Zillow CEO on growth in a housing crisis

Today, I’m talking with Zillow CEO Jeremy Wacksman.

Zillow is one of those apps that really exemplifies what you might call the smartphone era of software: the company built a great mobile app for looking at real estate listings, and it turned into not just entertainment for so many of us, but what has become a vertically integrated platform for buying, selling, and renting real estate.

If you’ve been listening to Decoder lately, you’ve maybe already guessed why I was so eager to talk with Jeremy — a popular mobile app like that is really the front-end to a big series of databases, and the politics of who gets access to databases and how much that access should cost is newly up for grabs if you think the next era of software will be defined by AI. After all, you don’t need an app like Zillow if your AI agent can just go look at the multiple listing service, or MLS, database directly.

Or maybe you do — maybe you need an app like Zillow more than ever, which is of course the case Jeremy made in this conversation. And the politics of Zillow’s database access are as complicated as it gets — the company is involved in big lawsuits over when and how houses get listed in MLS versus Zillow, and realtor groups have an uneasy relationship with the platform as a whole. Jeremy’s argument is that the future of Zillow looks a lot like an end-to-end business platform for real estate agents, and we spent a lot of time talking about whether a business as local and as relationship-driven as real estate can benefit from platform-level scale in the way he’s proposing.

Yeah, there’s a lot of Decoder in this one — it’s going to surprise you, I think.

Okay: Zillow CEO Jeremy Wacksman. Here we go.

This interview has been lightly edited for length and clarity.

Jeremy Wacksman, you’re the CEO of the Zillow Group. Welcome to Decoder.

Thanks for having me. Happy to be here.

I’m excited to be in person. I love doing these episodes in person. Thank you so much for coming to our office here in New York.

We’re here for the week. Zillow’s 20th anniversary was yesterday; we’re celebrating it today. So it worked out really well to be in person. I’m excited to do it.

It’s always a little looser when we’re in person, so I’m excited to do it. I’m excited for this episode. Zillow is one of those tech companies, there are platform dynamics, there’s a lot to talk about. My favorite Decoder episodes are the ones where you point out that a database is very complicated and full of political machinations and big economic regulatory disputes. And there is no more complicated database than the databases of houses that are for sale in America.

I think that’s well said.

And so I’m dying to just dive into that. I think most people understand Zillow is the app where you judge other people’s taste in interior design. Especially in 2026 America, it feels like its primary purpose.

Entertainment on Zillow has been a growing phenomenon since I’ve been there. Absolutely.

You’ve been there since 2009, I want to say?

16 years. And you only just became the CEO, I think last year or year before?

Yeah, about a year and a half ago. August 2024.

That shift from, “Hey, we can put all the houses for sale in your area on your phone and you can just see what’s happening,” to, “Oh, we’re going to transact a lot.” At one point, I think Zillow was buying lots of houses. So now you’re the CEO and the whole market has changed. Describe the arc of Zillow briefly for people, what the app was, what the company is now, what you wanted to be as CEO.

Zillow started 20 years ago today as we’re recording this. And it was founded on the principle that real estate information wasn’t available to the average buyer and seller. So our co-founders, Rich Barton and Lloyd Frank had come off of Expedia. They had built a bunch of marketplaces. They looked at real estate. They were shopping themselves and they said, “Wait, I can’t find out what homes are for sale or what houses are worth. The data is locked in a professional’s database and I have to go pay for access.” And so that was the founding inspiration: “Let’s turn on the lights. Let’s give people access to information.” The other categories on the internet by then had gone that way. You could start to get access to travel and all the other verticals, but real estate really hadn’t come to the internet circa the mid 2000s.

We launched in February 2006 with the Zestimate. That was our launching product. And that was just trying to answer one question: just what are homes worth? Not a perfect answer, but an estimate of, okay, take some comps and build a machine learning model to figure out what these houses might be worth to help you be smarter when you start to shop. And Zillow grew out of that into an advertising business. The first decade plus, it was just a classic ads marketplace. It was, okay, let’s offer great products to buyers and sellers. They come and use it. They dream, but they also shop and they buy. And advertisers, professionals are going to want access to that audience. So we’ll just sell access to the audience and sell leads and sell advertising. 10, 11 years in, we had built this great brand.

We were very proud that Zillow was searched more than “real estate” on Google. We have become a category term. People love the brand. The cocktail party story was always, “Wow, can you change my Zestimate?” Or, “How do you do that?” But we’re customer obsessed and we would ask customers how it was going and they’re like, “I love using Zillow.” Then what? “I cried when I bought my house.” And the data point I always remember was we surveyed the average home buyer. We do this every year now as a housing trends report and percentages and demographics and trends. And one of the things that stuck out back then was that more than half of buyers cried during the process. We’re like, “That’s not great. We’re the preeminent brand in the category. We’re helping people with the start, but the rest of it’s still broken.” So we shifted and pivoted the business to be more transaction focused.

And we obviously still have ads as a part of the business model, but what we’ve been focused on for a while now is actually building the software to help the buyer, the renter, and the seller transact. Not just find the agent, but coordinate and collaborate with the agent, the software for the agent to help you buy and sell the house, the software for the agent to help you list the house, the services for you to get the loan from us, the software for you to coordinate all that together. And we measure our growth and our success now as the amount of transactions that we actually participate in with the buyer and the seller.

That’s been the journey we’ve been on. We’re not done. Like you said, there’s no more complex database than the real estate one. That’s what makes building this integrated experience so hard: you have to coordinate so many things, which we can talk about. But that’s really the history of Zillow. And it’s fascinating because we’re celebrating a 20-year anniversary and the brand has come so far, but the goal of building a one-stop shop for real estate, we’re really just getting started.

I want to come to measuring transactions, especially in our current real estate market where it seems like transactions are just going to stay low for a while. That initial piece of Zillow, where it’s going to just show you what houses are available, how much they transacted for. That is just a public database. This is just a thing that happens in your area and you can have access to that database. And I would say, broadly, during the early internet and particularly early mobile, the innovation was often just exposing the database to people, packaging it up, making it pretty, saying, “Here’s some information that’s available to you that certainly now on your phone.” I’m at a restaurant, in some town I’m visiting, I can just see what the houses are worth.

Or you’re walking the beach and you can say, “Wow, I want to move here! What’s that house worth?”

That’s a big public database that created a lot of value for Zillow, created a lot of value for Zillow’s users. Over time, it seems like Zillow wants to have its own database. And the one that’s right next to it, that’s important. Over time, as you want to take control of more of that transaction journey, you need control over many, many more amounts of data through the process. Right next to that is the real estate industry in the United States, which has the National Association of Realtors. They’ve been in some regulatory issues that I want to talk to you about, but they maintain the multiple listing service, where it’s tempting to think of that as one database, but it’s actually like hundreds of databases.

Maintained by little groups of regional realtors all doing their thing.

That’s right. That’s right.

What’s your relationship with that? Because that feels like the most contentious back and forth of this all.

It’s a great question. From the outside, you would assume it’s contentious because we’re the innovator, we came in, we’re the technologists, but the reality is we love the MLSs. They are the marketplace. We have this real estate database, that is the MLSs and it’s available to everyone. Whenever I talk to people about our industry and they don’t work in our industry, what they are surprised to learn is that the listings that are for sale are broadly shared by every person in the real estate market to everyone else. So the listings that are on Zillow, the listings that are on realtor.com, that are on our competitor websites, we all have the same listings. There’s no differentiation of what’s for sale. And what that means is it’s an incredible public good. Buyers and sellers, big brokerages, startups, and innovators can all get access to the same information because the real estate industry has agreed to cooperate and share it.

We’re the only country in the world that works this way. Every other country, they are private databases. They are somewhat controlled. That creates less competition, that creates less transparency, and it actually creates a worse consumer experience. So we have grown up and the internet coming to real estate actually allowed the MLSs to build this so that it’s commoditized. Zillow then has to differentiate on top of that. We have to fiercely compete for an audience and win that audience with better products on top of commodity data. Many marketplaces, you build content other people don’t have and you have this lock-in benefit. We don’t have that. We have to build better touring experiences. We have to build better AI-powered virtual touring. We have to build better financing. We have to attract buyers and sellers to want to use our services because one click away is the same listings on another website.

Do you think if I went and asked the National Association of Realtors or any of these small realtor groups that maintain these databases, if they think their product is a public good, that they would think it’s a public good?

I think many MLSs would say that.

The transaction data is definitely a public good. That’s literally public information.

Well, the transaction data is actual county public records that you get.

The MLS is a bunch of realtors who’ve all decided to share their listings together in very specific ways. They have their own agreements. There’s stuff in those databases that I can’t see as a member of the public.

When you described it as a public good—I’m pushing on this for a reason—they maintain control of it and there is friction in that relationship.

How do you think they see their relationship to you?

I think many see it as complementary. I think the challenge is there’s 500 plus groups. So we can’t have one conversation. There are many MLSs that are very consumer focused and most of their rules, as you put it, are consumer friendly because they want to support broad membership. Their interests are, “How can we have more members? How can we have more folks come and use our service?” So there’s some that are more industry focused and we would actually take issue with some of their rules, and we do. We’re allowed when we think they have anti-consumer rules.

But I guess to answer your question, on balance, if we did not have a shared cooperative database and we went and signed up and got all the listings—which, by the way, we did before all this. We did not get our listings from the MLS until six or seven years ago. We still had most of all the listings on Zillow because people want to put their listings on Zillow. It was just way less efficient to get it that way. You’re talking about tens and thousands of feeds and relationships and changes and all the things. So on balance, the idea that the database is coordinated and broadly available is 80%-90% good, and there’s 10%-20% friction and rules that we don’t love that we work hard to lobby and advocate and change for.

One of the themes on Decoder that comes up over and over again is obviously aggregation theory, where on the internet, particularly people that have the consumer demand get to set a lot of the terms. Google is maybe the ultimate aggregator. There’s just incredible demand for whatever’s on Google and then Google gets to set a lot of terms for suppliers. There are many, many other examples of this. Credits to Ben Thompson who came up with aggregation theory. Zillow is one of those aggregators. And so even in the case where there was another database of information, you’re saying the real estate agents knew that the audience was on Zillow, so they should put their listings on Zillow as well. At what point did you realize that you had the ability to go say to the MLS providers, “Hey, this is way more efficient.” Because at some point they didn’t think that was efficient. When did you realize it would be more efficient or when did they realize it’d be more efficient to connect directly?

I don’t have the history perfectly, but the internet broadly helped them evolve that way. There’s a standard that they created that I’m sure if we go back, back to when Zillow was founded, it wasn’t available where brokerages, the real estate industry, the MLSs could make available a feed for any internet site or any internet brokerage based site to display. That came about as the internet was coming, as sites like Zillow were coming. They reacted to that. In a pro-competitive and pro-consumer way, it probably happened over the last 10, 15 years. The other part of Zillow’s history is we didn’t start that way. Because back then, there wasn’t this internet-powered set of MLS rules. They were before the internet. They hadn’t thought about this. So we had to go sign up and get all these things on our own and we did that and we built these great brands and consumer experiences and then I think the industry caught up.

The reason I’m asking all these questions is, one, I think the politics of databases are fascinating and in particular, who gets to connect what, who creates the data and populates the database and who gets to display the data in the database. In the age of AI it is actually more up for grabs maybe than ever before. The idea that you have an interface as opposed to a chatbot that’s directly reading your database—we’re going to come to that, but I see all these politics as upper grabs. The other reason is I wanted to set up the Decoder questions here. You had a business that was, “Hey, there are a bunch of public or semi-public databases that we can access to.” And your job, as you’ve described it, is to build the best product on top of those databases and to say, “Here’s the listings, here’s the best tour, here’s the best photo viewer.”

And I don’t mean this to sound reductive, but that is like, “We’re going to make a great mobile app. We’re going to have the fastest photo viewer and worry about caching images, where they are and all that stuff you’re going to do. We’re going to have the best 3D tours.” The things you’re describing now are real estate transactions. It’s not just pure software design on top of databases. It’s going to actually go and address the business logic that someone is having when they’re looking at a listing. That seems like you would have to change the company pretty dramatically. So tell me how the company was structured before you became CEO, and then this is the Decoder question: How is Zillow structured now?

I’ll answer both those questions. The company structure pre- and post- my taking over CEO didn’t change too dramatically, mostly because the CEO transition was pretty planned. I was COO before and we really planned for succession. We started organizing the company back then. The question probably is better [framed] as, “How did we set up the company at that stage?” And that’s largely how it’s structured now. We are—I’m sure you get this a lot—we’re a classic matrix. We have functional teams and functional expertise that then sits alongside business groups for our most mature businesses. Great product marketing, engineering, and design teams that are fully integrated with dedicated leaders that matrix into the business team and most of our mature businesses are run in one group. That’s most of Zillow.

What are your mature businesses?

Most of our for-sale business is how we think about it. We think about for-sales and rentals. Most of the buyers and sellers that are using the Zillow brand that you see, and most of the agents that are talking to them and most of the loans and home loans that you’re getting in our residential real estate, is the most mature part. And I don’t say “mature” in terms of market share and growth, but in terms of how long we’ve been at this. We also have startups, acquired businesses, smaller businesses that we intentionally run more independently for some period of time until we need to think about how to better integrate them. Rentals are a great example. We have a rentals marketplace that’s now north of 20% of our company, but that was largely a startup inside of Zillow for a long time and it operated and still operates relatively independently with a GM of rentals and had a lot of its functional teams fully owned.

As it’s matured, now it’s going to be our next billion-dollar business. That’s one of our goals, is we are starting to intentionally think about, okay, well, how do we fold some of that back into the functional expertise capabilities? StreetEasy here in New York, which I know you all know well, that was something we acquired and maintained its independence and have continued to maintain its independence because there’s not a lot of value and there’s actually a lot of more value in keeping it independent and thinking about New York because it’s such a unique market. We don’t have a one-size-fits-all, but broadly speaking, we have a functional business matrix that we look at. Does it make sense for the next thing we’re doing?

As you’ve evolved, there have been periods of growth in Zillow, there have been periods of layoffs. Again, one of the boldest attempts that will make home buying easier was Zillow would just own all the homes. That did not go well. How big is Zillow now?

We are approximately—don’t get my account exactly right—approximately 7,000 people.

And how’s that organized? Or how’s that distributed?

We think about it as [both] fixed and variable headcount. “Fixed” is largely product engineering, marketing, design, and shared services, and then “variable” is sales, operations, support, and things that grow with revenue. So loan officers, rental sales, account managers in our agent business, those kinds of things. And it’s, loosely speaking, about half-and-half.

And are you mostly invested in product stuff or are you all account executives? How does that work?

Product and engineering is probably the largest group. We are first and foremost technology innovators, and that’s where we got our start. As we talked about earlier, yes, the business model’s evolved, but building the best products and services—even if that’s becoming software for the real estate agent, or if that’s becoming software for loan officers—is still how we get to our end goal of more transactions. That’s still the largest investment we have. But as the businesses are growing, staffing more loan officers, staffing more account managers to work with agents, that is variable growth that comes as we see more share come [in]. So those are growing as well. And those are probably on a percentage basis growing more, even if on a dollar basis, that’s still a smaller part of the company.

The other question I ask everybody who comes on Decoder is about decisions. How do you make decisions as CEO? What’s your framework?

I’m a big believer in “Find the truth, don’t find the right answer.” There was a famous quote that I’ll get wrong that was like, humans are naturally biased to want to be right, and sometimes they want to make the argument versus actually seeking the truth in the data. So our version of customer-obsessed is we’re looking for the answer from the buyer or the seller or from the agent using the software. And ideally the data helps us make sure we don’t have biases or we rechallenge assumptions. We push for data where we can get it and data to make decisions. That doesn’t mean we wait for data. Sometimes you have to make a gut decision, but often we want to be as data informed as we can.

Let me ask you about a big recent decision that’s somewhat controversial. We’re talking about databases. And one of the rules that Zillow has is if a property goes for sale, it has to be on Zillow within 24 hours or it’s not going to be on Zillow at all. Explain that decision. There’s a lot of controversy. There’s a lawsuit next to this decision. Explain that decision. You’re like, “This is the policy and it’s going to lead to some controversy.” How did you decide, “Okay, we’re going to hold to it?”

That policy is really about protecting transparency. Back to your earlier question, we all benefit—Zillow and our competitors—from these MLSs having most all available listings for free. For free for us, but [also] free for the agents who work and free for the buyers and sellers to use any site. There are companies in the space that want to take those listings off the internet and hide them for themselves and not put them as part of the database. We don’t think that’s a good thing. If you look at the other real estate markets in the world, that’s how they work. Private databases where you have to pay for access leads to consumers having to pay more. We think that’s a bad idea.

What we can do as an audience in the industry is say, “We don’t think you should be able to free ride and only give the internet the listings that you want and still get access to everyone else’s listing. So we won’t let you do that on Zillow. If you want to do that, then that’s fine. You can operate that way, but we don’t want to have that listing.” It was more of an educational tool so that people understood there was a trade-off if they did that. And ideally in a sense, what we already see, is that the majority of sellers and sellers agents, if given a choice, want their listing on the internet and want their listing on as many internet sites as possible, and that’s what they’d choose.

The company you’re talking about is Compass, which is a reasonably high-end real estate firm. They have a bunch of software of their own. They will happily tell you that they’re a tech company. You’re obviously in a lawsuit with them. In every other business I can think of, there is some windowing. Netflix and TV is a great example of this. The movie comes out in theaters, Ted Sarandos has to promise Congress he’s going to maintain a 45-day theatrical window. He’s going to buy Warner Brothers. Then it goes to, I don’t know, it goes to HBO and then to Netflix and then to the Delta seat back, and every one of those is a different payment. And you can see windowing happens across different businesses. Why do you think windowing is inappropriate in the real estate market?

I think you should think about real estate more as e-commerce than as premium content. When you’re selling a good, you want to advertise to as much demand as possible, as quickly as possible. The number one thing, the reason people cry is because it’s so hard to buy and sell and it takes weeks and sometimes months to do it and you have people traipsing through your house and you basically have to hire someone as a part-time job on your behalf to help coordinate all this stuff. So the idea that you wouldn’t try and maximize price when we ask sellers, most people say, “No, no, no. I want to sell my house for the most that I can in as short a time as possible.” So that doesn’t really fit with trying to window or hold things out or take longer.

For us, the goal would be to help sellers do that, to help sellers understand that if you want to maximize price, the way to do that is to broadly market your home as quickly as possible. If you don’t, and there are rare cases where people don’t, you can do that. You can sell your home privately, you can opt your home out of the internet because you have privacy concerns. All those mechanisms exist today, and we support all those things. We’re not saying you can’t do that. You can window. You just can’t window by only showing your home to some people and then forcing those people to go work with certain companies over other companies because the toll on all those customers is going to go up.

Wait, how can you window if you can’t ever get on Zillow unless you list within 24 hours? I think this is a thing that I don’t quite understand.

Got it. The only thing we take issue with is selective marketing. There are plenty of ways for you to be private or sell your home in a private fashion today. You can never put your home in the MLS and you can sell it off market. And that happens in privacy situations or luxury situations. You can put your home in the MLS and opt out of the internet because you actually don’t want people traipsing around, driving around your street. You can hide the address if you have privacy concerns. You can take all these steps today. And then of course you can put your home on the MLS and it comes to all the sites. You can also list your home as coming soon in the MLS and say, “I’m actually not ready to sell. I have to do a whole bunch of work, but I’m going to be selling my home when the spring selling season comes, so I should start to tell people about it.”

That’s a version of windowing. The common theme across all those is that a company is not choosing which customers can get access to this stuff and making it broadly available for some, but not for others, which can lead to all kinds of bad behavior. And you’re not telling the seller that by doing so, you’re going to get even more money. If you’re going to forgo some demand, that you’ll somehow increase the price. That’s the only thing we say, “If you want to do that, we don’t want those listings.”

The reason I’m pushing here is—I think I understand a broad dynamic in this industry, which was that there was a pretty big settlement in a lawsuit about realtor fees and broker fees, and now those fees are up for negotiation. And so you can see the big agencies are actually trying to close their little markets. They have a lot of listings and if they can just sell in between themselves, they can pre-negotiate those fees. You can disagree with that reaction, but it’s a rational economic reaction to go, “Okay, we’re not going to negotiate broker fees the way we were. We’re just going to create a little closed ecosystem.”

That’s right up against you saying, “Actually, the ecosystem has to be maximally transparent,” because you’re not actually doing the listings. You’re not the seller’s agent in many cases, you’re not the buyer’s agent in many cases. I know that there’s some amount of referral there. So they’re trying to protect their cost of actually managing the inventory. How do you reconcile that? If that industry was destabilized by a legal settlement and they’re trying to find a new equilibrium and you’re saying, “Actually, we’re not going to let you make changes,” how do you reconcile that?

We reconcile it by what’s best for the seller. If we asked 90% of sellers, “Do you want to do something different?” And they said [yes], we’d go work on that. But you ask sellers, “Do you want to broadly market your home? Do you want to sell faster and for more money by public marketing?” They say yes. By the way, agents do too. That’s part of the other nuance of this industry that folks don’t quite get: a lot of these industry imaginations are between the brokers and the MLSs and the associations and NAR and all the legal work.

The industry is actually run by over a million independently licensed real estate professionals who are independent contractors. They inherit a lot of benefits from the companies they affiliate with, but they make their own decisions. They are fiduciaries in many cases, and they are supposed to disclose the benefits and risks of how you sell your goods to your buyer and seller. So that’s what we’re trying to help support. But you’re right. It is not surprising that companies want to find a way to find more margin, maybe at the expense of their end buyer and seller if they’re finding their margin challenged.

I’m reminded of a quote we get often from the various Amazon-adjacent executives who come on the show, that other people’s margin is their opportunity. Is that how you see it as you think about expanding the business?

No, only in as much as the agent’s margin is not really high, the end agent environment. We think about it like this: there are a lot of people in the middle of this real estate, and when you go back to the real estate database, there are a lot of folks in organized real estate. Zillow is very, very focused on the buyer and seller, but we’re also very, very focused on the individual agent at the end of the chain because those are the folks actually producing. And that is what’s led us into building a bunch of software for them and building a bunch of data and workflows for them.

Even though they work at tens of thousands of these brokerages and across tens of thousands of these markets, we don’t really think about the companies they affiliate with as much as we think about their end agents. And those folks are all basically small business entrepreneurs. They are running their own business. They aren’t getting a lot of business support from the inherited entities. We’re trying to help them run their business better. That’s how we help them help buyers and sellers sell more homes, and then more transactions are better for us.

One thing that’s interesting to me, and I’m sure you’ve thought about this, you’ve mentioned the seller a lot, like maximizing the value to the seller, which I think very, again, reductively means the prices are high, but the prices in the home market are staggeringly high right now, just across the board.

And I’m guessing that a bunch of people listening to this episode are going to say, “Well, I’m a buyer. I would love to buy a home. The prices are high. Who’s pushing the prices down? Who’s maximizing my value as a buyer?” Do you think about that side of the equation?

Absolutely. Yeah. We got into Zillow because we talked about listings and databases, but most all marketplaces start from demand and Zillow is mostly focused on the demand side and helping buyers. We are absolutely in an affordability crisis in this real estate market. That’s why, as you mentioned earlier, we are seeing such a depressed volume of transactions, because it is so hard to buy. That’s a combination of, yes, mortgage rates. Everyone loves to talk about mortgage rates; pre-pandemic, they were really low, and now they’ve run up. Although ironically, they’re still not nearly as high as they have been in historical times. The challenge though is home prices.

As you said, home prices are up nearly 100% in many markets from pre-pandemic levels, and that’s not sustainable and incomes aren’t rising that fast. So trying to ease that supply-demand and balance in the real estate market is really the only way to get back to a more normalized housing market. The way I always like to describe it is that in any given year over the last three, four decades, about six million homes would trade on average, five and a half to six million homes. So that’s 10, 12 million people are going through this part-time job process of moving. The last couple of years, it’s been 4 million homes. So we’re 2 million off a normalized market and it’s not really budging and affordability is the reason.

When you say affordability, I know you started by saying mortgage rates, everyone wants to talk about them. I have any number of friends who have pandemic-level 2% mortgages. They have no incentive to sell their house. They’re just not going to do it until… Maybe never. I don’t think they’re ever going to get another 2% mortgage.

No, but they may get more comfortable with a 5% mortgage. If you asked them three years ago, “What rate would you have to see?” They would say, “Three.” Now, three or four years go by and there are reasons people want to move that they’ve been putting off and now they might say, “Four,” or at some point they’ll say “Five.” So the pent-up demand and rates easing will eventually converge and you’ll start to see folks get unstuck from their mortgage more often.

Do you see the amount of building that you would need to see to address this directly from the supply side?

No. I’m glad you asked that question because that is… We get asked about rates so often on the demand side, but the affordability crisis is an availability problem and that’s the problem. The US, Zillow estimates. is nearly 5 million homes underbuilt. We’re just coming out of the global financial crisis back in 2008, 2009, which many folks listening would remember. We’d stopped building enough houses. It’s to keep up with new household formation and we definitely stopped building affordable homes. And so we have this accumulated deficit of supply.

That was a problem. That has been a problem that is now exacerbated by the mortgage rate lock you talked about. You need more new construction to help balance the market, but you also need existing home sales to come on and folks to say, “Well, I will trade out of my 2%-3% mortgage because I need to or because rates have come down or because enough time has gone by.” So you need both those things, but it is a supply problem. There are places in the country that are starting to make progress, but it’s very marginal progress and we have a long way to go.

You were mentioning that the key metric for the company is transactions, the number of transactions that you facilitated. How do you measure that accurately or well in a market that’s as frozen as the one you’re describing? Because it seems like it’s just a percentage of a number that is flat.

Having a small percentage of a number that is flat allows Zillow to grow and gain share even when the market is depressed. That’s a short version of how we think about it internally. Despite being a brand that is so well known and 60%-70% of all home buyers are on Zillow in any given month or quarter, our transaction share—as best we can measure it, and there are ways we have more so than others—is in single digits. The number of times a buyer and seller moves from using us as a shopping and dreaming site to hiring one of our agents, using one of our loan officers, using the software to help sell and list their home, and then we participate with the agent, that’s single digit. You’re right, while the housing market has contracted from the pandemic, Zillow’s share has grown. And that’s why our revenue and our business is able to grow in a flat housing market.

So you’re just taking more dollars out of a fixed pie?

And more transaction volume. We’re taking more share of transactions because we’re able to grow. You grow from five, six, seven on a percentage. 7% of four is still going to be a lot more than 3% of six.

Yes, that’s right. Okay. Well, you mentioned all the other software building. The software for the agents, the software to get the loans. Those are places where there are other companies with existing big margins, particularly loans. I always think about how I can tell when there’s excess margin in a market and it’s when I experience Glengarry Glen Ross on the other side of the phone. I whisper to the internet that I might want a mortgage and 500 companies show up at my doorstep. That means there’s enough money to do all that marketing to support all that lead generation, lead conversion. That does seem like an opportunity for you. You have the interface, you have the demand. “We’re just going to show you the rates, you’re going to pick a mortgage.”

Have all of those companies been willing to participate in that system? Because that would reduce their margin.

That’s a place where we are competing more directly because now we actually originate mortgages. That’s also part of that transaction journey we talked about. In the first chapter of Zillow, we were just an ads marketplace for mortgages too. And a lot of people came to Zillow and said, “I need financing.” And we said, “Great. There are a bunch of mortgage providers that will advertise to you and go work with one of them.” Then all the advertising inefficiencies that you talked about happen. We now are actually in the loan origination business and we more and more are originating mortgages ourselves. And you are right, we have a customer acquisition cost advantage there because we already have the shoppers on Zillow. There’s also a customer benefit there. Again, we’re back to asking the buyers what they want. What they don’t want is to be aggressively marketed to. They also don’t want to have to hire a bunch of different companies to do pieces of the transaction and have no idea where they all are and have 37 Gmail threads and text messages about all the pieces.

They want to be able to open an app and go, “Okay, here are my conversations with my agent, here’s my offer, here’s my financing, here’s my closing. I just want to see it all there.” That’s what we’re trying to build. You have to originate the mortgage yourself, really, to do that. In that case, we are competing more directly. We are an incredibly small mortgage originator. Mortgage is a massive category. We are basis points of share, but the growth for Zillow is going to come from offering that more often because when a buyer comes to Zillow, they’re asking two questions. They’re saying, “Can I afford it and is there anything I want to buy?” And then in asking those two questions, they realize, “Oh, I need a professional.” One of those two things leads them to hire an agent and we’re trying to offer answers to those questions all in one place. That gives us this tremendous customer acquisition cost advantage to offer mortgages as one of the things they could choose.

I’m just thinking about when we bought our house. Our agent, our buyer’s agent—this is before the settlement. So the seller was going to pay my buyer’s agent’s fees. This very nice woman showed up and she just helped us buy a house and I paid her no money because the seller paid the fee for her to do everything. And she was very local to where we live and she knew all the local mortgage brokers, she knew all the local attorneys. And I was like, “This is great.” I’m just going to call Pam. “She’s going to do it.” And she just did it. Can Zillow offer that kind of experience when there’s no guarantee that it will get a fee the way that the buyer’s agents use to get a fee, or are you just going to collect fees at every step?

There are two questions there. Yes. I think part of that is, can we offer a mortgage service that your buyer’s agent would want? We think the answer to that question’s yes, because one, it’ll be more economical. It’ll be integrated into the software. If you hired Pam through Zillow, ideally she’s using a follow-up boss. That’s our customer relationship management software that many of the top agents use. So if you were to message her inside of Zillow, she’s messaging back from our software, your loan officer shows up in there, your pre-approval letter shows up in there. We win her business because we are efficient, we are priced competitively, but the most important thing, the reason that she recommended that person, if she’s doing her job right, is she just knew that the lender was going to close.

Buyer’s agents close a handful of transactions a year. If you ask them what they want, yes, they want to help their customers save money for sure, but they also want to make sure the deal doesn’t get wrecked. What we are doing is building a capability set that we’re taking to more and more agents to show them, yes, Zillow, even though we are a national brand, we have local expertise and we have loan officers everywhere that can help. And that’s the number one question we have to solve, to get the deal done for them even when a hiccup happens at the 11th hour or even when someone calls on Saturday night. That’s what their value prop test is. And then of course the buyer’s value prop test is trust the agent, trust the loan officer, but also they want to be efficient.

One thing I’m really curious about in this dynamic is the value of data. And I promise, we’ll get to AI in a roundabout way. I’m previewing that we’re going to arrive there.

We’ve gone a long time before we’ve gotten AI.

I do my best to bring up AI late, because people’s actual businesses are interesting to me and the idea that we’re all just going to yell about AI in some mythical ways, it comes up. I promise it comes up, but I like to understand the actual business first. And the reason I’m saying I’m going to get to AI with this next question is because real estate is such a local data-poor business. You’re just in whatever market you’re in, and those are the houses and every transaction is different in a way that maybe is not even remotely comparable to the next door transaction that’s going to take place, or maybe it is, but that’s why everybody’s mad about this estimate.

There’s just something about this market, this business where the presence of additional data might not actually be helpful, especially if you have to collect that data on the national scale and then tell some local broker or some local agent or some local buyer, some local seller, “Hey, we know everything. We can definitely close because the data says this national broker that we’ve partnered with, this national mortgage company we’ve partnered with, is going to close.” How do you reconcile that? Because this feels like in order to make the AI tools work, you definitely need to feed them lots of data, but the presence of the data might not actually matter at the local level.

It depends on the type of data. Let’s take two examples you brought up there. In the mortgage process, the mortgage is, for lack of a better term, a government-mandated commodity. At the end of the day, a conforming loan is eventually sold to a bunch of Government-Sponsored Enterprises (GSEs) and so it has to fit a bunch of standards. It has to be underwritten exactly the same way. All the data collection for a mortgage is every company’s collecting the same stuff. And so in that case, maybe AI or whatever for more data collection is less interesting on a national or a local scale.

But I think your question was also more about, “I’m buying a house in Seattle and there are only so many houses in my price range in the neighborhood I want to look at in Seattle. And so what’s the value of going after more data?” In that case, it’s actually more data to help the buyer get a better sense because what’s unique about real estate is you are making a set of trade offs. It’s not like you go to Amazon or you’re even shopping for a hotel and you actually have to pick from more supply than you could. You have to make a set of trade-offs when you’re buying to say, “The perfect house is probably never going to come up.” It’s like, “What’s available in the timeframe I have and the price rate I have then?” And if you did it 10 times in a row and in 10 alternate universes, it would play out differently.

When you’re making a set of trade-offs, more data about how to make that trade off is immensely valuable. Take a product that Zillow offers called Zillow Showcase, which allows AI tools to generate more of a virtual tour and generate more data to learn about whether that house is good for you before you’d have to go get in a car and waste your Saturday going to the wrong houses. That kind of data is immensely valuable because people are in search of, basically, regret minimization when they’re buying in the category. It depends on the type of data, maybe in the transaction, whether it’s worth going after and investing in incremental data.

I’m just thinking about when you say, “We’re going to build a bunch of software for the agents,” or, “We’re going to originate a bunch of mortgages.” The value of the agent is perfect knowledge of their locality in many ways or like being able to stage your house. It’s in the atoms, it’s not in the bits.

It is in the atoms, agreed.

And those atoms are pretty local. And I’m just wondering as you say, “We can definitely say the data shows we’ll make you more efficient on X dimension,” if that’s actually convincing.

For the agent—[using] the example you just gave—letting them focus on the atoms and not the bits is incredibly valuable. That’s the other thing. If you talk to Pam, she probably wanted to spend time trying to figure out how she should buy, trying to figure out how to help you. The best agents are, yes, they’re staging the home, but more often they’re negotiators, they’re consultants, sometimes they’re therapists. They’re doing the professional service, client service work, but they’re stuck doing all the busywork. The customer relationship management software we offer, you go watch and agents are spending so much of their time on follow-ups, responding to the text that came in, checking the item off the to-do list. This is stuff that software and workflow can and should do for them. If we can generate more data to help take away the busywork, the back office work, so they can do the front office client work, that’s intensely valuable to them as small business owners.

One of the interesting dynamics, again, of Zillow, it’s been around for 20 years. There have been a lot of acquisitions, there has been a lot of growth. You’re talking now about pretty much vertical integration in the stack. We went from “We have a demand and we can showcase this public database,” to now, “We work more directly with the hundreds of databases that have whatever dynamics they have.” Okay, now it’s the banks. We’re all the way up to, “We’re going to make this transaction and that’s our key metric.” As you integrate vertically, are you running into regulatory concerns? And I know the Biden FTC would have a lot of concerns about a lot of things. The Trump FTC, the Trump DOJ, they have different concerns and they express them differently and very differently in some cases, but with just as much emphasis. This FTC and this DOJ, they’re not shy about their opinions. How has that changed and are you running into those concerns?

We have a benefit in that we’re so small in that market. The market which is—well, one, we would argue a local share of transactions, but even if you want to think about a national market, a national share of transactions, we are so small. Then the components of the transactions, mortgage is one of the most regulated markets out there, and we are basis points of share relative to the large mortgage providers. We can grow nicely and still be very, very small. We need to be aware, but we try to build all of our products backwards from the customer outcome and from consumer choice. It’s in our DNA and it’s in our business. Again, as we talked about earlier, more volume is better for Zillow, therefore that’s why we focus on transparency. That tends to align us well with avoiding any regulatory scrutiny.

I’m sure there will be a point, and we definitely have various individual legal challenges we can talk about, but from how we think about regulations as we build, we think, “Let’s build something that’s pro-transparent, pro-consumer,” and there are going to be places where we want to innovate. Innovation is going to be stuff that hasn’t been thought about from regulatory frameworks, but if we can show our work and explain why we want to do what we want to do, we think that’ll end up being a good conversation regardless of administration.

Have you guys been having those conversations with just administration?

We have conversations with both types of administrations over the 20 years we’ve been here because housing, as you said, is such a fascinating database. The other challenge is, you asked about DOJ, FTC, it is more of a local issue. A lot of times these things play out at the state regulator level or state associations and even local markets. And when we get back to macro lobbying for supply, that’s a county-level challenge.

It’s funny, I’m talking about this and I’m ping-ponging between the experience of using Zillow and then asking if you’ve talked to Donald Trump about mortgage rates. He has very strong opinions about mortgage rates. I don’t know if you’ve heard he would like them to be much lower and that’s his solution. There’s building supply. All of your ability to do all these things comes from being that aggregator. Having that demand. And that is under threat in different ways. As I look at the different aggregators, we’ve done a lot of episodes about companies like DoorDash. We literally call it the DoorDash Problem here on the show, that the next great interface might just be a chatbot and the mobile revolution will be where we create new interfaces for databases on our phone and build a new economy.

Jony Ive is doing something at OpenAI. It is unclear what, but it is almost certainly going to be voice-activated in some way. You can see where all of this is headed. That disaggregates you. Quite literally, it takes your interface out of the mix and maybe presents it in someone else’s app or doesn’t present it at all in some other way. Have you thought about this problem? It seems like the most unlikely one to be disaggregated first. I think food delivery is much closer. Getting Ubers is much closer, but it’s coming for you as well. Have you thought about it?

We have. You started to answer the question. It’d be easier for me to cop out and say, “Well, our category’s different,” but we do think our category is different. And then we’ll get back to the business model that we think we’re building that’s pretty durable for any potential aggregation change, I would say, over time. Real estate is probably the most, as you said, complex, regulated, localized, licensed category out there. It’s very hard to think about whether there is a better way to shop for shoes or a better way to find the restaurant I’m going to order delivery from in the next 20 minutes. Every category has its challenges and how aggregation might change, but I think real estate is different for the structural reasons, also for the consumer reasons. You’re not doing something [like this] every day. You are buying a house. The average home buyer right now is once every 14 years.

But if you think about Zillow as entertainment—say I’m walking down the beach and now instead of opening Zillow to see how much that house costs, I just asked Gemini or Siri or whatever, and I get an answer and I get some transaction listing from the public database, that means I’m not opening your app. And then maybe I’m less likely over time to open your app when it’s time to do the other thing. That cycle is already quite good for you.

Or they become a new source of demand and sure, maybe they want to ask the horizontal LLM a question that they’re asking us right now. And then when they get into the category of actually transacting, which, again, is the business that we’re trying to build, they’re going to want to use our software workflows, data and people to transact. It’s just a different source at the top of the funnel. This is the age-old question when Google came around and when mobile and apps came around, the top of the funnel versus the transaction flow.

That is why we get really excited about the software, the workflows, the data. The licensed professionals that are in the category. We think that building that vertical integration, as you put it, is really the most durable component of any evolution, even if the way people get to our front door might change over time. And then on the flip side, we also are going to look to innovate on those capabilities because there is the horizontal ask, a horizontal “do it.” There’s also the ask of vertical to do it, which we think in this category also gets really interesting.

I’m curious about vertical searching; who really knows where that’s going. I think that’s wide open because the models just keep getting better. You see those opportunities open up and close down really fast. It sounds like you’re saying, when you say it’s more durable, that serving the enterprise customer, the business customer to actually make the money is where the growth is, and maybe it’s not in the consumer experience. And I wonder about that because it echoes things I’ve heard from so many people. The consumer experience of using the internet is maybe consolidating, maybe it’ll all just be chatbots, it’ll be AI powered browsers, who knows what’s going to happen. But at the end of the day, the real estate agents still need to buy and sell the properties and you can sell them software forever.

I would have said it a little differently, which is that maybe we care more about the transactors than the dreamers over time, but that’s already where our business is. We make money with the transactors. The dreamers and the shoppers and the browsers are efficient tomorrow’s transactors and we love filling that and the brand stretching across both is a huge Customer Acquisition Cost (CAC) advantage and a huge cycle, lifetime value customer. So that part we can talk about, but I don’t think we would pivot to be B2B focused over B2C focused. I think it’s more you would probably think about the transactors as the customer segment, as the more durable part of the customer segment.

It’s those customer acquisition costs I’m actually really interested in. If that starts to go up, because we have fewer dreamers just looking at houses.

80% of Zillow’s traffic comes to us free and direct now. Yes, if that percentage went down, that’s a different customer acquisition equation and we need to be good at where the traffic’s going to come from when they get into the category of transactions.

All right. Let me ask you the question I think everybody thought I was going to ask when I talked about AI. It’s the first question my wife asked me. It’s the first one that comes up. Every platform that has user content on it—and I would count Zillow as a platform with user content; it’s a little different, but it’s still user content, it’s not yours—is overrun with slop, like pure AI slop. And I see this in listings now. Actually, you initially saw it with over HDR photos, so then the photos you were looking at had no relationship to reality, and now you’re seeing virtual staging that’s getting maybe one tick too aggressive. It’s not actually so easy to just rip out the floor, but here we are, we’re just going to rip out the floor. And now you’re doing fully virtual tours, like generated 3D tours that maybe show you reality and maybe don’t. How are you guarding against that? Because this is, I think, the other big threat to your demand.

Two ways. One, to go back to our earlier conversation, having a professional database that has some rules around it is good for this. One of the benefits of an MLS is we are one member; there are many members and we all want to solve that problem. And so you can have compliance and rules and violations and fines. And you have licensed professionals who are endorsing and stamping that the content is accurate. When it gets through, that’s maybe a licensed professional who didn’t do the right job. You have a way to police or manage that. So that’s one way.

As you talked about earlier, we’re not going to just rely on the mechanisms of 500 plus [groups]. We hope, and we’re going to lobby for that, but we’ll also look at the data ourselves. And our rentals marketplace example of that where you don’t have the MLS structure and it is long-tail users enter directly into Zillow’s tools until we have to build good compliance tools. The nice thing is the benefit in doing it in the real estate market is probably lower because ultimately you’re trying to sell the home or rent the place. If you do it, it might drive top of funnel volume, but the incentive will quickly get squashed because you won’t actually rent or sell the place.

Reality shows that that dynamic has not played out. You see it everywhere. Every listing I see now has at least over HDR and photos, if not the full—

Not every [listing], but you definitely have the content challenges. The reality is that at the end of the day, someone’s going to go tour the home before they buy it. People aren’t buying… Well, actually that’s not true. Some people do buy houses sight unseen, but then they’re usually entrusting a professional to go see the house while they’re doing it. So someone’s putting eyes on the house. Again, you’re not buying shoes off Amazon and they’re like, “Well, wait, that’s not what I thought I bought.”

It feels to me like the counter-incentive to not lie about the photography has not actually resulted in the photography being more accurate.

But the flip side of that is we’re trying to use the AI tools to generate a more realistic version, which I think is part of the challenge of getting things like Showcase going is people who want to market the home maybe don’t want to show. So we take a good example, like we take drone photography, they’ll shoot it in the best possible light and then we’ll generate a full flyover. You can joystick around the house and see it from all angles. And we fill that in with LLM tech. That showcases things that maybe they didn’t want to show and that gives you a more realistic sense of the home.

Trying to navigate that balance of, “Maybe I didn’t want to market it that way,” but “Oh, if I Showcase it this way, I’ll actually drive more demand.” That’s one way we’re really trying to drive the right incentive is we can show that when you generate a more realistic version of things, it actually provides more demand and more pull through from buyers if you give them more realistic content, not just staged content. And we’re getting to 3%, 4%, 5% of listings are now doing that and we just started that a couple of years ago. That’s a good example of tools we’re bringing to bear and using AI to actually create a more realistic version to maybe counter that incentive.

Would you ever do the opposite and penalize the fake imagery? A lot of platforms are having to consider this.

I think MLSs are going to have to rule with it. We’re not in the business of trying to figure out how to create the rules and the fines for licensed agents. We’re not the associations of MLSs. They are grappling with this concept right now and we for sure will have a point of view on it.

What’s your point of view?

We want to create the most fair and transparent and accurate version of a listing as best we can. We started the conversation, talked to the seller. We started with the buyer and a lot of our innovations are around, “Look, more transparency is good. Yes, you want to try and hide the bad stuff, but they’re going to find it when they come see the house anyway.” So now you’re just wasting the seller’s time with people walking through the house who aren’t going to buy the house. So virtual touring is actually a good example of trying to drive incentives to the buyer, to give buyers more empowerment which actually helps the seller, even if it is an indirect thing for the seller.

All those assets are really complicated. I feel like I could do another full hour on just what technology we’re using to create virtual tours. Let me ask you about that. Every view of a database with its own custom extensions—and this is just again, very reductive, very abstract—there’s always the opportunity to extend the standard, to build the proprietary version, to capture an audience by saying you can only get this here. It sounds like Showcase is one of those examples. How broadly do you think about that dynamic? There’s the public database and then there’s, “We’re going to add to it. We’re going to add our own proprietary information to this,” and say to every buyer, “If you want the cool drone flyover, you’re only going to get that in Zillow.” How does that dynamic work? Because even—I don’t know, you had a dispute with Matterport. There are these other providers that would love to be in your views and then there’s the opportunity for you to capture that value yourself. How do you think about the bounds?

I think our broad philosophy is: any innovation today is going to become commoditized in table stakes. So we want to lead with innovation and create the right incentives for the marketplace to use the innovation. Because back to you said, the biggest challenge we’re trying to get Showcase more used is it’s really hard to do. People have to change their workflow. Now, the photographers are changed on it. There is this, “I don’t necessarily want to show all the things that Showcase would show in the house.” It’s like, “But you should. It’ll sell the house faster.” “But it’ll show this room I don’t want to show.” There is that incentive. Innovate and create the right incentives for people to use it, but it’s going to get quickly commoditized. And that’s part of why we’re even open with all the rich media stuff.

So yes, we have our own tech. We support everyone else’s tech. We support a bunch of 3D interactive floor plans on Zillow because if we can just get more folks using the tech and the tools, more transactions will move from offline to online. That’s still the biggest shift we’re working on because our business is transaction volume based. We benefit from getting more transactions flowing online and digitized. We’re going to bend towards open and we’re going to bend towards platform support because we still need to get all this tech, whether it’s Zillow’s or not, to be across more of the content.

Let me ask you one last big-picture question. Then I want to ask you what’s next for Zillow, broadly.

Whenever I talk to people who use platforms, big platforms, there is this sense that they’ve lost agency. And I think this is very true of the consumers on consumer platforms. It is extraordinarily true of creators on the consumer platforms. You ask a TikToker if they feel a tremendous amount of agency about TikTok, they do not feel that way. I think more and more consumers do not know what the algorithms are showing them.

Real estate is a different business, a different market. You’ve said many times Zillow is small, but to many participants in this ecosystem, you’re the platform and your rules are what make things happen. And as you build more software for the agents, the way your platform is designed will shape more of how they do things. Certainly for buyers, the idea that their housing stock options are limited to what they see in Zillow feels like a constraint. How do you give everybody in this ecosystem more of a sense of agency? Because you’re the big platform and they don’t get to yell at you the way that I get to yell at you.

The really unique thing about this category is—and we’ll talk about transactors, buyers and sellers, not people who are just using it as entertainment—you are selecting from a too-limited set of what you want and having to make a set of trade-offs. In that marketplace where supply is so constrained, the buyer actually has a tremendous amount of agency. It’s more about actually making sure you hire the right professional and making sure you end up making the right decision, which there’s no A/B test for. You won’t really know until you do it and then you look back. That’s actually where the agency comes from. Maybe if we’re ever in a market again where there is an extreme abundance market and buyer choice is not so constrained, it would feel differently. But right now, if I’m a buyer, the agency I have is that eventually I’m going to see all the homes that I possibly could buy before I make my decision.

What about on the seller side, on the agent side specifically? Because I know they feel constrained.

Yes. The big challenge they are trying to solve is, “How can I run my business more efficiently?” The agents that are still growing in this market are the better agents because they have been able to thrive in probably one of the lowest levels of volume that we’ve seen in a long time. Helping give them more agency is—yes, you could say we’re the platform, but making them more efficient, whether it’s our tools or other tools, is a really good way to give it. And I’ll give you an example of how we’re trying to do that.

For a long time, we just thought about, hey, you’re getting a bunch of customers from Zillow. Zillow has buyers and sellers and you collect them and then, okay, we benefit if you convert them and they think about Zillow’s customers and they think about their other customers and they’re just two separate worlds.

And they’re like, “Well, I really like a lot of the data I get about Zillow’s customers. Can you help me with those insights about the database of all my other customers?” And so we’re now launching a software tool that doesn’t just give them the thing they’re used to, which is if they want to use data on Zillow buyers and sellers, they can, and if they don’t have Zillow buyers and sellers, we can’t help them. We’re actually helping them connect the intelligence and insights from Zillow to the rest of their data, even if they’re not Zillow customers. And so you could say that’s Zillow providing more software and tools, but it’s also us opening it up and giving them more agency to run the rest of their business the way they want.

Zillow’s 20 years old; you’re about to leave here and go ring the bell in the stock exchange. What’s next for the company?

Building out the rest of this integrated transaction. We want to get to a place where you can open the app—the experience you had that you talked about, the next time you do it, you’re messaging with your agent side of the app, you’re getting pre-approved in the app, you’re closing the app, you’re signing in the app, and you feel that it’s more digital. And it actually does work the way the rest of the apps on your phone work. It does feel more like a one-stop shop.

Jeremy, this has been great. You’re going to have to come back. I have many, many more questions about database.

_{Questions or comments about this episode? Hit us up at decoder@theverge.com. We really do read every email!}