How the market actually prices an agency.

For decades, agency valuation ran on a single convenient number: 1.5 to 2 times annual revenue. It was simple, and it was wrong. A $10M agency at a 15% margin and a $10M agency at a 30% margin are not worth the same — yet the revenue rule valued them identically, ignoring margin, carrier mix, retention, and owner dependency. As institutional capital flowed into the sector, that heuristic collapsed. This piece maps what replaced it: normalized EBITDA, the modern multiple bands, and the platform-premium arbitrage that explains why consolidation keeps accelerating.

§ 01 · The death of revenue multiplesWhy profitability won.

The revenue rule produced two structural problems for sellers. The first is the silent discount: an owner of a $10M-revenue agency earning a 25% margin who accepts a "1.5× revenue" offer of $15M is actually accepting roughly an 8× EBITDA multiple — well above the institutional baseline of 4–6× at the time, but only if they realized institutional buyers had already moved to EBITDA. Many didn't, and left value on the table through pure information asymmetry. The second is valuation fog: without the ability to calculate normalized EBITDA or benchmark against comparable deals, an owner simply cannot tell what the agency is worth. Both problems trace to the same root — pricing on the wrong metric.

§ 02 · Normalized EBITDAThe metric buyers actually use.

Institutional buyers value agencies on normalized EBITDA — operating earnings adjusted to reflect the sustainable, ownership-independent cash the business generates. Starting from reported EBITDA, buyers add back owner-specific costs (above-market salary, personal vehicle, owner travel) and non-recurring items (one-time consulting, litigation, an irregular IT build), then normalize lumpy capital spend to an annual average.

The $125,000 adjustment looks small — 6.25% — but at a 6× multiple it is a $750,000 swing in valuation, produced entirely by correct normalization. Buyers demand the normalized figure for two reasons: it lets them compare dozens of agencies on a like-for-like basis, and it tells them the cash the business will actually throw off to service debt once the owner's personal costs are gone.

§ 03 · The multiple bandsWhere agencies trade.

Multiples are not random — they encode market consensus on profitability and growth. Three bands, plus the competitive tail.

The baseline band reflects standard books — flat-to-modest organic growth, typical retention, no carrier concentration. The premium band rewards demonstrable advantage: 5–12% organic growth, 95%+ retention, diversified carriers, documented processes, and a book that doesn't depend on the owner's personal relationships. The 19×+ tail appears almost exclusively in the "kill zone" — agencies at $3M–$10M of revenue where competing PE platforms see immediate tuck-in value and bid each other up.

§ 04 · The platform premiumThe arbitrage beneath the bands.

The most important dynamic in modern agency M&A is the platform premium — the valuation uplift from consolidating agencies into a larger entity. A standalone standard agency trades near 11.2× EBITDA; the same agency inside a platform firm is valued near 14.0× — a spread of roughly 2.8×.

That spread is a built-in arbitrage. A PE platform acquires a standard agency at ~11×, consolidates it (integrating clients, merging back-office, eliminating duplicate overhead), and exits the combined platform at ~14× — creating value without any revenue growth at all. The "kill zone" agencies are the prime fuel: too large for micro-acquirers, too small to be platforms themselves, common in the market, and exactly the right size to tuck in and consolidate up. A platform that buys ten kill-zone agencies at 11× and exits the combined entity at 14× captures the 2.8× spread on every one. That arithmetic — not just growth — is why institutional consolidation keeps accelerating, and why the rate cycle that sets the cost of leverage moves deal volume so directly. The companion deal-structuring piece covers how the price is actually paid.

◆