial-lines book. The aggregate exposure matters operationally.
Single-client threshold.
- No single client > 15% of revenue.
- 15–25%: 0.5× multiple haircut.
- 25%+: 1.0× multiple haircut, possible deal restructure.
Top-12 aggregate exposure.
- Top 12 < 25% of revenue: diversified book.
- 25–40%: producer-dependent book.
- 40%+: integration risk material; retention bonus targeting essential.
Earnout against retention.
- 10% earnout against top-12 retention.
- Aligns seller / producer / buyer incentives.
- Converts concentration risk into structural protection.
When new business hides churn.
The leaky-bucket pattern is the most under-detected book-quality finding. An agency reports 7% top-line revenue growth — looks healthy. The decomposition reveals 4% rate, 3% retention growth, 0% net new business. But the gross new business is 12%, and the lapse rate is 12%. The net new business is zero because the bucket is leaking.
The leaky bucket signals operational fragility. The agency is working hard to acquire new business, but the same effort is required just to replace churn. Producer capacity is consumed on replacement rather than growth. Customer-acquisition cost is high, customer-lifetime-value is depressed, and the post-close pro-forma based on gross new-business velocity is wrong.
The diligence catches the pattern by decomposing top-line growth into rate, retention, and new-business components — both gross and net. Sellers report headline numbers; the buyer's diligence digs into the source data.
The five measurements — policy retention, premium retention, account density, concentration, the leaky-bucket diagnostic — together complete the retention and quality layer. They feed the book-defensibility analysis (the next layer) and the verification work (the third layer in the cluster). The Pillar — Customer Due Diligence for Buyers — covers the broader framework.