We’ve spent plenty of time trying to figure out what’s fair to pay Redfin agents. As part of that exercise, we analyzed the gross commissions for all Seattle-area (King County, to be precise) agents who closed at least one transaction over the past year (May 12, 2009 to May 11, 2010). The data surprised us, so much so that we thought we’d ask this community if we’re making any obvious mistakes.
We sorted the agents by gross commission, assigning percentiles to each. When we didn’t know the commission on a deal, we assumed it was high: 3% for each side.
Agents at the 50th percentile of pay earned $29,820 in gross commissions. Agents at the 75th percentile earned $75,018. You don’t hit $100K in commissions until the 82nd percentile. Then we graphed the data, showing the gross commissions on the vertical axis, and the percentile of the agent earning those gross commissions on the horizontal axis. The result was a hockey stick:
But then we reasoned that a lot of part-timers are closing one or two deals on the side while working another job; so we excluded all the folks who earned less than $25K in gross commissions. This shifted the graph to the right a bit, but otherwise we still saw a very small number of agents earning a huge proportion of the total commissions in a market:
Then we asked ourselves how much money a good agent — say, someone earning $100,000 in gross commissions — has to shell out in costs each year:
| Type of Expense | Traditional Agent, Annual Costs |
|---|---|
| Brokerage fee | $10,000 |
| Health insurance | $10,500 |
| Marketing | $10,000 |
| Social Security, Medicare Taxes | $6,500 |
| Transportation | $3,600 |
| Cell service | $1,200 |
| Equipment | $1,000 |
| Dues, education | $783 |
| IT | $1,000 |
| Total | $44,583 |
All told, the data left us scratching our heads. In a fairly wealthy market where sales volume has been increasing, a good agent — someone among the top 15% of his peers — is probably netting less than $60,000 per year. Does that sound right?
Greg Swann says:
If you want to double your dismay, break the adjusted gross income number out to an hourly wage. I normally work at least 14 hours a day, seven days a week, 360+ days a year. Call it 5,000 hours a year, which might be low. Eleven bucks an hour?
The countervailing benefits are huge, though: I control my own destiny, I take shit from no one and I’m making decent money right now, when 10% of Phoenix is unemployed.
Where the curve starts to look like a short head, you need to divide GCI by headcount. The people at the top of the food chain work less than the grunts on the ground, but they do it by delegating the low-wage scut work. That’s every serious Realtor’s goal, to get a reliable-enough cash-flow to be able to add — and pay for — cheaper labor.
May 24, 2010 — 9:05 am
Jeff Brown says:
Greg hit the nail on the head.
My model has been a winning one for me since first tried back in 1998. My so-called ‘team’ consists of assistants, both part/full time and in/outa town. Unlike Greg, I never feel compelled, nor do I love doing every-freakin’-thing, as he does. I average roughly 35-50 hours weekly. So far this year, the production has been solid. That’s code for, ‘My year’s expenses are paid for, but haven’t been meeting my Herculean goals just yet’. 🙂
May 24, 2010 — 9:44 am
Dan Scher says:
For Redfin, I think you are missing the point. The Internet capture model, due to all of the leads it provides to agents, plus the discount commission and buyer rebates is supposed to greatly increase the number of transactions each agent can produce. With those two advantages plus the help of contract managers, agents should be able to complete ten times the number of transactions that the average agent accomplishes. If agents completed 50 – 60 deals a year the percentage of the commission they receive would be less important. If you think about it most discount models (soft goods, food, etc.) greatly increased volume has always been one of the most important aspects of success. Also, the agents are mostly concerned about how much money they bring home. I don’t think you will find many agents bring home a six figure income that are complaining about the percentage of the commission they are receiving.
Of course, no company of whom I am aware, has yet accomplished that goal. Foxtons did the volume, proving it could be done, but messed up other places (bad service, over expansion too fast, too much money spent on too much advertisig and mostly {and this is true with all of the companies with a simuliar business model as yours} terrible inadequate agent training and management.
Contact me and we can talk about the solutions.
May 24, 2010 — 9:46 am
Tim and Julie Harris says:
If the public only knew what it actually takes to get a home sold…how much time we have to put into getting it sold…all the hard and soft costs etc…
Gregs demonstration of the hourly rate makes the point perfectly.
Tim
May 24, 2010 — 12:41 pm
Glenn Kelman says:
@Greg: great points. You always help me understand this industry better and I thank you, sir.
@Jeff: good luck meeting those Herculean goals! It’s early yet…
@Dan: we were asking about the industry not ourselves, though I agree completely on agent training and management as being the key to success.
May 24, 2010 — 1:29 pm
stephanie crawford says:
I think you might be over-estimating some costs, and under-estimating others. As a single agent I only pay $196 per month in health insurance. But your brokerage and tax estimates seems low to me. I don’t take my brokerage splits into consideration when looking at my net. I mean – it’s not real money and its not taxed.
May 24, 2010 — 6:31 pm
Jesse D. Moore says:
I think that you’re underestimating most of the costs. Brokerage fees are generally much higher, health care might be lower, marketing is higher, cell is much higher, and dues are higher. In all, I think that based on your model an agent that makes $100,000 might net more like $50,000.
As for your data, I think that you should exclude agents not on income earned, but on production. An agent could have made $25,000 on one transaction. Shouldn’t you instead define what a “productive” agent is and exclude agents below a certain number of transactions a year? I think you would start to see that graph flatten out dramatically, although the 80/20 rule still applies.
Thanks for this post Glen. I’ll refer to it the next time a homeowner asks me why I charge a full commission.
May 24, 2010 — 9:05 pm
Joe Spake says:
It’s the old 80/20 rule, except in real estate it’s more like 85/15. The top 15% producers rake in 85% of the commissions.
I can’t bring myself to calculate my hourly rate, because, being self-employed (independent contractor), I can choose how little or how much to work.
May 25, 2010 — 6:32 am
Sean Purcell says:
Glen,
Interesting article, but you’ve got a statistical problem and Jesse nailed it on the head. Trying to estimate the average income of an agent, your variables should revolve around the agent and not the income. Dropping agents who grossed less than $25,000 didn’t change your results much because it didn’t change any true variables. For example, you might graph income distribution by number of transactions as Jesse suggested (create groups: agents doing less than 10 transactions, agents doing 10-20, agents doing 21-40 and so on). It would be interesting to see those graphs overlayed. Another variable you might control for (although it would take some work) goes to Greg’s point. Graph income distribution by hours worked. Agents working less than 20 hours/week, agents working 20-40, agents working 40-50, agents working more than 50. I don’t know for sure, but I’d bet the results of that overlay would be very interesting.
Other variables might include listing agents vs. buyers’ agents average income distribution or income distribution measured against years in the business, or how about income distribution as a function of certifications! (I bet that’s one graph the NRA and a lot of real estate teachers would rather not see the light of day). You could probably come up with a dozen variables of interest, the key is to control for variables revolving around the agent and not the income.
As you’ve no doubt noticed (and your study would seem to confirm), real estate is a profession strikingly lacking in professionals. Could you imagine the same study for doctors or financial advisors? The 80/20 rule would most likely still apply, but I’ll bet dollars to donuts the base would be at a “profession” level rather than a part-time cashier’s level.
May 25, 2010 — 11:25 am
Thomas Johnson says:
Glenn: Most brokerages cap the company dollar around $20000 for productive agents. So, a 100k GCI producer would keep 80k and cover his costs, marketing, health insurance, taxes, car,telephone,technology, food, housing and Starbucks card with that amount. Above 100k it all goes to the agent, less the incremental taxes. This helps explain the hockey stick as the incremental revenue is ever more profitable to the agent. This is where the money and the leverage for the grunt work comes from. In my understanding of your model, I do not see where there is incentive for a top agent to pour on the GCI(Gross Commission Income) for maximum profit, since happy surveys are where the agents’ rewards lie.
May 25, 2010 — 12:02 pm
Lauren Dugger says:
To get a better understanding of what amount of “pay” each agent should get, the time would be better spent figuring out what percentage you are really going for, or just going with the state/city average. There are agents that have better splits due to the fact they have been with a specific company for far longer then other agents. When doing your graph, this is one of the variables that should be included, because there is a flood of agents just starting out, with a standard split, while the top dollar agents with the better splits are selling homes as well. The “older”(I use this term referring to agents in the game longer) agents will commission more, so your graph, without accounting for this, will be lopsided.
May 25, 2010 — 12:43 pm
Ashlee says:
I think in every market there are going to be those few agents that blow everybody else out of the water.
May 25, 2010 — 6:54 pm
Robert Worthington says:
Interesting graph you have Glenn. Your estimates seems right on in my eyes. Great read. Greg Swann makes a super point as well. I’m with you 100% Greg about 5000 hours a year.
May 26, 2010 — 6:41 pm
Joe Dallorso says:
Like all national real estate blogs you are over looking the fact that real estate is local. The median home price is what effects an agents income the most. If you are in a market with a $400K average sale that’s very different from here in Florida where my median home price is just below $100K. I did 13 sides in 09 but only averaged $125K per side. It’s not about volume or even hard work, it’s all about the average sale price.
My expenses are actually lower than your chart. I’m 100% commission for a $349- a month desk fee. I’m on my wife’s health care. My only marketing is my blog, web site & postage.
May 27, 2010 — 5:41 am