Splits only hurt when you close. Fixed fees hurt precisely when you don't. Dry months are where brokerage models diverge most — and where agents get wrecked.
Published July 2026 · figures in USD, modeled
Every brokerage bill is a mix of variable cost (split, royalty, transaction fees — you pay when you close) and fixed cost (monthly fees, desk fees, annual fees, plus every tool you had to buy — you pay no matter what). Variable cost scales with success. Fixed cost scales with time.
That distinction is invisible in a good year and brutal in a bad one. An agent who closes nothing for three months pays $0 in splits everywhere. What they owe in fixed fees ranges from $0 to $5,400 depending on the model.
Here's the fixed bleed for one quarter of zero production, using our modeled mid-range figures (tools the brokerage doesn't include are counted, since those subscriptions don't pause either):
| Model | Fixed / quarter | What's bleeding |
|---|---|---|
| RE/MAX (desk fee) | ~$5,430 | $1,500/mo desk fee + own CRM/site/tools |
| Century 21 | ~$1,455 | Monthly fee + own tools |
| eXp Realty | ~$255 | $85/mo cloud fee (tools included) |
| HomeSmart | ~$1,005 | $65/mo + own full tool stack |
| Real Broker | ~$188 | Annual fee prorated only — no monthly fee, tools included |
| Local independent | ~$930 | No brokerage fees, but your whole tool stack |
Quarterly figures = (monthly + desk fees) ×3 + annual fee ÷4 + out-of-pocket tools ÷4, standard needs (CRM, IDX site, TM, e-sign). Run your own months in the simulator.
A 95/5 desk-fee model is the best deal in real estate if you close every month — and the worst if you stall. At $250K GCI the RE/MAX agent's high fixed base is diluted to irrelevance. At $60K GCI with a slow spring, the same desk fee can consume 30%+ of gross income. The desk-fee decision is a bet on your own consistency, and new agents systematically overestimate theirs.
Three structural features cushion dry months: zero monthly fees (Real, most flat-fee plans — the brokerage only earns when you do), included tool stacks (your CRM and website don't become a second rent), and passive income layers — a revenue-share check that arrives whether or not you closed is the only brokerage income that's truly seasonality-proof. The rev share article covers how realistic that layer is.
Ask one question: "What do I owe you in a month where I close nothing?" Then model a realistic bad year — 8 or 9 producing months, not 12 — in the career simulator and compare the same bad year across models in the compare engine. If a brokerage only pencils in your best-case year, it's not a fit; it's a wager.