Every brokerage has a number: the count of agents it can add before its own back office stops it. Almost none of them know it, because the number does not announce itself. It shows up as a review desk that backs up, an onboarding that slips from days to weeks, and a broker-owner whose calendar quietly converts from recruiting to file work. Most brokerages find their capacity number by hitting it.
The belief worth correcting is that growth problems are recruiting problems. Past a certain agent count, they are arithmetic.
Why it matters: AccountTECH’s Labor Cost Index (July 2025) makes the stakes concrete: profitable brokerages spend 4.7 percent of income on non-agent labor; unprofitable ones spend 8.6 percent. Nearly double. Profitable shops do not support agents less. Their support costs track their transaction volume instead of running ahead of it. The capacity number is where those two lines cross.
Dig in: the number is computable from four inputs a broker already has. Agents, sides per agent per year, hours of back-office work per file, and productive hours per support person per month. Multiply the first two and divide by twelve for monthly files; multiply by hours per file for monthly demand; divide by productive hours for the headcount the operation actually requires. Then add the agents you plan to recruit and run it again. The gap between the two answers is the hire you have not made yet, or the ceiling you are about to hit. As a reference point with its limits stated: in July 2026, Empower’s compliance line ran at 0.97 labor-hours per file across 3,816 files (one month, one work type, not a blended figure), and demand at that rate moves in lockstep with every agent recruited.
Dig in: the ceiling arrives before the math says it should. Support demand is not smooth: closings cluster at month-end, recruiting classes onboard in cohorts, and one person covering review, onboarding and marketing is three ceilings wearing one salary. This is why the same brokerage that “ran fine” at 60 agents describes chaos at 85. The average said there was room; the month-end peak said otherwise. At Empower we review on the order of 12,000 transaction files a month, and the month-end pile is always bigger than the monthly average predicts. The structural version of this failure, and where it bites first, is laid out in the bottleneck test; what it does to the promises made in recruiting is in the staffing-ratio piece.
What to watch (the honest limits): the 0.97 hours-per-file figure is one month of Empower’s own timesheets (July 2026, compliance work only; excluding one implausible entry gives 0.79). It is a reference point for the method, not an industry standard. AccountTECH’s figures are that index’s sample, stated as published. Sides per agent vary enormously by model: the big flat-fee platforms’ filings work out to about 5.2, while the median in NAR’s Member Profile is nine. Run your own number, not either of those.
The short version: the capacity number is the honest version of every growth plan. A brokerage that knows it can time hiring to recruiting instead of to crisis, or move stepwise costs onto a per-file basis so the ceiling moves with volume. That is the choice between building the next seat and buying the next file. A brokerage that does not know it will meet it anyway, at month-end, with a recruit’s first file sitting on the pile. Growing into a second state raises the same arithmetic with a second requirement set attached. That version is in the four approvals.
Also related: the capacity number arrives fastest when the owner is also the reviewer: the most expensive file reviewer in any brokerage is the owner.
Related: Scaling a Brokerage Without Hiring Admin Staff Is Arithmetic, Not a Job Posting
Frequently Asked Questions
How many agents can a brokerage support without adding staff?
It is computable, not generic: agents × sides per agent ÷ 12 gives monthly files; × hours of back-office work per file gives monthly demand; ÷ productive hours per support person gives required headcount. When planned recruiting pushes required headcount past actual headcount, that agent count is the capacity number. Month-end clustering means the practical ceiling sits below the averaged one.
When should a brokerage hire back-office staff?
When the capacity math says the next cohort of recruits pushes demand past current headcount, not when the backlog makes it obvious, because by then the recruits have already experienced the backlog. The alternative to the next hire is moving stepwise costs to a per-file basis, so support capacity scales with volume instead of in salary-sized steps.
What breaks first when a brokerage grows?
The line where demand moves per file but cost moves per person, usually compliance review, because every recruited agent adds files to the same desk, and the desk sits between the agent and their commission check. Onboarding is typically second, because recruiting classes arrive in cohorts rather than smoothly.
