How to set brokerage commission splits

Set brokerage splits by working backwards from two numbers you control — your cost to serve an agent and the value you deliver them — rather than forwards from whatever competitors post. A split structure is a price, and like any price it fails in two directions: too rich for your costs and the brokerage quietly loses money per deal; too thin for your value and recruiting stalls while producers drift out. The craft is a ladder that prices each agent segment honestly and a bundle that makes the price feel obviously fair.

Start from cost-to-serve, not folklore

Before designing anything, compute what an agent actually costs you annually: staff time, office and tools provided, insurance, admin, compliance overhead, errors absorbed. That number — which varies wildly between a high-touch mentorship model and a lean virtual one — is your floor. Every plan you publish should clear it at realistic volumes, including the flagship split you advertise. Brokerages that skip this step discover their star recruit class is unprofitable only after building the fleet around it.

Design the ladder, not a number

A single split fits nobody; a ladder prices segments. Workable rungs follow production or tenure: newer agents on richer brokerage shares that fund their heavier support needs, mid-producers stepping up as volume proves out, top producers on caps or near-flat economics that acknowledge how little marginal service they consume. Two design rules keep ladders healthy. Make thresholds objective and published — agents evaluating you model their net anyway, and opacity reads as trickery. And make each rung genuinely attainable; a ladder whose top tier requires volume nobody hits is a marketing claim, not a structure. (Teams inside your brokerage run their own second-layer economics on top of yours — the team split mechanics are a separate design space worth understanding as their landlord.)

Caps, fees, and the arithmetic of honesty

The same revenue can be shaped as percentage, cap, or flat fees, and sophisticated agents translate between them instantly. What they punish is disguise: an advertised split that nets far less after stacked transaction fees, desk charges, and mandatory add-ons. Publish the all-in reality — or better, structure so the all-in and the headline are close. Fee opacity wins exactly one negotiation per agent, and then becomes the story they tell about why they left.

Bundle value the split visibly pays for

The strongest defence of any split is what it demonstrably includes. Providing the working stack — website, CRM, lead capture and routing — moves real costs off the agent’s personal budget and into the brokerage’s bundle, where scale pricing favours you: platform seat economics (BrokerFold’s brokerage plan carries twelve seats, with extra seats flat-priced) cost the brokerage far less than the same tools cost thirty agents retail. A recruit comparing your slightly richer take against a “high-split” shop where everything is bring-your-own can see your math; programs like brokerage-provided agent websites exist to make that comparison vivid. Bundled value also survives copycat pricing — a competitor can match your percentage overnight, but not the infrastructure behind it.

Illustrate before you publish

Stress-test the final design with worked examples across personas — a first-year agent at modest volume, a mid-producer, a capped top performer — and confirm each nets sensibly for both sides at realistic and pessimistic volumes. Then publish those examples in your recruiting materials. A split structure you can show working, agent by agent, in plain arithmetic is a recruiting asset; one that requires a meeting to decode is a warning label. And keep the worked examples current as fees and bundles evolve — an outdated example circulating in recruiting materials is a promise you will be held to.