Redfin CEO Glenn Kelman has a reputation for bold statements. He once said “dropping a turd in a punch bowl” on an earnings call and more recently joked, “Plan B is to drink our own urine or our competitors’ blood.”
But behind the quips are often real insights about where the industry may be heading.
In a recent conversation with Andrew Flachner on the Playmakers Podcast by RealScout, Kelman revisited Redfin’s turbulent experiment with W-2 agents, explained why the Rocket acquisition makes sense, and shared his vision of a vertically integrated future where AI, mortgage, and brokerage work together.
Here’s what stood out.
The Promise and Pain of W-2 Agents
Redfin’s W-2 agent model has been one of the company’s most controversial decisions. Unlike traditional brokerages that rely on independent contractors (1099s), Redfin employs its agents directly. The idea was simple: if one company could oversee both the technology and the service, the consumer would get a smoother, more consistent experience.
Kelman still believes that’s true. But the reality of housing cycles has tested the model.
“In 2022, we did I think five layoffs. It was awful, and I just don’t want to go through that again.”
Those layoffs highlighted the downside of tying payroll too closely to the ups and downs of the market. To protect against future swings, Redfin now sends some business to competing agents at firms like RE/MAX and Compass, rather than hiring too aggressively during hot markets.
Even so, Kelman isn’t ready to give up on the employee model.
“Even 15 years in, we considered switching to a 1099 model… but if you’re all employees of one enterprise, well, we’re going to take care of the whole enchilada.”
That “whole enchilada” means more than just selling the home. It’s about capturing revenue across the entire transaction: mortgage, title, and closing.
Which is where Rocket comes in.
Why Rocket Bought Redfin
Rocket’s acquisition of Redfin signaled a new chapter for both companies. For Rocket, the appeal wasn’t just Redfin’s traffic; it was the potential to combine a leading mortgage lender with a tech-driven brokerage and a popular consumer search platform.
Kelman was blunt about what matters most to consumers:
“One insight that I’ve had is that there are people in this world who want to save money on a real estate agent, but not everybody. And I feel differently about mortgage rates. There is no one who prefers a higher mortgage rate to a lower mortgage rate.”
Translation: commission discounts are nice, but they’re not the big driver. Lowering the monthly mortgage payment is. By pairing Redfin’s W-2 agent model with Rocket’s lending machine, the combined company hopes to deliver savings where every consumer feels it.
Why iBuying Didn’t Work
The Rocket deal also underscores why Redfin is moving away from another bold experiment: iBuying. Like Zillow and Opendoor, Redfin once bet that consumers would pay for instant liquidity. But rising rates have made that model a lot harder to sustain.
“Anyone who wants that level of instant liquidity is really going to get a haircut on the amount of money they get for the property. And that’s just a trade-off that makes iBuying much less attractive.”
In other words, when money was cheap, companies could afford to buy homes and hold them until resale. With interest rates closer to 7%, that holding cost is too high.
For Redfin, the better play is shifting focus to mortgages, title, and services that can be delivered at scale without sitting on inventory.
Where AI Fits Into the Picture
The Rocket-Redfin partnership is also about building the next layer of efficiency through AI. Redfin already uses machine learning in its Zestimate rival, the Redfin Estimate, and in its listing recommendations. Behind the scenes, AI also drives lead routing and agent follow-ups.
Kelman described one example where AI tracks consumer behavior: if a client views the same home multiple times in one morning, the system prompts an agent to reach out.
The idea is to help agents focus on moments that matter most, not busywork.
And while some industry voices predict AI could replace agents entirely, Kelman’s not buying it. He sees AI as a way to eliminate “baloney work” so agents can focus on advising clients during high-stakes decisions.
A Rallying Cry on Productivity
Kelman also didn’t shy away from challenging how much the average agent produces.
“Most agents should not just be doing six deals a year. They should be doing 12 or 24 or 36 deals a year, because we all know you can close two or three deals in a month when you have to.”
It’s the kind of remark that makes some agents bristle, but it highlights Redfin’s approach: higher productivity, salaried compensation, and revenue spread across multiple services.
The Bottom Line
Redfin’s early gamble on discount brokerage didn’t reshape the industry the way Kelman once hoped. But with Rocket behind it, the company is betting on a vertically integrated model where W-2 agents, mortgage, and AI all work together.
For agents, the lesson isn’t to copy Redfin’s playbook, but to watch how consumer expectations evolve. If Kelman is right, the real competition won’t be over commissions; it’ll be over who can deliver the most value across the entire transaction.
