Leaky-bucket attrition — when retention compounds down.

Retention is the pre-acquisition analogue of post-close attrition — it's the historical number a buyer reads as the most reliable predictor of how the acquired book will behave. And because retention compounds, small differences in the current rate produce large differences in the book's value over a holding period, which is exactly why a buyer can't treat a retention gap as a rounding error. Below a clear threshold, a low rate signals not bad luck but a structural problem the buyer would inherit.

§ 01 · Three retention bandsWhere the discount starts.

Retention sorts a book into three bands. Above 90% is top tier — durable relationships, a working service model, and compounding in the buyer's favor, earning premium multiples. The 83%–90% band is industry-standard, drawing typical market multiples and standard attrition assumptions. And below 83% is the leaky-bucket band, which earns valuation discounts, conservative pro-forma modeling, and retention-linked earnouts. The 83% threshold is the meaningful line: it's where retention stops looking like normal variance and starts looking like an operational or cultural problem — a book that can't hold 83% is leaking faster than a healthy service model should allow. The customer-diligence framing of retention as a metric is in leaky-bucket syndrome; here the focus is what the band does to the synergy model and the price.

§ 02 · The compounding mathWhy the discount surprises owners.

The compounding is the part that catches sellers off guard. Two agencies with nearly identical current-year financials can be worth very different amounts if their retention rates differ, because the rate determines the book's trajectory, not just its current size. A 5-point retention gap compounds into a 16-point book-size gap over five years — 92% retention keeps 66% of the book at year 5, while 87% keeps only 50% — and a 15-point gap (95% versus 80%) becomes a 2× book-size delta over ten years. Any valuation that capitalizes projected cash flows has to reflect that divergence, which is why a retention-based discount is larger than an owner expects: they're comparing this year's revenue, while the buyer is comparing the present value of two very different ten-year trajectories. The retention number, in other words, isn't a footnote to the valuation — it's a primary driver of it.

§ 03 · Five drivers and the DD readWhat moves retention, and how to measure it.

Five drivers explain a book's retention performance, and reading them tells a buyer whether a low rate is fixable. Service-model design (stratified per client tier), producer-versus-principal relationships (a producer-owned book retains better through a sale than a selling-principal-owned one), carrier stability and product fit (a book on a retrenching carrier shows deterioration unrelated to service quality), multi-line penetration (cross-sell clients retain materially better), and communication cadence (proactive reviews versus reactive-only). The DD discipline has three layers. Measure historical retention on a policy-count and a revenue basis separately — 88% policy retention with 83% revenue retention means the agency is losing its larger accounts. Segment the retention trend by account type, line, and producer — a portfolio 85% often hides 92% and 77% subsegments. And project forward with an explicit transition adjustment: even a strong book experiences a 2–3 point dip below trend for one to two years post-close before returning to its underlying rate. That two-to-three-point post-acquisition dip is the modeling rule, and it's the bridge to the attrition forecast in client attrition.

§ 04 · Remediable or structuralAnd why you can't bury it.

The decisive question for a low-retention book is whether the cause is remediable or structural. A 12–18 month remediation window works when the deterioration has identifiable causes — a specific weak producer, a specific carrier issue, a specific service-model failure can take 80% retention to 85%–90%, and the resulting valuation uplift typically exceeds the opportunity cost. But structural retention weakness is not remediable: a book where clients have shifted to a segment the agency can't serve effectively, or where the service economics have been squeezed past the point of repair, won't recover with effort — and there, honest disclosure wins while remediation is wasted time. For a seller, burying the retention number is never viable, because a buyer calculates it independently from the raw book data, and discovering an undisclosed gap reshapes the deal more negatively than the underlying number would have on its own. Retention is the same metric the post-close discipline in client-retention work manages, read here as the pre-acquisition predictor — measure it both ways, segment it, diagnose remediable versus structural, and price the band honestly. That's how a buyer keeps a leaky bucket from draining the deal's value after close.

◆