The seller's revenue number is an assertion until three independent sources agree on it. Triangulation is the discipline of making them agree — and where they don't, the gap is the most valuable thing diligence produces, because a normalized, defensible revenue figure agreed before the price is set is what makes the rest of the deal hold.
§ 01 · The three sourcesBilled, paid, received.
| Source | Measures | Blind spot |
|---|---|---|
| Management system | What was billed (policies, expected commission, rates) | Masks downstream payment problems |
| Carrier statements | What was paid | Carrier-timed and classified; misses fees and direct bills |
| Bank deposits | What actually arrived | Unlabeled — requires reconstruction |
Each is real and each is incomplete: the system is closest to intent, the carrier statements closest to economic reality, the bank closest to truth but hardest to read. A clean agency reconciles to within 1–3% across all three — usually pure timing, commission earned in December and paid in January. Pick one time basis (cash received is standard) and normalize the other two forward or back to it, because timing differences alone produce several points of apparent variance that evaporate once you align the basis.
§ 02 · SamplingTop-20 weighted plus a random tail.
Pulling every carrier statement is tedious — a larger agency has 10 to 30 carriers in PDFs, spreadsheets, and portal exports — so sample in two layers. A top-20 weighted sample gives 100% coverage of the top 20 clients by commission, where concentration risk lives and each error carries the most valuation weight. A random tail sample of 25 to 50 additional policies, stratified by line and carrier, surfaces the systematic problems the weighted sample misses — misconfigurations, ghost policies, entire missing categories. Together they cover 40–60% of revenue in most small agencies. At the LOI stage, the top five carriers plus a tail sample is acceptable; the comprehensive pull is the pre-closing confirmatory standard, over a trailing 12 months aligned to a quarter- or year-end close.
§ 03 · Reading the discrepanciesFive patterns.
When the three sources diverge, the shape of the divergence is the diagnosis. AMS commission earned that doesn't match the carrier statement paid is a rate mismatch — innocent if a rate change wasn't updated, serious if it's sustained over-reporting. Policies in the system with no carrier-statement entry are ghost policies. Carrier paid but no bank deposit points to timing, a producer receiving personally, a diverted account, or an uncaptured direct-bill arrangement. Bank deposits the system doesn't classify are usually fee income — real revenue, but it trades at a different multiple. And contingency in the system that doesn't match the 1099 is misclassified contingency income, which inflates the multiple if it's treated as base commission.
The ghost-policy signature is specific: the system shows it, no carrier statement confirms it, and no bank deposit corresponds to the expected commission. Removing ghost policies often takes several points off EBITDA — which makes this the primary retrade moment in a customer-side review.
§ 04 · When to escalateQuality of Earnings and the seller's pre-listing audit.
Escalate to a third-party Quality of Earnings review when the deal exceeds about $5M of enterprise value or the first-pass triangulation surfaces a variance over 3–5% of revenue; the firm cost typically runs $15K–$75K for a small-agency deal and most lenders require it above certain debt thresholds. On the other side, a seller who runs the same triangulation 6–12 months before listing gets three things: errors fixed before a buyer finds them, a normalized EBITDA that needs no adjustment to defend, and clean reconciliation workpapers that let a buyer's analyst verify rather than reconstruct — which compresses diligence and friction. The framing matters throughout: this isn't a fraud hunt, which is rare. It's the production of one revenue number both sides agree on before anyone names a price.
◆
Terminology on this shelf
- Triangulation
- Reconciling the management system (billed), carrier statements (paid), and bank deposits (received) to one defensible number.
- Reconciling variance
- The gap across the three sources — 1–3% is clean (timing); larger is the finding.
- Time-basis normalization
- Aligning all three sources to one basis (usually cash received) so timing noise isn't read as a discrepancy.
- Ghost policy
- A policy in the system with no carrier-statement or bank-deposit confirmation — a primary retrade finding.
- Top-20 weighted sample
- Full coverage of the top 20 clients by commission, paired with a stratified random tail.
- Quality of Earnings escalation
- A third-party review triggered above ~$5M or a 3–5% variance, typically $15K–$75K.