Zillow Group Inc (Z) 2017 Q3 法說會逐字稿

Good day, ladies and gentlemen, and welcome to the Zillow Group Third Quarter 2017 Earnings Conference Call. (Operator Instructions) As a reminder, this conference call is being recorded. I would now like to turn the conference over to your host for today, RJ Jones, Vice President of Investor Relations. You may begin.

Thank you. Good afternoon, and welcome to Zillow Group's Third Quarter 2017 Financial Results Conference Call. Joining me today to talk about our results are Zillow Group's Chief Executive Officer, Spencer Rascoff; and Chief Financial Officer, Kathleen Philips.

During the call, we will make forward-looking statements regarding future financial performance and events. Although we believe the expectations reflected in the forward-looking statements are reasonable, we cannot guarantee these results. We caution you to consider the risk factors described in our SEC filings, which could cause actual results to differ materially from those in the forward-looking statements made on this call.

The date of this call is November 7, 2017, and forward-looking statements made today are based on assumptions as of this date. We undertake no obligation to update these statements as a result of new information or future events, except as required by law.

During the call, we will discuss GAAP and non-GAAP measures. We encourage you to read our financial results press release, which can be found on our Investor Relations website, as it contains important information about our reported and non-GAAP results, including reconciliation of non-GAAP measures.

In our remarks, non-GAAP financial measure adjusted EBITDA is referred to as EBITDA, which excludes other income, depreciation and amortization expense, share-based compensation expense, acquisition-related costs, Internet interest expense and income tax benefit.

This call is being broadcast in the Internet and is available on the Investor Relations section of Zillow Group's website. A copy of management's prepared remarks has been posted to the Quarterly Results section of our Investor Relations website. A recording of the call will be available later today.

We will open the call with prepared remarks followed by live Q&A. In addition to taking questions from those dialed into the call, we will answer questions asked via Slido. We encourage you to visit www.slido.com, where you may submit questions by entering the event code #ZEarnings. You may also vote on which submitted questions you want us to answer.

I will now turn the call over to Spencer.

Thank you, RJ, and to all of you for joining us. We're hosting this call today from our office in Irvine, home to around 300 employees, including our Premier Agent sales and support team, our data acquisition team, listing feeds team, B2B software tools -- internal software tools team and RealEstate.com.

Zillow Group delivered terrific third quarter results. Total revenue for the third quarter of 2017 grew 25% year-over-year to approximately $282 million, which exceeded the high end of our guidance range by nearly $4 million. Q3 GAAP net income was more than $9 million or more than 3% of revenue. EBITDA for the third quarter exceeded $70 million or more than 25% of revenue, ahead of expectations due to strong revenue growth across all marketplaces and cost savings.

In addition to strong financial results in Q3, we also made progress on our key corporate priorities. The first priority is to grow our audience size and increase consumer engagement. Traffic to Zillow Group brands' mobile apps and websites reached more than 175 million average monthly unique users in the third quarter of 2017, an increase of 6% year-over-year. In July, we reached an all-time record of more than 187 million unique users, which was up by more than 17 million unique users from the same period last year. Our advertising and other marketing initiatives continue to be effective in driving more consumers to Zillow Group's mobile apps and websites, even at such a large audience scale.

Further down the funnel, visits reached nearly 1.7 billion during the third quarter, up 19% year-over-year. Visits growth was driven by a variety of product improvements that increased consumers' propensity to return to our mobile apps and websites, including initiatives to encourage mobile web visitors from search engines to download our apps. As a reminder, the visits metric helps us evaluate progress towards our goal of increasing audience engagement. Users who visit frequently have a greater intent to buy, sell or rent a home, which ultimately means more high-quality leads for our agent advertisers.

Our next priority is to continue growing our Premier Agent business, which makes up about 70% of our total revenue. In the third quarter of 2017, Premier Agent revenue grew 24% year-over-year.

We continue to focus on helping top-performing agent advertisers grow their businesses by advertising with us and by using our software tools. Last year, we introduced 2 exciting new advertising options to better align our Premier Agent advertising products with the way real estate professionals run their businesses, including the Premier Broker program and team accounts. These programs allow agents and brokerages to purchase advertising with us in a new way that results in a better consumer experience, since they can more efficiently service the leads received from our platform. This strategic shift in the way agents and brokers advertise with us made the total advertiser account metric less meaningful because it did not represent the total number of agents advertising on our platform. However, we still get many questions about how many agents are using the Premier Agent program, so I want to clarify that, today, more individual real estate agents receive leads from Zillow Group than ever before, thanks to the growth of Premier Broker and team accounts.

Since the third quarter of 2016, the number of Premier Broker accounts has grown by 260%, and the number of team accounts has grown by 87%. Premier Broker and teams range from 2 to more than 130 agents under one account.

Just a few weeks ago, we hosted nearly 2,000 agents at our annual Premier Agent Forum in Las Vegas. At the forum, we announced several exciting new features designed to help agents and brokers convert more leads into transactions. I want to share a few of the products and features that we announced at the forum.

Last year, we launched Premier Agent Direct, a program that allows agents to advertise their listings and their brand on Zillow, Trulia and Facebook. This year, we expanded Premier Agent Direct to include postal mail, which is still the advertising medium used by most real estate agents. With this program, printed postcards are automatically generated and mailed on behalf of Premier Agents, providing another marketing touch point with consumers.

We also announced some exciting updates to the Premier Agent App. For the first time, we're adding features to the app that expand its use beyond lead management and into business and transaction workflow. For example, our newest feature, called Tasks, allows agents to create a checklist to stay on top of their commitments to clients and prospects. We're also making the PA App a single-source communications platform. Premier Agents can now text and e-mail clients right from the app, allowing agents to manage their client communications all in one place.

The Premier Agent App is now a powerful productivity platform, and we believe it is the most widely used B2B app in the real estate industry.

Another feature we're excited about is called My Agent. We know that after initial contact via a Zillow Group mobile app or website, Premier Agents typically move clients to another internal system to share listings and communicate, and yet the consumer will come back to Zillow or Trulia almost 30 times in the next month. Now, when we observe through the app that a Premier Agent is actively working with a buyer, we will replace our traditional Buyer's Agent List that the consumer sees on Zillow and Trulia listings with a contact box that features only that consumer's chosen agent. Since no other agents will be visible, it will solidify the agent-consumer relationship, making it much easier for them to work together.

This 2-way contact module, where agents and their clients can communicate about homes, will help agents convert more leads into transactions by keeping that agent in front of the consumer throughout their search. For the consumer, they are able to search on a platform they already are familiar with and love. In addition, this will make the Zillow Group experience even better for Premier Agents by improving the quality of leads.

Prior to the forum, we announced an exciting new product that allows agents to create 3D home tours on their iPhone and post them for free to their Zillow Group listings. The Zillow Group Home Capture App will bring an incredible 3D home tour experience to buyers, letting them narrow down their home searches before visiting in person. I'm especially excited about this app because, up until now, capturing 3D images was really expensive and time-consuming for agents. We've been working for the past 2 years on a way to democratize this technology and make it freely available. It's hard for me to overstate the level of creativity and sophisticated technical knowledge that has made this product a reality. Stitching together still photos taken from an iPhone to create an immersive 3D panorama tour requires teams intently focused on machine learning and image recognition. We're currently testing the app in one market and anticipate its nationwide rollout in 2018.

Moving on, as you may recall, we're currently experimenting with a product called Zillow Instant Offers. The Pilot is testing in 2 markets, Orlando and Las Vegas, and soon, Phoenix. Instant Offers provides motivated sellers with an opportunity to easily submit information about their home through Zillow. Those potential sellers then receive a comparative market analysis and a listing presentation from a Premier Agent alongside actual offers from investors. If they choose an investor offer, the consumer gets control and certainty on the sale of their home in a predictable time frame. The purpose of the test is to determine where sellers see the balance between convenience and price. It is still early in the experiment so we're not sharing specific results, but we are seeing a positive response from agents, direct home buyers and sellers. We identified an opportunity to innovate on this emerging trend and to provide more transparency for consumers while delivering high-quality seller listing leads to Premier Agents. We are the only player in this growing market category that keeps agents at the center of the process. We believe that an agent's role as trusted adviser is incredibly valuable to consumers and will be a permanent part of most real estate transactions.

Our third strategic priority is to grow our emerging marketplaces of mortgages, rentals, New York City and new construction.

Our Mortgages revenue is growing at a healthy pace ahead of the industry, largely because of 2 key advantages. First, most of the activity in Mortgages is driven by homebuyers looking for purchase loans, and that is increasing, thanks in part to sustained low interest rates and strong housing demand. While overall industry mortgage lending volumes are down due to historically low refinancing lending, we continue to expect our revenue growth to outpace the growth rate of the industry, as the bulk of our mortgage activity comes from consumers seeking purchase loans.

Our second key advantage in Mortgages is its connection with our real estate sites, where potential borrowers are already shopping for homes. These 2 marketplaces support one another, and just like with real estate, we are continuing to test and experiment in the mortgages space. The mortgage and home shopping experiences are inextricably linked, and we feel that we can bring significant innovation to the space for the benefit of home shoppers and borrowers.

In our Rentals marketplace, revenue grew 56% year-over-year for the third quarter of 2017. Our long-term rentals strategy is to drive engagement through personalization and machine learning to get the right properties in front of the right renters fast.

Last month, we launched Rental Inform, a cloud-based data dashboard with exclusive real-time, robust rental market and aggregated consumer insights. Rental Inform is powered by anonymized, aggregated search data from the approximately 33 million monthly unique rental users that visit Zillow Group's brands, mobile apps and websites. Property managers can now access that data to help them better understand consumer preferences. The data provide insight to what type of floor plan, amenities or price points are most desirable in an area as well as details on the current rental market and pricing trends. This is similar to our Builder Inform data tool set that we rolled out earlier in the year.

We are also making the Zillow Group Home Capture App available for property managers and landlords to shoot 3D tours for their rental units.

In our New York City marketplace, the rollout of Premier Agent on StreetEasy and rentals paid inclusion for the New York City Rentals Network had exceeded our expectations. The addition of these 2 new advertising products contributed to our strong third quarter 2017 total revenue performance.

Our New York City industry relationships remain strong. During the third quarter, we expanded our multiyear agreement with Realogy, the parent company to Corcoran Group, Citi Habitats and Sotheby's, to have their for-sale and rental listings displayed on StreetEasy. This was a big win for consumers and Zillow Group, as these are some of the leading brokers in New York.

In a unique market like New York, it's very important to have local expertise, and StreetEasy's hyper-local focus has always been a major competitive advantage for us. New Yorkers love the StreetEasy brand. For the past decade, StreetEasy has focused on developing products that are specifically designed to capture the unique needs of New York City buyers, renters and sellers, and has spent tens of millions of dollars on New York City-specific marketing. Agents will always want their listings displayed where the largest audience goes for their home search needs, and in New York, that's StreetEasy.

For-sale listings in New York City across all Zillow Group brands attracted almost 13 million visits during the third quarter, clearly making us the brand of choice for home shoppers there.

Specific to the rental side in New York, agents realize the value that StreetEasy provides to their business and are choosing to advertise their rental listings with us. More than 10,000 local agents have already opted in to the New York City Rental Network since we launched in the second quarter of 2017. Consumers shopping for a rental in New York benefit from StreetEasy's high-quality inventory, since paid inclusion eliminated many stale and fraudulent listings. Agents benefit from the large audience on the New York City Rental Network, which received almost 14 million visits during the third quarter of 2017.

In our New Construction marketplace, we are encouraged by the addition of many new promoted communities to our platform as we work toward our goal of being the leader in residential new construction marketing. During the quarter, we continued to invest in this emerging marketplace with the acquisition of New Home Feed, a streamlined listing management technology that allows builders to input, manage and syndicate their listings across the web. This acquisition makes it easier for our builder partners to send their listings to Zillow Group and ultimately improve the quality and accuracy of new construction listings across our consumer brands. We look forward to partnering with more builders to help them build their brands and increase discoverability of their new construction homes.

Finally, our fourth priority is attracting and retaining the best talent and maintaining Zillow Group's unique company culture focused on innovation. We know that our employees are our most important asset. In September, Zillow Group was named to Fortune Magazine's list of the best workplaces for women. This was an important recognition for us, as diversity and inclusion at Zillow Group is a major initiative. We have invested significant resources to ensure that all employees feel valued and know that their achievements are appreciated, and we're pleased to be recognized for these efforts.

It has been a really exciting year so far for Zillow Group and the real estate category as a whole, and we're excited about what's next. Fortune Magazine, again, recognized us just a couple of weeks ago as a company poised for breakout growth by placing us on their Future 50 list, and we're focused on priming our company for that growth.

Technology is evolving faster than it ever has before, and we're seeing consumer expectations rise in every category, including real estate. We know that delighting the consumer will mean something different in 2018 than it did in 2008. For home buyers, sellers and renters, the desire to simplify and shorten the process is growing. We're already exploring ways to satisfy that desire, with 3D tours and our experiment with Instant Offers as a couple of examples. You can expect that kind of innovation from us to continue in the coming years.

Our position as a category leader, along with our employees' technical prowess and creativity, put us in a position to continue growing and strengthening our own business and to help our valued industry partners do the same.

Before I turn the call over to Kathleen, I wanted to discuss the impact of the catastrophic hurricanes and fires that have devastated regions across our country over the past few months. Our thoughts are with the people of Texas, Florida, Puerto Rico and California, where thousands of homes and businesses were damaged or destroyed, leaving countless people without a place to live and where lives were lost. Many Zillow Group employees have volunteered their time and energy to help with the recovery and to raise money, and I want to thank them all for this dedication.

With that, I'll turn it over to Kathleen for a financial review and our outlook for the remainder of the year.

Thank you, Spencer, and hello to everyone joining us on today's call. Let's dive into our financial results.

Total revenue for the third quarter increased 25% year-over-year to a record of $281.8 million from $224.6 million in the same period last year.

Premier Agent revenue grew 24% year-over-year to $197.1 million in the third quarter, resulting in a 5% year-over-year increase in Premier Agent revenue per visit. Our auction-based pricing platform continues to perform well. As traffic and engagement increased across all brands, demand from agents to advertise with us increased in some of the country's hottest real estate markets. Nationwide adoption is trending as expected based upon our experience with the pilot that we conducted in a small number of markets. We continue to see top-performing agents increasing their spend with us to maintain their competitive advantage in their markets, while other growth-oriented agents are expanding their reach and brand by spreading their ad spend to neighboring ZIP Codes where their potential ROI may be higher. In both cases, we have more agents increasing their advertising spend with us over time.

Top-performing agents, brokers and other teams continue to increase their spending with us as they realize the benefits of advertising on our platform. For example, revenue from same agent advertisers, or those who have been on our platform for more than 1 year, grew by 45% compared to the prior year. New sales to existing advertisers made up 52% of total bookings in the third quarter. The number of Premier Agent accounts spending more than $5,000 per month grew by 98% year-over-year and increased 88% on a total dollar basis during that period.

In response to the hurricanes and the Northern California wildfires that occurred late in the summer, we worked closely with our Premier Agents, Premier Brokers and other advertisers in the affected areas to manage their advertising budgets efficiently during those challenging times. We estimate that these relief initiatives, which included billing credits and other forms of advertiser assistance as well as lost sales, impacted our third quarter 2017 Premier Agent revenue by more than $800,000.

Additionally, we expect that our fourth quarter 2017 Premier Agent revenue will be impacted by almost $1 million. We also experienced a temporary decline in traffic to our mobile apps and websites from consumers in affected areas during the month of September.

On our last earnings call, we shared that we were just in discussions with the Consumer Financial Protection Bureau related to a possible settlement concerning their investigation. We continue to have ongoing discussions with the CFPB but have not yet come to a mutually agreeable settlement. As a reminder, the CFPB has indicated that if a settlement is not reached, they may pursue legal action against us. We believe that our practices are lawful and that our co-marketing program allows agents and lenders to comply with their legal obligations. We will continue to engage in discussions with the CFPB and help to put this matter behind us as soon as possible.

Getting back to our financial results for the quarter. Other real estate revenue grew 55% year-over-year to $44.8 million. Other real estate revenue primarily includes Zillow Group Rentals, New Construction, dotloop as well as revenue from the sale of various other advertising and business software solutions and services for real estate professionals.

Mortgages revenue reached $20.9 million in the third quarter, which represents a 6% increase year-over-year. Average revenue per loan information request increased 31% year-over-year. Display revenue was $19.1 million, an increase of approximately 8% over the same period last year, driven by strong brand sales initiatives and increased traffic.

Shifting now from revenue to our expenses. Total operating expenses for the third quarter were $267.1 million. Our cost of revenue was $22.2 million or 8% of revenue. Sales and marketing expense was $107.1 million or 38% of revenue. Technology and development costs were $83.4 million or 30% of revenue. General and administrative costs were $54.2 million or 19% of revenue.

We ended the quarter with more than 3,000 employees across all of our offices.

Now turning to our outlook for the remainder of 2017. We are updating our full year revenue outlook to a new range of $1.068 billion to $1.073 billion, which represents 26% year-over-year growth at the midpoint of the range. We also are updating our full year EBITDA outlook to a range of $233 million to $238 million, nearly 22% margin at the midpoint of the range.

For a detailed fourth quarter and full year 2017 guidance, I encourage you to review our press release that was issued this afternoon and is available on our Investor Relations website.

With just one quarter left in 2017, we expect to end the year strong. It has been an exciting year for Zillow Group as we completely changed our Premier Agent pricing model and experienced significant growth in our emerging marketplaces. We look forward to the opportunities ahead for Zillow Group as we further solidify our position as the real estate industry's most trusted marketing and technology partner.

With that, we will now open up the call for questions.

(Operator Instructions) Our first question comes from Michael Graham of Canaccord.

I just wanted to get an update on the self-serve platform, specifically, you just came through a really seasonally strong period there. Ostensibly, there was a lot of activity on the platform. Just update us if you could on what you learned there, and on what level of seasonality you're expecting in Q4, if any, like how the possible sort of downward seasonality in Q4 impacts how you thought about guidance.

Sure, Michael. So the goal of switching the Premier Agent model to market-based pricing was to create a pricing model where agents with the highest lead conversion were the ones buying the most impressions. And it absolutely had that desired effect. Market-based pricing had the effect of increasing CPM in high ROI areas. Money flowed towards where there was ROI to be had, so it works. Now the focus on Premier Agent is on increasing lead conversions because improvements in lead conversion result in improvements in monetization through the beauty of the market-based pricing option. And so the initiatives to improve lead conversion touch on things like better customer selection; selling products like Premier Broker, where brokerages are the ones buying the impressions and then they turn around and hand those leads to their agents; investing in the Premier Agent App; investing in My Agent, which improves lead quality. So a whole host of initiatives around improving conversion. And then, secondly, in terms of where the Premier Agent business goes from here, we're investing more resources in Premier Agent Direct, which is a listings promotion product that helps listing agents feature their listings on Zillow, Trulia, Facebook and in DirectMail. And as you know, most other international real estate sites monetize their audience in this way by helping listing agents merchandise their listings, and we think there's a lot of opportunity for us to invest here. The third major initiative for Premier Agent is investing in the sales team so that we can widen our focus to sell more to smaller agents with growth potentials, smaller teams, smaller brokerages and smaller agents and setting those new accounts up for success. So all that -- taking all that together, I'm very pleased with our Premier Agent results in Q3 and our guidance for Q4, especially given the approximately $1 million headwind that we faced in Q3 and the about-$1 million headwind that we predict for Q4 due to natural disasters. In terms of seasonality, this is the first Q4 that we're entering where the whole Premier Agent business is now on this market-based pricing model. You'll recall that, a year ago in Q4, we had only a small portion, I think around 1/4 of the country, if I remember correctly, at about this time last year. So our Q4 guidance for Premier Agent reflects our best guess at this point in time of how Q4 results will come together based on our estimation of seasonality and all the other puts and takes that I just ticked off in reviewing Premier Agent.

Our next question comes from John Campbell of Stephens Inc.

Spencer, on the My Agent feature, I just want to make sure I understand that correctly. So I mean it sounds like that you guys are, I guess, attempting to improve the conversion trends. That's maybe going to take some repeat leads out of the mix, so maybe a near-term, medium-term kind of impact to rev but it sounds like the right pivot long haul. But my question is, the My Agent feature, is that going to be priced differently? Or are you thinking that the self-serve model just kind of naturally normalizes that spin in ROI, I guess, as the lead quality improves?

Yes. So there are 2 benefits to My Agent as it -- with respect to lead conversion. The first is that consumer that's working with the single agent will now see the same agent the next 30-odd times they come to the site, and therefore, they're more likely to transact from the agent they're connected with. The second benefit is once the consumer is working with a particular agent, she no longer contacts a different agent who is unable to convert that lead. So there is less -- there are fewer bad leads that can't be closed. So it's a double benefit to lead conversion. In your question, you said something like it might be a near-term hit to revenue. I don't -- we don't see it that way. It might be a near-term hit to lead volume, absolutely. We hope it is a hit to lead volume but it's a dramatic improvement in lead quality. In terms of how it will be monetized, for the time being, we've given it free to attendees to our Las Vegas Premier Agent Forum events, and we haven't yet announced how the feature will be appreciated by the rest of the Premier Agents who didn't attend the Vegas forum. So stay tuned.

So it just sounds like pricing, I guess, is TBD?

Yes.

Our next question comes from Ron Josey of JMP Securities.

This is Andrew Boone on for Ron. So for 2Q and 3Q, new bookings to existing agents was 52%, the lowest since you started to report this metric. So that's just as the new agents are joining the platform? Or is this happening because of the auction format, which makes pricing more transparent, the expansion to New York or something different? And then, within the context of Premier Broker and the team account stats that you highlighted, can you talk about how that's influencing the self-service auction? Are teams more dynamic in their allocation, which is kind of evening out ROI, or are you seeing anything else kind of within that manner?

Sure. Good questions, Andrew. So the biggest determinant of the metric around what percent of our impressions are bought by existing agents versus new agents is how much sales resources we devote towards bringing on new accounts versus, relatively speaking, focusing on up-selling existing accounts. And we -- if -- I think we probably overfocused for -- during parts of 2017 on growing our existing accounts and ensuring that they bought more impressions and, relatively speaking, under-invested in bringing on brand new accounts. And as I mentioned, that's something that we're focusing on rebalancing and making sure that we restock the advertiser base with new accounts that are growth-oriented and can become the big spenders down the road. With regards to how Premier Broker impacts market-based pricing, it is a -- incredibly additive to the market-based pricing model. Because when a Premier Broker advertiser buys $20,000, $30,000, $50,000, $100,000 of impressions in a group of ZIP Codes, that of course creates a market dynamic where the CPM increases in those ZIP Codes based on the new advertising dollars that are being brought into the marketplace by the Premier Broker. It also has the benefit of this new larger advertiser, the Premier Broker, providing the tracking and the lead follow-up to their agents, which, as I've already discussed, is so important to the revenue flow-through of the model. If more of our leads become transactions and become commissions, then people will pay more for the leads. And Premier Brokers have the tools, the training and the incentive to focus their agents on lead conversion. And we work with them closely to do that. One of the requirements of the Premier Broker program is using our technology stack including the Premier Agent App, and that allows us to help the brokers focus their agents on lead conversion. And for most of those Premier Brokers, they're using dotloop as well as part of the lead conversion tracking and transaction management also.

Our next question comes from Nat Schindler of Bank of America Merrill Lynch.

Yes, I know that you're still in ongoing discussions with the CFPB so nothing to report there, but if -- could you elaborate a little bit if there's been any change in behavior from your Premier Agents or mortgage brokers that are using the system because of this controversy?

Sure. So the nature of our platform is such that advertisers come in and out on a fairly regular basis. So -- and we don't speak about particular advertiser behavior. What I think is probably most instructive for all of you is that we still have robust participation in the co-marketing program. And that is a sign that our lender advertisers, own compliance departments, are looking at the program and they believe that it's compliant and that their broker participation in it complies with the laws that apply to them. So we haven't seen any significant impact, as I said, though that is with a caveat that advertisers do tend to come on and off our platform just in the ordinary course.

Great. And just to clarify earlier comments you have made on the call on this. You have said in the past that this is a small percentage of revenue that comes from this program. Is that how you would clarify it? That you would...

Yes. We have said it's a small percentage of overall revenue.

Our next question comes from Brad Erickson of KeyBanc Capital.

So first -- just 2. When you think about the composition of Premier Agent revenue now coming from the teams you called out earlier in the prepared remarks versus just the individual agents, is it kind of roughly even at this point? Or is one side larger a contributor than the other there?

I don't have the data at my fingertips to answer that. I'd say, certainly, among our largest spenders, it's tilted towards teams. A large portion of the Premier Agents spending between $10,000 and $100,000 a month are teams. I would venture to guess there are very few individuals agents at that spend level. In terms of total dollars though, there are still many, many, many individual agents spending many thousands of dollars. So it would -- by that metric, it would be tilted more in favor of the agents -- individual agents not teams.

Got it. That's helpful. And then just in terms of the growth rates of Premier Agent versus some of the existing cohort spending you called out, those are still really far apart. Obviously, very, very high for that existing cohort versus, I guess, the 20 -- mid-20s you reported in the quarter for overall Premier Agent revenue. Is that something we should expect to converge at some point? Maybe just help us reconcile that a bit better.

I'm trying to think about how best to answer that. I mean, clearly, we've been focused on growing high spenders' spend. We've been doing that by focusing our sales energies on those accounts, by changing our product strategy and our software tool strategy to focus on those accounts and by changing the way the pricing model works to market-based pricing to focus on those types of accounts, and those initiatives you're seeing and the numbers that we shared around growth of teams and Premier Broker. In addition, there's a broader industry trend towards teams who are gaining share from individual agents. And the reason for that is changing consumer expectations, where consumers now expect agents to be 24/7, and it's very difficult for a single agent to be 24/7. That's a lot easier for a team of 3 or 5 people to offer that kind of service. In terms of, over time, how -- might those trends converge? Sure, I suppose, over the very long run, I would expect those trends to converge, by which I mean the spend growth of teams to converge closer to the overall spend growth. But I still think that's many years away.

Our next question comes from Lloyd Walmsley of Deutsche Bank.

This is Kunal for Lloyd. A couple, if I may. One on the New York market. What are you seeing in terms of pricing and overall spending trends in the New York market? How big is it as a percentage of your Premier Agent revenue?

So when we brought Premier Agent to New York earlier this year, we were dramatically under-monetized in New York relative to the size of our leads and size of our audience. So there was a bit of a gold rush in New York where agents and teams were buying impressions at a very low price and leads at a very low price. The cost per lead and cost per impression in New York has gone up materially over the last 6 months as the auction dynamics have had their effect, but still, if you were looking at ZIP Codes throughout New York, you would find outsized ROI as compared with many other parts of the country, because it's still dramatically under-penetrated relative to the size of the market and the size of our volume and our lead flow. The -- I guess the other way I'd answer it is we have outsized consumer audience share in New York as compared with even in other major cities, and yet our revenue in New York is still under-penetrated relative to even the percent of total commissions or total real estate activity in New York as compared with the country. So we still have a long way to go to sort of earn into our fair share and then some in New York. I'm going to take 1 or 2 questions from Slido, and then we'll come back to the call, please.

Kathleen, another question about CFPB. How are real estate agents and mortgage agents responding to the uncertainty around co-marketing? And specifically, what percent of mortgage agents that buy agent co-marketing also spend in the mortgage marketplace?

I think you -- I guess, I think you answered the first portion of this. I would add that the percent of agents that have a co-marketing partner hasn't changed materially in the last year or 2. And it also hasn't changed since we announced that we are in settlement discussions with the CFPB. Do you agree with that?

That's correct. Yes, Spencer. And in terms of overlap of advertisers, there certainly is overlap of advertisers. Some lenders simply prefer to be in the mortgage marketplace and not do co-marketing. Many do both. We don't break out what the overlap percentage is there.

There's another question from Slido about Premier Agent revenue deceleration year-over-year. Even adjusting for that hurricane impact. I mean, there's obviously a law of large numbers which makes it difficult to keep growing Premier Agent revenue at the same rate. I think Q3 revenue growth for Premier Agent was very strong. That having been said, I do think that we need to grow the size of the sales team a little bit more in Q4 and going into Q1 because there is still so much opportunity and we're still so under-penetrated. Adding more salespeople is highly accretive, and we intend to do so.

Operator, we'll go back to the call and then I'll go back to Slido before we wrap up.

Our next question comes from Jason Helfstein of Oppenheimer.

This is Jason's associate. So I'm asking, the -- so the impact of the business slowdown in home sales metrics on the -- on Zillow.

So I take that question to be asking about generally sort of how is the housing market and what impact does -- do macro housing trends have on our business result. And so I'd say that housing overall is very strong, which is to say it's a seller's market, home values are appreciating more than 5% year-over-year. We are -- the market is inventory-constrained in most major cities. That puts pressure on buyer agents on lead conversion, meaning that a buyer lead is less valuable if a buyer's agent has to work with that buyer for 2, 3, 6, 10 months and write 10 or 20 offers before their offers are being accepted because inventory is so tight. So in some ways, the hot housing market is a bit of a headwind on our business because it takes longer to convert a buyer lead. And that having been said, our listing lead-generation business, Seller Boost and Premier Agent Direct, benefits from the tight listing environment, and our New Construction business benefits materially from the inventory constraints because homebuilders are anxious to market their listings on our platforms to sell those new homes. So taken as whole, this is a pretty good housing market for us, with the caveat that tight inventory makes it a little bit harder and more time-consuming to convert a buyer lead.

Our next question comes from Tom Champion from Cowen.

Engagement in terms of visits per user continues to decline. It looks like this is the fifth consecutive quarter this has been the case. Just curious if you could share your thoughts on what might be driving this very favorable trend. And then shifting gears a little bit to expenses, sales and marketing was a little bit lower than we had modeled. It looks like the decline was a little sharper sequentially than maybe it was last year, and just curious if there was anything that you'd call out on that.

So on sales and marketing, we've been focused on cost control company-wide with respect to headcount. And starting about 2 quarters ago, we were -- we have been very careful about adding headcount. And I think you're seeing that in the sales and marketing numbers. When we look at our efficiency, our sales efficiency, it is significant. We were looking at the numbers yesterday, and adding $10 million to $20 million of sales team salaries and total compensation year-over-year drove over $200 million of incremental revenue. Am I getting those numbers right, Kathleen?

Yes. Yes.

Thanks. So add $10 million to $20 million of headcount expense on sales and get $200 million plus of incremental revenue, that was a pretty good efficiency improvement in 2017. And I think that's what you're seeing on the sales and marketing numbers. Now as we go into 2018 planning, we're of course asking ourselves whether that is the right trade-off. And as you know, we manage this business by constantly evaluating trade-offs between revenue growth, market share gains, product innovation and margin expansion. And so we're going through that exercise right now for 2018. And of course, next quarter, we'll share with you where we net out on it all. In terms of visits per user, we -- across all of our brands, we have countless initiatives focused on driving increased engagement. Everything from better short order to better e-mail and push notifications to more personalized search results and similar homes to improvement in listings quality, which has been a dramatic step-change improvement over the last 2 years, where, today, listings quality on Zillow and Trulia are dramatically better than they were 2 years ago, and that improves user engagements. So it hasn't been one thing. It's been many things. And of course, there's always a bit of a headwind to that engagement metric, which is advertising dollars, which -- of which we spent more than $100 million in 2017, and ad dollars typically drive new users but typically less engaged users. And so that's a headwind on engagement metrics but a benefit on overall traffic growth.

Our next question comes from Mark May of Citi.

No problem, Mark. Feel free to follow up with us after (inaudible)

Hello?

Oh, there you are. Hey, Mark.

Sorry about that. Sorry. Just a quick question. Sorry if this has been addressed already, but your sales and marketing expense ratio, I think, was down year-on-year in the quarter. How much of that was driven by sort of efficiencies in sales force as opposed to kind of your working -- ad budget? And how should we think about the sustainability of that leverage?

Yes. So at the risk of repeating a little bit what we said, that -- the savings was due to a mix of things. I think the primary driver there was discipline around headcount that you see reflected in those numbers. There are some other things in there, just efficiency and headcount. But as Spencer just noticed -- noted, one of the things we're looking at in Q4 and into 2018 is we do think that we can make some additional strategic investment in sales, headcount that could yield additional revenue, primarily widening our focus to emphasize on small-spending agents who have the potential to grow into being larger-spending agents and setting them up for success. In terms of the mix with advertising, again, as I said, primarily the leverage that we're seeing was on the headcount line and not the advertising number.

Okay. We'll have that question from Slido. Update on Seller Boost. So we don't break out revenue from Seller Boost, as you probably know. Agents that buy Premier Agent above a certain price point get seller leads, so that's listing leads off not-for-sale homes as an added benefit, but we don't break out revenue from it. As I mentioned, seller leads are more highly valued than buyer leads, especially in an inventory-constrained market like this one. There's another question from Slido. Do you envision deeper integrations of Premier Agent with other leading agent platforms? Yes. So we are investing very heavily in rewiring the way data flows in the real estate industry through tools like Bridge. We acquired two start-ups, Bridge and Retsly, and have woven them together such that now Bridge allows a listing agent to enter a single listing and have it go to overlapping MLSs, and it allows brokerages to get output of their listings from overlapping MLSs. And this single point of entry and single point of data output solves a huge industry problem that plagues agents, brokers and MLSs, and we're committed to devoting resources to solve it. Meanwhile, our Premier Agent App connects to dozens of other CRMs so that agents have choice in terms of how to best convert leads to transactions. So through APIs on the Bridge side and on the Premier Agent side, we have many integrations with other technology companies, and it's an important part of the real estate data ecosystem, of which we're a leader.

Operator, any other questions from the call?

We do have a question from Shyam Patil of Susquehanna.

Just had a few quick ones. Spencer, you kind of talked about this a little bit. But in the past, you talked about how, when you look at revenue growth and margins, that you generally expect that number to be flat to up kind of going forward. Do you see why -- is there any reason why 2018 would be different? And then in terms of your mortgage products, one of the things you've talked about is how if there's an impact from co-marketing that you could shift lender spend to other products. Can you just talk about some of these other products and what kind of traction you've seen thus far? And then last one, Kathleen, just in terms of co-marketing, you've talked about it being a small portion of overall revenue. How would you characterize it as a percentage of Premier Agent revenue?

Great. So I'll take the 2 co-marketing-related questions first and then pass it back to Spencer. In terms of overall revenue, we haven't given any guidance in terms of the portion of co-marketing and how that measures up. We look at it on a percentage of overall revenue, in large part because of the second -- the other question you asked, which is, to some extent, this advertising spend is pretty fungible between the Zillow Mortgage Marketplace and co-marketing. And we are working with advertisers every day to figure out the optimal mix for each of them based upon their goals and whether they want to spend in a co-marketing arrangement or in the Zillow Mortgage Marketplace. So beyond that, we don't break it out because we don't think it's instructive.

On the revenue growth and margin expansion, the point I was trying to make a couple of years ago when I first brought this up was that those 2 -- there are trade-offs between those 2 metrics. And so a point of revenue growth versus a point of margin, these are 2 dials and we control them both. And of course, in the aggregate, we endeavor to raise sum total of the two. Now as you get larger and larger revenue numbers, and we're passing $1 billion of total revenue this year, it becomes difficult to continue to raise the sum of those 2 numbers because revenue growth tends to decelerate when you hit larger numbers. So I wouldn't draw anything -- this isn't a rule that's written in blood. It's an observation that these 2 numbers trade off against one another. And as we're in the midst of our 2018 planning season, we're looking at all potential investments, and we'll come out in 3 months and tell you where we netted out as we go into 2018.

There's a question from Slido about NOLs and their expiration. So I guess the question is what amount of outstanding NOLs do you expect to use? They don't expire for 20 years or so, so we hope to use them sometime in the future. I think it's uncertain. It depends how quickly net income grows, but they expire very far from now, so it's not something...

And the balance is very large, so we have a lot of time and a lot of NOLs to use up.

Are there other questions from Slido? Or no? Okay. Operator, are there any questions from the queue?

(Operator Instructions) Our next question comes from Brian Nowak of Morgan Stanley.

Okay.

Hello?

Brian, did you have a question?

(Operator Instructions)

Okay. Brian, feel free to follow up with us afterwards. I think, operator, we'll wrap up with that. Thank you for joining the call. We look forward to talking to you again when we report Q4 and full year results in February. Thanks very much, everybody.

Thank you, all.

Ladies and gentlemen, thank you for participating in today's conference. This concludes today's program. You may all disconnect. Everyone, have a great day.