Online real estate marketplace Zillow has brought to home buying and selling what a previous generation of travel websites provided to shoppers wanting to compare the prices of hotels, rental cars and airline flights — transparency. But finding data that is reliable across the board can be difficult, according to Zillow CEO Spencer Rascoff. And the key is not just simply to offer the information, Rascoff said during a recent conversation with Knowledge at Wharton and Wharton real estate professor Susan M. Wachter; it’s about the various sub-models the company’s software incorporates.  

Zillow has a database of more than 110 million U.S. homes that includes homes for sale and homes for rent. It also incorporates Zestimate, which helps calculate home values. Zillow is the leading real estate website, Rascoff notes, but only about 12% of Americans have heard of it. His challenge is to bring that number up to a 90% level.

An edited transcript of the conversation follows.

Knowledge at Wharton: We’re starting to hear that the housing market is recovering. What do you see based on the data at Zillow? How even is the recovery across different regions?

Spencer Rascoff: Well, housing is back. In 2012, home values were up 6% year-on-year. In 2013, Zillow is forecasting 3% appreciation, which is a bit on the conservative side, as compared with other economists. But we’ve had 16 months in a row of appreciating home values, which is great. There is very tight inventory across the country. And one of the reasons for that is so many homeowners still have negative equity. It’s difficult for them to sell. There’s not that much inventory. And that’s because of low mortgage rates.

That’s creating bidding wars in a lot of areas. So, we’re seeing some parts of the country appreciating at unsustainably high rates — Phoenix is over 15%. South Florida is appreciating very quickly. That’s a little worrisome. I think those rates of appreciation are unsustainable. But, overall, this is a much better housing market than it has been in the past several years.

Knowledge at Wharton: What do you think, Susan?

Susan M. Wachter: I agree. I think, though, the months of appreciation — 16 months and ongoing — restore confidence. The negatives in the market have been a lack of confidence — an unwillingness to put money into housing. Most people know someone who has lost $40,000 [in the housing market.] But this is a different world. This is a world where housing price rises are sustainable. They will continue. I also concur with 3% — I’m forecasting 5% — price rises this coming year.

Knowledge at Wharton: That’s interesting. Spencer, let’s go back to before Zillow was launched. You founded Hotwire, which you sold to Expedia. Then you worked for Expedia as vice president of lodging. How did your background in technology and marketing help shape Zillow?

Rascoff: When we sold Hotwire to Expedia, I and my management team at Hotwire joined the folks who had created Expedia within Microsoft, about 10 years earlier. The whole thesis behind creating Expedia was to provide information transparency to consumers. Prior to Expedia, as you probably recall, you would have to speak on the phone to a travel agent to make travel arrangements. You could hear them type-type-type over the phone [looking for the lowest rates.] You’d want to jump through the phone and see for yourself: What secret database is that travel agent looking at that I don’t have access to?

That’s what Expedia did. It turned around the screen, so that you as a consumer could have access to the secret database. When we were at Expedia, we started looking at other industries that hadn’t been revolutionized by the Internet yet. It was obvious to us that the real estate industry was among them. There were these secret databases that only professionals had access to. There was all this great data locked up in these databases and also in county courthouses.

At Zillow, we aim to unlock that information and make it readily available to consumers. The difference, though, is that in the real estate space, unlike in the travel space, there will always be an intermediary — somebody involved in the transaction [besides the buyer and the seller]. That’s fine by us. We actually think that real estate transactions are so infrequent, so expensive, so highly-emotional and so complex, that there will always be and should always be a professional involved in the transaction.

So, we have a very different business model at Zillow. We sell advertising and software tools to professionals. That’s how we make money, rather than being involved in the transaction, which is how Expedia makes money.

Knowledge at Wharton: Susan, what do you think of the business model that Rascoff just described?

Wachter: Zillow’s a game changer. I can’t talk about the profitability of the company, but what it has brought to the consumer, to the homeowner, in terms of valuation of their home, is extremely important, especially in these years of extraordinary volatility. People want to know what their home is worth. Has the value in fact fallen by 50%, which homes in many parts of the country have?

Do you have to actually go to someone, to an expert, to get several different views [of the market]? No. Zillow is getting the power of all the data and bringing it together in very sophisticated analytic models. It’s there for you to test. And, if it appears to be off in some ways, there is now a way of going back. This is extraordinarily important.

You need to know the value of your home today for many reasons: Are you going to sell? How does it compare to your mortgage amount? This democratizes data, and it makes the market more transparent, and more efficient. All of this is very important to a market that has seen its troubles. This helps bring liquidity back.

Knowledge at Wharton: One of the things that I found most interesting on Zillow was the Zestimate. I promptly, of course, went and looked up the value of my own house. How do you figure out what the house is worth? How does your algorithm work?

Rascoff: Well, the short answer is the database. The software looks at the property attributes of your home, and compares it with like homes. The much longer answer is, at each census-track level — census track is basically a couple of neighborhood blocks — we have a meta-model that sits on top of many other sub-models. For example, one sub-model might look at tax assessment, another might look at distance to an arterial, a third might examine dollars per square foot and a fourth would look at just regressions against the number of bedrooms.

The meta-model, at the census track-level, tries different weightings of each of the sub-models, going back 15 years. And it’s constantly improving itself by changing the allocations, the relative weights of the different sub-models, to be as accurate as possible. So, as time passes, as more and more data comes into the system, it becomes more accurate.

About a third of homes in the U.S. now have had their property attributes changed on Zillow by homeowners or their agents. So, much as Wikipedia is a living database of the world’s knowledge, Zillow has become a living database of all the property parcel information in the U.S. That allows Zestimate and other data on Zillow to become more accurate over time.

Knowledge at Wharton: Does it apply only to existing home sales or also new homes?

Rascoff: We also come out with Zestimates [for] new homes, based on the property records that either come to us from the county [where the home is located] or from the listings feeds [sent by] home builders. Also, when we get new information about a home, the Zestimate will change.

Knowledge at Wharton: Susan, what is your view of the process Rascoff just described? How sound, from a pricing standpoint, is it?

Wachter: It’s a great process. It’s using what we call hedonic analysis. It’s taking the attributes of the property and weighting them, to efficiently use it statistically to come out with the best estimate. There are other ways of doing this. You could do repeat sales, and that avoids the structure of the home, and the differences in the structure of the home. What’s interesting about this is it incorporates not just the value changes in the neighborhood — or the nearest neighbors — but also the attributes of your own home. And that’s an advantage.

Knowledge at Wharton: A question for both of you: I’ve heard some criticism, especially from realtors, that the Zestimate, or the pricing model that Zillow uses, may have some flaws. There are some factors that just can’t be captured, and automated. Do you agree with that criticism?

Rascoff: I agree with that. I mean, it’s computer-generated on every home in the country three times a week. No employee of Zillow has ever been in your home. For a computer-generated valuation, we believe it’s quite accurate, and it’s a good starting point. We call it a Zestimate, not a Zappraisal. It’s meant to be a starting point. If you want to figure out a more accurate determination of your home’s value, we encourage you to talk to a real estate agent or an appraiser, who will come to your home and refine the valuation.

Wachter: Absolutely. If you have something special, like a pool for koi fish, let’s say, we’ll never know. And you’ll know it’s very important to you, it might be important to some buyer. So, that’s why it’s a basic estimate. And there may be some negatives as well, that might not be incorporated.

Knowledge at Wharton: If you take some of the other companies that are active in the same space as Zillow —, or Trulia — everyone has access to the same basic data, which is from the Multiple Listing System, or the MLS. What is your strategy to create a sustainable advantage for Zillow, based on the fact that you all start with the same data?

Rascoff: Well, you’re right. At the core, it’s fundamentally commoditized listings information. This is why we don’t focus exclusively on that. If you think of the entire housing stock of the U.S., 110 million homes, about 3%, are for sale in the MLS at any point in time. Everybody has that information. And that’s [available] on tens of thousands of other websites.

We tried to innovate on top of that. For example, by having Zestimates that show historical Zestimates. By showing sale data going back 15 years. Tax information. Making predictions of what’s going to happen to that home’s value going forward. But fundamentally, the price, the photos, the square footage, that’s on every website. And so, that’s not enough.

What we then do is have information on hundreds of thousands of “Make me move” listings — when an owner posts a dream price for their home, new construction listings, for sale by owner listings…. We’re the only website with over a million foreclosure and pre-foreclosure listings posted for free for consumers. And then, of course, we have market context. We have information on every home in the country. We have ratings and reviews, 250,000 reviews of real estate agents that consumers have left on Zillow. So, we do our best to innovate not on that commoditized information, but also on all the other stuff, which we think is what differentiates us from competitors.

Knowledge at Wharton: Susan, do you think this is enough to create a sustainable competition advantage for Zillow? Or can others catch up?

Wachter: The key is the algorithms, which one can improve over time, and which become self-improving with additional data. Adding data — particularly individual input that others don’t have or can’t organize — makes it a sustainable model.

Rascoff: Right. The platform shift from desktop to mobile has also been a great boon to our business.

Knowledge at Wharton: How so?

Rascoff: We’re a software company. We can devote massive amounts of resources to create software, especially on mobile, where it’s very complex to write software across multiple mobile platforms. So, for example, we went public about a year-and-a-half ago. When we went public, 20 homes were viewed every second on Zillow, on mobile. A year-and-a-half later, 75 homes are viewed every second. Then, we had five mobile apps. Today, we have 24. More homes are now viewed on Zillow on mobile than on the desktop.

As more and more usage shifts toward the mobile, it’s difficult for individual brokerage websites, at a local level, to compete, because it requires so much software to create great mobile experiences, and great mobile apps. It’s difficult for even other national players to compete. So, we definitely benefit from that.

I just saw some interesting data from Domino’s Pizza. At their investor day, they put up a slide that showed the national market share of Domino’s and Pizza Hut over the past 10 years, versus local pizza places. It was basically flat between national and local for the past 10-plus years. And then, about two years ago, something interesting happened. The national guys gained a couple points of share, really out of nowhere, versus the local pizza stores. The reason is the smartphone. All of a sudden, people are now ordering pizzas from Domino’s and Pizza Hut’s mobile apps. But the local pizza guy can’t really compete with that. A big shareholder of ours showed that to me, and said, “I bet you never thought you’d have something in common with Domino’s Pizza. But the way you’re dominating on mobile is very similar, and you’re seeing this in other verticals as well.”

Knowledge at Wharton: That’s really fascinating. One more question: How do you insure the accuracy of your data? I was talking to some real estate developers, and they said that for some markets, your data is absolutely spot-on accurate. And other markets, maybe not so much.

Rascoff: This is an industry-wide problem. The data on Zillow comes partly from county records. There it can be corrected by owners or agents. But it also comes to us from agents. For example, a listing agent posts a home for sale. If the home sells, he doesn’t take it down, and we have a stale listing.

So, we’ve deployed a lot of software, a lot of business expertise, and a lot of money toward solving this problem. We’ve come a long way. But, it’s an industry-wide problem. I think it will continue to be something that we work on for years.

Wachter: It is a problem. It’s a problem in part because our sales records are at the county level. It’s actually in some ways stunning that we as a country do not collect the data and don’t verify the data. Some states and some cities have really poor collection methodologies.

Knowledge at Wharton: You’re also doing some interesting work in the area of rental housing. Could you tell us a little about some of the challenges involved in that?

Rascoff: Rental is a very big area of investment for us. We bought three companies in the past 18 months for over $60 million. We have about 100 people out of 600 at the company focused exclusively on the rental [market]. The reason that we’re so excited is because it’s a $5 billion market. Property managers and landlords spent $5 billion advertising their listings to consumers. The structure of the industry is quite different from the structure of the for-sale industry. Notably, it’s disorganized.

In most markets — outside of Manhattan — they’re not broker-controlled markets. So, property managers and real estate developers market their listings directly on a many-to-many approach, rather than going through a centralized MLS, or a centralized brokerage system. That allows us, we hope, the opportunity to create the marketplace.

To give you some numbers, there are four million rental units available in the U.S. at any point in time. Craigslist has about 750,000 of them. We have about 600,000. A year ago, we had 300,000. Most other sites have about 100,000. So, the strategy is to give away property management software to the industry for free in order to out-Craigslist Craigslist — to end up with many more listings than Craigslist. Then, hopefully, we will become the marketplace, and we can monetize it. So, today, we don’t monetize rentals at all. We’re delaying that gratification until we are the marketplace, and then we can focus on monetizing it.

Knowledge at Wharton: Susan, what do you think of Zillow’s strategy?

Wachter: The rental market is growing in the U.S., so it’s important to have that coverage. It’s less organized, so that’s a real strategic move. But even more so, going forward, there’s a lot of information, the comparisons of rents to values. And to be able to get that accurately, that, too, would be a game changer.

Knowledge at Wharton: Let me end with one last question. In so many ways, Zillow has been highly successful. And to take just one metric, since your IPO in 2011, your stock price has gone up to $55. I just saw a report from Goldman Sachs setting a $60 target for it. But regardless of what happens to the stock, what is your definition of success for Zillow?

Rascoff: My definition of success is about creating an everlasting and enduring brand that my children, or maybe someday grandchildren, will know. There are very few technology companies that have created brands that endure for 10, 20 or 30 years. Microsoft is really the only one, and I suppose Apple.

To me, it feels like we’re just getting started. We have less than 1% revenue share across the $35 billion in advertising in our space. We have a relatively little-known brand, relative to the size of the opportunity. Only about 12% of Americans have heard of Zillow. And yet we’re the leading real estate website. If you look at other categories, 90% will say E-Trade, or, 90% will say Expedia, 90% will say WebMD … [only] 12% will say Zillow.

That, to me — as CEO of Zillow — could be depressing. But I actually get very excited about the opportunity. To me, success is closing that gap. If we do that, a lot of other great things will follow.