By Keith Dunham, Founder & CEO of Empower Transactions. Keith built Empower after running operations at HomeCity Real Estate, where he helped scale the brokerage to hundreds of closings a month before its sale to Better Homes & Gardens Real Estate. He also co-founded Opcity, later acquired by Realtor.com.
Twenty-one percent of National Association of Realtors members work on a team, which makes the team the most visible answer to how an agent scales. NAR’s 2026 Member Profile describes those teams as averaging four members, and team-based brokerage specialists reported a median of 32 transaction sides in 2025 — a figure we read, as HousingWire does, as the team’s production rather than one member’s. Divide the second number by the first and the team closes eight sides a head. The median individual agent, working alone and counted in a different survey cut, closed nine.
Read that as a productivity statistic and teams look like a bad trade. Read it as a job description and it is the whole point. A team does not exist to make four people each sell more than one person would. It exists to move two of an agent’s three jobs onto other desks so the agent can do the third at volume — and the first job to move should rarely be selling, for the three reasons below.
Why it matters — the agent at 15 or 20 sides deciding whether to build a team is really deciding which of three jobs to stop doing, and in what order. Getting the order wrong means hiring a producer before the paperwork is off your desk, which is how a team of four ends up closing barely more per head than one agent does alone.
What are the three jobs of a real estate agent?
Every agent does three distinct jobs. The first is finding clients: prospecting, marketing, database, referrals. The second is serving them: pricing, showing, negotiating, advising — the licensed work, and the work the client is actually paying for. The third is closing the file: deadlines, disclosures, signatures, document routing, the closing checklist. Every side an agent closes carries all three, and the third one is the same on a $300,000 file as on a $900,000 one.
A solo agent’s constraint is that all three jobs draw on one calendar. NAR’s median member works a 35-hour week; sales agents report 30. At nine sides a year, the file work does not fill it, which is why a coordinator is a hard buy at the median. Somewhere between ten and seventeen sides, on the arithmetic we ran separately, it fills, and the agent who wants to keep growing has to give a job away. Teams are the institutional form of giving jobs away.
Why does the median team close fewer sides per member than a solo agent?
Because the members are not all doing the same job. NAR reports team size and team production but does not publish how many of the four are licensed producers versus support staff, so the eight-a-head figure is a ceiling on producer output, not a measure of it. Consider the two readings.
Dig in — if all four members produce, the median team is less productive per person than the median solo agent. If one of the four is a coordinator or administrator, the three producers average roughly 11 sides each; if two are support seats — a coordinator and an inside sales agent, a common pairing — the two producers average 16. That second reading is what a team is for: two people who find and serve clients at nearly twice the median rate, because two other people are doing the first and third jobs. The eight-a-head number is not a productivity failure. It is what a per-head average looks like when two of the four seats do not sell.
The two populations are different NAR survey cuts — team-based specialists and individual agents — so the comparison is directional; NAR’s release describes teams as “averaging four members” while HousingWire reports the figure as a median, and the production figure is a median either way. It is a rough instrument, and the data cannot tell the two readings apart: four modest producers and two producers plus two support seats give the same eight a head. What the instrument does say is that a team’s per-head production is not the case for teams. The case is what the support seats let the producers do.
Which job should a growing agent give away first?
The third one. Three reasons.
Closing the file is the only job that can be bought by the unit. Prospecting is bought by the hour or the month — an inside sales agent, a marketing VA, a lead source. Serving clients is bought with a split, which is what hiring a buyer’s agent onto the team actually is. The file is bought by the file. We publish $400 per contract-to-close file in most states and $500 in California, with a $500 monthly minimum. A 20-side agent buying the third job by the file spends $8,000 a year and adds no headcount. The same agent hiring a full-time in-house coordinator takes on a seat that costs about $69,100 fully loaded at the Bureau of Labor Statistics’ May 2025 median administrative wage and employer-cost ratio, or roughly $23,000 in wages alone for a half-time hire — and either one is a fixed cost in a variable business.
Second, the file is the job with the least upside for the agent and the most risk for the brokerage. An agent’s hour spent on a disclosure deadline earns nothing a coordinator’s hour would not earn, and a disclosure deadline missed becomes the broker of record’s problem, not only the agent’s, under state broker-responsibility rules such as TREC’s in Texas. That asymmetry is why some brokerages pay for coordination themselves rather than leaving it to the agent or the team.
Third, giving away the third job first is reversible. A per-file arrangement scales to zero in a slow quarter. A hired producer on a split does not, and an inside sales agent on salary does not. Hiring for jobs one and two before job three is handled locks in fixed cost before the variable work has proven it can carry it.
What to watch — the first hire is where team economics are decided. A team that opens with a coordinator has bought back the lead agent’s file hours at a few hundred dollars a side and can decide later whether the freed hours go to more clients or to a second producer. A team that opens with a buyer’s agent has added a person who also does all three jobs, and now has two people doing paperwork. The per-head arithmetic goes the wrong way from the first month.
When does it make sense to build the team instead of buying the file?
When the freed hours have a buyer. At 32 sides, the median team’s production, per-file coordination runs $12,800 a year against $69,100 for a loaded in-house seat, so the file is still cheaper to buy than to staff, and the question is what the lead agent does with the hours. If the answer is “more clients I already cannot get to,” the next hire is a producer and the team is real. If the answer is “I would finally have time to prospect,” the next hire is a lead-generation seat and the coordinator stays outsourced. If there is no answer, there is no team yet — there is an agent with a coordinator, which at 20 sides is exactly the right size.
The brokerages that scale operations without adding administrative headcount follow the same sequence at a larger scale: buy the file work by the file, add fixed seats only when the variable work has proven it will fill them, and keep the licensed people on the licensed work.
The bottom line — the median team closes eight sides a head, a figure consistent with teams that have moved the first and third jobs onto non-producing seats so the producers can do the second at well above the median rate — and equally consistent with four modest producers, which is the version to avoid building. An agent deciding whether to build one should copy the mechanism before copying the org chart: give away the file first, because it is the only job priced by the unit and the only hire that scales back down. Build the team when the freed hours have somewhere to go.
Method note: figures are from NAR’s 2026 Member Profile (2025 business year, released June 25, 2026): 21 percent of members on teams, teams averaging four members, median 32 sides for team-based specialists, median nine sides for individual agents, median 35-hour week for all members and 30 for sales agents. We read the 32-side figure as team-level production, as HousingWire does; the per-head figure divides a median by an average across two survey populations and is directional only. NAR does not publish team composition, so the producer-count readings above are illustrations, not findings. Hours figures are as reported by HousingWire from the survey release. Seat costs use BLS Occupational Outlook Handbook median wage for office and administrative support occupations (May 2025) and BLS Employer Costs for Employee Compensation (March 2026). Empower’s rates are as published and change without notice. Empower Transactions is not affiliated with, certified by, or endorsed by the National Association of REALTORS®, and this article is not legal advice.
Frequently Asked Questions
Why do real estate teams use transaction coordinators?
Because a team’s economic function is to move two of an agent’s three jobs, finding clients and closing files, onto other desks so the producers can serve clients at volume. NAR’s 2026 Member Profile reports a median of 32 sides for team-based specialists, read as team production, with teams averaging four members: about eight sides a head against nine for the median solo agent, a figure consistent with non-producing seats being counted in the average.
What are the three jobs of a real estate agent?
Finding clients (prospecting, marketing, database), serving clients (pricing, showing, negotiating, advising, the licensed work) and closing the file (deadlines, disclosures, signatures, document routing). Every closed side carries all three, and the third is the same size regardless of the price of the home.
Should a growing agent hire a transaction coordinator before building a team?
Usually, yes. Closing the file is the only one of the three jobs bought by the unit rather than by the hour or the split, so it adds no headcount and scales down in a slow quarter. At any per-file rate the market publishes, a 20-side agent buying the file work is spending a fraction of a fixed in-house seat, which runs about $69,100 fully loaded at BLS median administrative wages. Build the team when the freed hours have a buyer.
How many transactions does a real estate team close per member?
Roughly eight, using NAR’s 2026 Member Profile: a median of 32 sides for team-based specialists divided by an average team size of four. NAR does not publish how many team members are licensed producers, so the figure is a ceiling on per-producer output, not a measure of it, and the two figures come from different survey cuts.
When should an agent build a team instead of outsourcing coordination?
When the hours a coordinator frees have a clear use: more clients the agent already cannot reach means the next hire is a producer; a need to prospect means the next hire is a lead-generation seat with coordination still outsourced. If there is no answer, the right structure is an agent with a coordinator, not a team.
For team leads
Price your first support seat.
Tell us your team's side count and who does the paperwork today. We will cost the file work by the file against a fixed seat, so the decision about which job to give away first is made on your numbers rather than a rule of thumb.
