Why transforming placement is the next — and largest — lever of brokerage margin expansion.
Herald | April 2026
The best-run US insurance brokerages have reached an economic ceiling. After three decades of technology adoption, offshore outsourcing, PE-backed consolidation, and operational discipline, the industry’s most efficient firms have converged on a single cost structure: roughly 50% compensation, 15% other operating expenses, and 35% EBITDA margin. No firm has sustainably broken through.
The productivity gains have been real. Revenue per headcount has grown 2–3x over the past 30 years. But compensation has held at ~50% of revenue even at frontier firms. Productivity rose, and so did the cost of talent.
Two questions now define the industry’s next chapter. What unlocks the next wave of growth? And how can that growth be captured at a new margin frontier?
AI answers both by transforming placement — the document-heavy process of matching risks to carriers that sits at the intersection of sales and operations.
Placement is the only function where productivity gains compound into both cost efficiency and revenue growth. A 10x improvement in placement productivity — together with natural knock-on effects like BPO elimination and commission optimization — creates a credible path to 50% EBITDA margins, a level never achieved in the industry’s history.
The history of the last 30 years of brokerage evolution shows us how we got to the frontier today, and why AI makes possible the path to break through it.
Every wave of technology adoption has delivered a step-change in brokerage profitability and productivity. The pattern is remarkably consistent.
| Era | Revenue per employee | Public broker adj. EBITDA margin | Private broker pro forma EBITDA (Reagan GPS median / Best Practices) |
|---|---|---|---|
| Pre-1995 (Paper & Fax) | ~$110–130K | ~15–18% (est.) | ~12–16% (est.) |
| 1995–2005 (Internet & Email) | ~$150–175K | ~18–22% (est.) | ~15–19% (est.) |
| 2005–2015 (Cloud, SaaS & BPO) | ~$175–215K | ~22–28% | ~17–20% / ~20–23% (from 2008) |
| 2015–2025 (Digital Platforms & Analytics) | ~$230–295K | ~27–35% | ~23%+ / ~26% |
Public broker margins from SEC filings; pre-2005 are estimates. Private broker margins are pro forma EBITDA with owner compensation normalized — not directly comparable to public figures.
Each era drove revenue per employee up roughly 50% from the pre-1995 baseline and expanded margins by several points. The defining shifts of each era:
Pre-1995 (Paper & Fax). Brokerages ran on mainframe terminals, fax machines, and typewriters. Agency management systems existed but were expensive, DOS-based, and confined to accounting. Client relationships depended on face-to-face meetings and phone calls. The defining changes of this era were organizational — mega-mergers among top firms — rather than technological.
1995–2005 (Internet & Email). Email displaced fax and mail, and rudimentary portals emerged. Adoption was uneven. The more consequential change was regulatory: the 2004 Spitzer investigation into contingent commissions forced transparency in broker compensation and accelerated the shift toward fee-based advisory models.
2005–2015 (Cloud, SaaS & BPO). Cloud-based agency management systems, digital document management, and early analytics reshaped daily operations. Equally important was the rise of offshore BPO: large brokerages moved policy checking, certificate issuance, data entry, and endorsement processing to low-cost centers in India, the Philippines, and Eastern Europe. By the mid-2010s, offshore operations were standard across the top 50 brokerages.
2015–2025 (Digital Platforms & Analytics). Digital placement platforms, centralized operating models, and data analytics reached maturity. The most disciplined firms hit the ~35% EBITDA ceiling during this period. Revenue per employee accelerated sharply, with the industry-wide average reaching nearly $296K by 2025.1
| Company | ~2000 rev/employee | ~2015 rev/employee | ~2024 rev/employee |
|---|---|---|---|
| Marsh McLennan | $179K | $217K | ~$280K |
| Aon | $145K | $170K | ~$268K |
| Arthur J. Gallagher | $140K | $245K | ~$290K |
| Brown & Brown | $174K | $211K | ~$276K |
Source: SEC filings, company reports.
A note on growth. The industry did not just become more efficient — it grew substantially.2 Organic growth hit a record 10.3% median in 2023,3 fueled by a sustained hard market. Top brokers scaled from single-digit billions to $10–25B enterprises through organic growth and aggressive M&A. More than 10,000 brokerage transactions have closed since 2008,4 and PE-backed firms now account for 70–87% of all deals.5
Well-run brokerages converge on a remarkably consistent cost structure: roughly 50% of revenue to compensation and benefits, 15% to other operating expenses, and the remaining 35% flowing through as EBITDA.
Some firms reached this frontier decades ago. Others arrived more recently through consolidation, centralization, and operational discipline. None has sustainably broken through. 50/15/35 is the ceiling — public and private brokerages alike.
SEC filings make the pattern explicit:
| Firm | Comp ratio trend (% of rev) | Other opex trend (% of rev) | Margin trend |
|---|---|---|---|
| Brown & Brown | Flat ~49–52% for 25 years | Already low (~13–16%) throughout | Already at frontier (~33–39%) — minimal expansion |
| Gallagher (brokerage) | Flat ~51–54% for 20+ years | Compressed from ~19–24% → ~13–15% | Expanded from ~23–27% → ~35% as opex caught up |
| Aon | Compressed from ~55–62% → ~52–53% | Compressed from ~22–24% → ~13–17% | Expanded from ~14–23% → ~31–34% |
Source: SEC 10-K filings, 2000–2024.
The story is one the industry already knows. Every firm that expanded margins over the past two decades did so by compressing opex toward ~15%. Once there, expansion stopped. The compensation line held at ~50%. Firms arrived at different times. They all arrived at the same place.67
Other operating expenses — occupancy, technology, BPO vendor fees, travel, professional services — have already compressed from ~20–25% to ~13–15% at frontier firms.8 Three decades of technology adoption, offshore outsourcing, COVID-driven real estate rationalization, and PE-backed operational discipline have done the work.
Little room remains to squeeze. The residual ~14–15% is largely fixed or semi-fixed: lease obligations, software licenses, compliance costs, E&O insurance, and essential infrastructure. Further margin expansion cannot come from this line.
The ~50% compensation line is not monolithic. It breaks down as follows:
| Employee category | % of headcount | % of payroll | What they do |
|---|---|---|---|
| Production (producers) | ~21–25% | ~40–45% | Selling, client relationships, new business, renewals |
| Service staff (AMs, AEs, processors) | ~55–60% | ~35–40% | Placement, post-policy servicing |
| Support (admin, HR, finance, IT) | ~10–12% | ~10–12% | Back office infrastructure |
| Executive & management | ~5–8% | ~10–15% | Leadership, strategy, oversight |
Source: MarshBerry 2024 Compensation Study.
Production compensation (~40–45% of payroll, ~20–22% of revenue) will not compress. Producer pay is commission-based by design — it moves with revenue. As revenue grows, producer comp grows proportionally. This is structural and desirable: producers are the revenue engine.9
Support and executive compensation (~20–27% of payroll, ~10–14% of revenue) is a small target. These are essential functions with limited headcount. Incremental efficiencies exist, but the absolute dollar opportunity is modest.
Service staff compensation (~35–40% of payroll, ~19–20% of revenue) is the opportunity. Service staff are the largest employee group by headcount and perform the most improvable work in the brokerage. The category splits into two distinct functions — with very different potential.
Service work at a brokerage divides into two functions.
Placement (~half of service cost, ~10% of revenue). Matching risks to markets, preparing submissions, formatting applications, shopping to carriers, chasing quotes, comparing terms, negotiating with underwriters. This is document- and process-heavy work that consumes significant hours but sits at the intersection of sales and operations.
Post-policy servicing (~half of service cost, ~10% of revenue). Certificate issuance, endorsement processing, policy checking, renewal preparation, data entry, billing support. This is routine, process-oriented work.
Both present productivity opportunities. Placement is the far more compelling target, because gains there reduce cost and drive revenue growth. Placement is part of the sales motion. A faster, more accurate, more comprehensive placement process delivers:
Post-policy servicing improvements are pure cost savings — valuable, but one-dimensional. Placement improvements are both a cost lever and a growth lever, making them the highest-impact focus area.
Placement is also exceptionally well-suited to AI. The majority of placement activity — data gathering, document preparation, carrier appetite matching, application formatting, quote chasing, and comparison — is structured, repetitive, and document-heavy. The human judgment that matters most — understanding a client’s risk profile, navigating carrier relationships, negotiating bespoke terms — represents a small fraction of total placement time. AI automates the former; humans focus entirely on the latter.
If placement-related service costs represent ~10% of revenue at a frontier firm, the margin impact of productivity improvements is significant:
| Placement productivity improvement | Placement cost (% of rev) | Margin impact |
|---|---|---|
| 1.4x (43% improvement) | 7.0% | +3 pts |
| 2.5x | 4.0% | +6 pts |
| 5x | 2.0% | +8 pts |
| 10x | 1.0% | +9 pts |
For context: the entire brokerage industry gained approximately 6 points of margin expansion over the last 15 years — across every technology and operational lever combined. A 2.5x improvement in placement productivity alone would match that. From a single function.
A 10x improvement — achievable if AI automates the ~90% of placement work that is data gathering, document preparation, carrier matching, and administrative follow-up rather than human judgment — would deliver 9 points of margin from placement alone. That exceeds any single era’s margin expansion in the industry’s history.
10x sounds aggressive. But practitioners themselves consistently estimate that ~80% or more of their day-to-day work is low-value and repetitive — data gathering, application intake, policy review, client follow-up. Compressing that layer is well within reach on a near-term horizon.
“In placement, the mundane stuff, that’s 80% of our entire team’s time. 20% we actually spend on broker and carrier relationships — our real value add.”
— Head of Specialty, Enterprise Brokerage
Transforming placement extends well beyond service staff productivity. Two knock-on effects further expand margins.
Placement-related BPO becomes unnecessary. A meaningful share of offshore outsourcing spend — estimated at ~1–2% of revenue — supports placement activities: submission processing, data entry, document preparation. This cost sits in the other operating expenses line, not compensation. When AI handles the work directly, the BPO vendor cost disappears. That adds ~1–1.5 points of margin on top of service productivity gains — and it comes from the opex line the industry already considers optimized.
Average commission per policy increases. An AI-standardized placement workflow enables brokerages to systematically route placements toward partner carriers with higher commission rates. This is not about pushing clients into inferior coverage. It is about ensuring panel carriers with top-tier products and strategic relationships receive the right flow of business. If average commission across the book rises from 14% to ~15.3% of premium — a 1.3 percentage point uplift — the resulting revenue growth on an essentially fixed cost base delivers ~4–4.5 additional margin points.10
Taken together, placement transformation creates a credible path to 50% EBITDA margin — a level never achieved in the industry’s history:
| Scenario | EBITDA margin |
|---|---|
| Today (frontier firm) | ~35% |
| + Placement productivity (10x) | ~44% |
| + Placement BPO elimination | ~45.5% |
| + Avg commission growth (14% → 15.3%) | ~50% |
All three effects flow from a single strategic investment: transforming placement. The productivity gain is the primary driver. BPO elimination and commission optimization are natural consequences of the same transformation — not separate initiatives requiring separate investment theses.
As AI automates placement, brokerages gain deeper visibility into — and control over — which carrier partners they prioritize, driving compounding commission gains over time.
The financial impact of moving from today’s margins to 50% EBITDA is substantial at every scale of brokerage:
| Firm size (revenue) | Current EBITDA (at current margin) | EBITDA at 50% | Incremental EBITDA / year |
|---|---|---|---|
| $50M (private, 26% margin) | $13M | $25M | +$12M/year |
| $100M (private, 26% margin) | $26M | $50M | +$24M/year |
| $1B (large/public, 33% margin) | $330M | $500M | +$170M/year |
The insurance brokerage industry has spent three decades optimizing its way to a 35% EBITDA ceiling. Every available lever — technology, BPO, consolidation, operational discipline — has been pulled. Opex is compressed. The compensation ratio has held at ~50% through every prior technology wave.
For the first time, a new lever is available. Placement productivity drives cost efficiency and revenue growth simultaneously — a combination no prior wave offered.
AI-enabled placement has the potential to break through the 50/15 frontier that has defined brokerage economics for decades. A 10x improvement in placement productivity — combined with the natural knock-on effects of BPO elimination and average commission growth — creates a credible path to 50% EBITDA margins. That would represent the largest margin expansion in the industry’s history, driven by a single transformation in how insurance is placed.
Herald builds AI solutions for brokerages. Visit our website at www.heraldapi.com to learn more.
Enter your email to continue reading — we'll also send you a PDF copy.
