Commission Splits and Cap Models: How Agents Keep More (2026)

Ask most agents what their split is and they will give you a number. Ask what they actually keep and the room goes quiet. The commission split is the most discussed and least understood number in a real estate career, because the headline percentage is only the first line of a longer equation that includes caps, desk fees, franchise fees, and per-transaction charges. Two agents on identical splits can take home wildly different amounts, and an agent who chooses a brokerage on the split alone is choosing on incomplete information. Understanding the full model, and matching it to your real volume, is one of the highest-leverage financial decisions in the business.

How Traditional Splits Work and Climb

At a traditional brokerage, you and the firm divide each commission by an agreed percentage. New and developing agents commonly start near a 50/50 or 60/40 split in the brokerage favor, in exchange for training, brand, leads, and support. As production rises, the split improves on a tier schedule toward 70/30, 80/20, or 90/10. The logic is that a productive agent needs less hand-holding and brings the brokerage more total dollars even at a smaller percentage, so the firm trades a bigger cut of each deal for a bigger pile of deals.

The trap is treating the split as the whole cost. A generous split paired with a heavy fee stack can net less than a modest split with none. This is why the split conversation belongs next to your GCI and pipeline planning: your split only matters in proportion to the gross commission income flowing through it, and the right split at 12 deals a year is frequently the wrong one at 40. Model the percentage against your real annual volume, not against the version of your business you hope to have someday.

The Cap Model: Why High Producers Love It

A cap is the maximum total the brokerage will collect from your splits in an anniversary year. Keller Williams popularized the capped model, and eXp Realty uses a widely published $16,000 cap. The mechanics are simple and powerful: you pay your normal split on each deal until the brokerage portion reaches the cap, and from that point you keep 100% of your commission for the rest of the year, minus small per-transaction fees. For a high producer, the back half of the year is dramatically more profitable than the front half.

The cap reframes the whole season. An agent on a $16,000 cap who closes enough early deals to cap by midyear keeps nearly everything from July forward, which makes the marginal value of a late-year closing far higher than an early-year one. This is exactly why capped models reward volume and punish inconsistency: an agent who never caps pays the split on every single deal and captures none of the upside. Before choosing a capped brokerage, benchmark your production honestly. A realtor performance benchmark against transaction-volume norms tells you whether you will realistically cap, which is the entire question the model turns on.

100 Percent and Flat-Fee Models

At the far end sit 100% commission and flat-fee brokerages, which charge a monthly desk fee or a fixed per-transaction fee instead of taking a percentage split. A high producer keeps almost everything; the brokerage makes its money on predictable fees rather than a slice of each deal. The trade is that these models typically provide little to no lead generation, marketing, or support, so you fund your own pipeline. That makes them ideal for established agents who already have a strong sphere and referral system and a website generating owned leads, and poor for new agents who still need the brokerage to feed them.

The 100% model is essentially a bet on yourself. You exchange the brokerage split for a fixed cost and take full responsibility for generating business. For an agent with a proven, self-sufficient pipeline, that bet pays off enormously; for an agent still dependent on brokerage leads, the fixed fees become overhead with nothing behind them. The deciding factor is not ambition but current self-sufficiency: can you reliably produce business without the brokerage handing it to you? If you are funding your own lead generation, embedding tools like a mortgage calculator on your own site turns that marketing spend into an exclusive, owned pipeline that justifies the flat-fee structure.

The Fee Stack That Erases Your Split

The reason split-shopping fails is the fee stack hiding beneath the percentage. Beyond the split itself, brokerages may charge a monthly desk fee, a franchise fee (often a percentage of commission off the top before the split, common at franchised brands), a per-transaction fee, technology fees, errors-and-omissions insurance fees, and more. A 90/10 split with a $1,000 monthly desk fee and a $300 per-deal fee can cost more than a 70/30 split with none, depending entirely on your volume. The percentage is marketing; the fee stack is the price.

The only honest method is to build the total-cost comparison. For each brokerage you are considering, take your real trailing-twelve-month production and compute total annual brokerage cost: split dollars plus every fee, accounting for the cap if there is one. Then compare net take-home across models on identical volume. This is the same net-after-everything discipline that separates listing-side and buyer-side economics, where the headline commission and the take-home are very different numbers. Do the full math and the right model is usually obvious, and frequently not the one with the most attractive split on the recruiting flyer.

A Worked Example: One Agent, Three Models

Put real numbers on the three models and the volume-dependence stops being abstract. Take an agent who closes 24 deals in a year and earns an average gross commission of $12,000 per deal, the same per-deal figure the gross-commission example above uses. That is $288,000 of total gross commission income flowing through whichever model they pick, and the only question that matters is how much of it survives to take-home.

Start with a traditional 70/30 split and no fee stack. On the post's own arithmetic, a $12,000 gross commission on a 70/30 split yields $8,400 to the agent, so 24 deals return $201,600. The brokerage keeps the other 30%, which is $86,400 across the year. Clean and simple, and for an agent who genuinely uses the brand, leads, and training, that 30% buys something. The weakness only shows up when the volume climbs and the agent stops needing what the split is paying for.

Now run the same 24 deals through a capped model on the eXp Realty $16,000 cap. Until the brokerage's share reaches $16,000 the agent pays a split; after that they keep effectively all of it. Suppose this agent pays an 80/20 split on the way to the cap: the brokerage collects 20% of each $12,000 commission, which is $2,400 per deal, so the cap is reached after roughly seven deals (7 times $2,400 is $16,800, just past the $16,000 ceiling). For those first seven deals the agent keeps 80%, about $67,200. For the remaining seventeen deals they keep essentially 100% minus small per-transaction fees, roughly $204,000. Total take-home lands near $271,200 before those small fees, well above the $201,600 the flat 70/30 returned, and the entire gap is the upside the cap unlocked once the brokerage stopped taking its share.

Finally weigh the fee stack the post warns about. Layer a $1,000 monthly desk fee and a $300 per-transaction fee onto a generous 90/10 model. The agent keeps 90% of $288,000, which is $259,200, but then pays $12,000 in annual desk fees plus $7,200 in per-deal fees across 24 closings, leaving about $240,000. Still strong at this volume, but notice the fee stack quietly removed $19,200, more than a full point and a half of the entire year's gross. Drop the same agent to 8 deals instead of 24 and that fixed $12,000 desk fee suddenly eats a punishing share of a much smaller pie, which is exactly why the headline 90/10 is a trap at low volume and a gift at high volume. The numbers move with production, never on the percentage alone.

Match the Model to Your Real Volume

There is no universally best commission model; there is only the model that fits your current production. At low and inconsistent volume, a higher split with no fees and real brokerage support frequently nets the most, because you benefit from the leads and training and never reach a cap anyway. At high and steady volume, a capped or 100% model wins decisively, because you capture the upside the split would have taken. The break-even moves as your business grows, so the right answer at one stage becomes the wrong answer at the next, which is why this is a decision to revisit annually rather than once.

Treat your commission model as a number you manage, not a default you inherited. Run your real production against each option once a year, the way you would re-underwrite any major cost. The agents who keep the most are rarely the ones who found the most generous split; they are the ones who matched the model to their actual volume and revisited the fit as that volume changed. The owned-lead systems that make the high-self-sufficiency models viable are laid out on the lead generation tools for real estate agents page, because the more of your own pipeline you build, the more freedom you have to choose the model that keeps the most.

Revenue Share and Equity: The Newer Variable

A category that has reshaped brokerage choice deserves its own line in the comparison: revenue share and equity models. Cloud-based brokerages such as eXp Realty, alongside others built on similar structures, pair a capped commission model with two additional levers, a revenue-share stream paid to agents who recruit other productive agents, and stock or equity awards tied to production and tenure. The pitch is that an agent builds a second income that is not strictly their own transactions, plus an ownership stake, on top of keeping most of their commission after the cap. For agents who genuinely build a downline or value the equity, this can add real long-run value the headline split never captures.

The discipline is to value these components honestly rather than on the recruiting presentation. Revenue share is only meaningful if you actually attract and retain productive agents; for most agents who never recruit, it is effectively zero and should be modeled that way. Equity is real but illiquid and variable, so it belongs in the comparison as upside, not as guaranteed take-home you can spend this year. The sound method is the same total-cost-against-production math from the fee-stack section: model the commission economics on their own first, then treat revenue share and equity as separate, probability-weighted additions rather than letting them paper over a split or fee structure that does not stand up on its own.

Splits Are Often Negotiable

Agents frequently treat a brokerage's posted split as a fixed price, but for a producing agent it is often a starting point. Brokerages compete for agents who bring volume, because a productive agent generates revenue even at a favorable split and lends credibility to the office. An agent with a track record and a portable book of business has leverage, particularly when a competing brokerage's total-cost offer is on the table as a credible alternative. The negotiation is not only about the split percentage either: a cap level, specific fees, technology costs, and transaction charges are all potentially movable, and a concession on the fee stack can matter more to take-home than the headline percentage.

The way to negotiate from strength is to arrive with the total-cost math already done. An agent who can show their real trailing-twelve-month production and a precise total-annual-cost comparison across brokerages is negotiating with evidence, not asking for a favor. This is where benchmarking your own numbers pays off directly: an agent performance benchmark that documents your transaction volume and production substantiates the value you bring to the table. The agents who keep the most are not only the ones who pick the right model; they are the ones who negotiate the terms of that model from a position of demonstrated production rather than accepting the recruiting flyer's numbers as final.