Description

ZILLOW GROUP (Z) - SHORT

Thesis: Zillow’s entire business model is reliant on third-party real estate brokers who not only don’t like them, but who own and control the listings data content without which the company couldn’t exist. This is the flaw in their model and it has been exposed by the deals they have been forced to cut with Realogy, Keller Williams and several large regional brokers that increasingly limit their saleable inventory creating a material existential risk. Furthermore, the company is under pressure from several other angles as outlined below. 

Note: Zillow recently closed its acquisition of competitor Trulia (TRLA); Trulia basically has the same business model and all comments below on how the business works, how the broker relationships are set up, etc. apply to both Zillow and Trulia. In addition, TRLA was an all-stock deal and post deal, the short interest % is now lower. There has not been an issue with borrow in the past and borrow got easier and cheaper post-closing.

Background
Basic home sale: a homeowner hires a real estate agent to sell their home - this creates a listing; that agent is an independent contractor who works for a broker who owns and runs the office (can be branded/franchised such as the local RE/MAX office or an independent). When a commission is earned by the agent, the broker gets a cut known as the split.

This agent inputs the listing data comprised of pictures and information about the house into both/either (i) their broker’s internal database and/or (ii) something called a Multiple Listing Service or MLS; there are more than 800 MLSs in the country, one for each submarket, the primary function of which is to aggregate all local listings into a database made available to all brokers in that market so that anyone potentially interested in buying the house knows it’s for sale. The key item here is that the listing itself is the legal intellectual property of the broker, not the agent. 

Zillow recognized that having 800 MLSs was inefficient, figured out how to get free access to the listings data and created a website to aggregate most homes for sale in the U.S. into one place. Zillow gets the listing data primarily in three ways, all of which are for free: (i) from a broker directly (i.e. a Realogy brokerage sends the listings data feed), (ii) from a MLS feed (note that the MLS is simply the aggregator, they do not own or control the data, the ultimate decision of who the listings are sent to is made by the broker via an opt-in/opt-out mechanism at each MLS), or (iii) from a syndicator called Listhub (they have agreements in place with most MLSs and provide a one stop shop for a party who wants access to the listings but doesn’t want to go through the process of signing up each MLS separately).

Zillow’s business model
Once the site was built, Zillow had to figure out how to make money and spent >$100 million dollars on advertising to establish their brand and drive traffic, becoming the most visited real estate site in the process. Go to these homes on Zillow (or input address into Google search box) and it will take you to a page showing what you will find on their website for any home for sale. (couldn’t post a screenshot to Value Investors Club; over time these homes will be sold or taken off the market, if so go to any listing with 3 real estate agent faces on it)

Listing 1: 2505 Worthington St, Dallas, TX 75204
Listing 2: 5403 Willow Wood Ln, Dallas, TX 75252
Listing 3: 5631 Richmond Ave, Dallas, TX 75206

There is nothing proprietary here in the listing information, it is a pass-through reproduction of the data provided by the broker; what Zillow does add to the page is the box in the upper right corner where you see the names and faces of three real estate agents. The idea is that if you the consumer are checking out this house and want to learn more, you can enter your contact information and click the box “Contact Agent”; once you do this, you become a “lead” and your information is sold to one, two or all three agents and they will likely call you very quickly. Most importantly, none of these three people are the selling agent who has the listing, these are competing agents buying that space on Zillow’s page and the clicking consumer is their next potential commission. Selling this space to agents for lead generation represents basically the entire business model, it is called the “Premier Agent” business and represents ~70% of Zillow’s current revenue. 

Here is a summary of the pitch from Zillow if you are an agent and you want to be one of those three Premier Agents that show up in that top right box on a for sale listing. The product is sold by zip code and by impression (not per lead), so your buy is based on how many times your face shows up in the box on homes for sale in your market. The Zillow pitch: for $360 per month or $4,300 per year, you will get about 180 leads per year; the typical conversion rate of a lead is only 3%, so 180 times 3% gets you an average of 5 additional closings; based on an average home price of $300,000 and your commission take of $5,200 per home, all-in you make an extra $26,000 based on your $4,300 spend; a return of ~ 6x your investment in Zillow leads.

On its face this oversimplified pitch makes sense and Zillow has built this lead generation platform into a business with about $240 million in revenue, but after expenses, huge stock comp expense and all the advertising spend, the company loses money on the bottom line. There is a large and undisclosed churn of Premier Agents, but as of the end of 2014, there were ~62,000 agents (“Agent Count” or subs) paying for leads spending an Average Revenue Per Agent of $360 per month (the company’s definition of “ARPA”). These are the two key metrics as (i) Agent Count times (ii) ARPA = total revenue. Agent count is the number of agents during a period purchasing impressions/leads and ARPA is a function of price/mix and volume (number of impressions are driven by traffic). 

Brokers Expose the Business Model Flaw
Zillow’s available inventory is defined as all of the space on current for sale listings on their site on which they can add the box and sell that space to the three agents. There are 5-6 million homes sold in the U.S. per year and about 2-2.5M active listings at any one time. This creates a finite number of leads but remember that Zillow smartly charges per impression and they can create more and more impressions with more and more advertising to drive eyeballs to the site. (Note that Zillow also has a page for most homes in the country including all of those that are not for sale; space on these homes is not sold to/paid for by agents as browsers viewing these homes are likely not in the market for a home and therefore they are not a lead that has any/very little value – third parties estimate that up to 50% of all of Zillow’s traffic is from people looking at their own home or checking out their neighbors with no intention of buying or selling).

So why do the brokers and the selling agents that work for them dislike Zillow? For the selling agent, having a listing is historically a great source of new leads for their next deal as he/she would typically be the first point of contact: anytime a call came in about the house for sale, this represented the opportunity to become that caller’s agent on their potential purchase: the agent would say something like “Sure come by and check out the house and I would also love to show you five more just like it.” The problem on Zillow is twofold: (i) go back to the listing above, the selling agent’s contact information is nowhere to be found, it’s not in the 3 agent box and you have to scroll all the way down to the bottom to even find it; (ii) what is in the box are three of the selling agent’s competitors who have paid Zillow to both show up as well as buy leads on his/her hard-earned listing. (Separately, what the consumer doesn’t know is that when they click “Contact Agent” to learn more about the home on Zillow’s site they are interested in --- the agents that have paid to have their faces in the box who call you back are not the selling/listing agent and therefore likely have never actually been to the house and can’t answer any questions you may have).

The brokers also lose out as their selling agents are losing valuable leads and potential commissions. Furthermore, the large brands/franchisors view portals like Zillow as enabling startup/unbranded independent brokers to gain share as it is less necessary to be affiliated with a brand if leads are available elsewhere. So who are the big brokers what have they done about it?

1. Realogy (RLGY) is the largest broker/franchisor in the country who owns some of the best brands in the business (Coldwell Banker, Century 21, Sotheby’s, ERA, Corcoran, Better Homes & Gardens) which collectively represent a huge 20-25% market share of all homes sold in the U.S. Zillow cannot exist without access to Realogy’s listings as the Zillow website would be missing 20%+ of all homes for sale in the country -- Zillow needs close to 100% coverage of all listings to be relevant as having all the listings in one place is their convenience/value add to the browsing consumer and if they are missing a big chunk of the listings in a market, the consumer will go elsewhere (there are many other places to go including realtor.com, any local brokerage site has complete market coverage via “IDX” sharing, numerous other portals like homes.com and homesforsale.com, parts of Facebook, more to come including www.brokerpublicportal.com; type an address into Google of a home for sale in your neighborhood and see how many sources of the same listing information come up). 

This gives Realogy all of the leverage (as the owner of the listings, they send an aggregated feed of all of their listings information to Zillow and they can pull this feed at any time) and they used this leverage to cut a deal with Zillow, Trulia and the other portals fundamentally changing the way its listings appear on these sites, check out these homes on Zillow: (again, when these homes are sold or taken off market, go to any Realogy branded listing with just one agent in the box – some like Sotheby’s listings have no box at all)

Listing 1: 10794 Edenoaks St, San Diego, CA 92131 (Coldwell Banker)
Listing 2: 15622 Jube Wright Court, San Diego, CA 92127 (Coldwell Banker)
Listing 3: 1236 Milan Ave Coral Gables, FL 33134 (Century 21)
Listing 4: 2600 Sw 64 Av Miami, FL 33155 (Century 21)

These links are to Realogy branded listings and look similar to the others above with one key difference, instead of three competing agent faces, there is only one face and that is the selling listing agent, featured in the box alone and not hidden at the bottom. Realogy cut a deal --- for all of its brands --- that blocks Zillow from selling leads in the box to its competitors, thus protecting their agents and their listings. The implications of this deal for Zillow are profound: Realogy has ~20%+ market share so overnight Zillow lost the ability to sell leads on 20% of the homes for sale in the country - this is a problem as their entire business model is based on selling leads. 

What did Zillow get in return? Realogy agents pay Zillow up to $15 per agent per month / $175 per year to protect this inventory – this is comparable to the $360 per month that Premier Agents pay for the same space, a >95% discount. It is estimated that the total amount paid by Realogy agents to Zillow per year is approximately $15 million so Zillow, a company with a $7.6 billion valuation, sold 20% of its available inventory for $15 million. There is a link at the bottom of this site for the full document that outlines Realogy’s program: http://zillowscrewsremax.com/

2. Keller Williams: this is a private company and the 2nd largest and fastest growing national broker with approaching 10%+ market share and it stands to reason that they would want a similar deal --- turns out they have a better deal. Call them up and they tell you about it or watch this video of founder Gary Keller describing his motto “My Listing, My Lead”.https://www.youtube.com/watch?v=EgulaQS6xMs

Here are some Keller Williams listings on Zillow: (again, when these homes are sold or taken off market, go to any KW branded listing)
Listing 1: 826 Lake Ave NE, Atlanta, GA 30307
Listing 2: 2925 Tall Pines Way NE, Atlanta, GA 30345
Listing 3: 1601 W Las Palmaritas Dr, Phoenix, AZ 85021
Listing 4: 14614 S 20th Pl, Phoenix, AZ 85048

You will notice that there is no box at all but rather a banner ad. Keller Williams has a similar aggregated direct feed to Zillow that they can pull at any time and thus they also get protection on all of their listings from competitive lead purchase. What does Keller Williams pay to Zillow for this privilege: ZERO. If Zillow said no, they don’t get the listings. Zillow would never do the RLGY deal for $15M let alone the Keller deal for zero but they had little choice – they need the listings content to exist and this is the only way to get it.

3. Large Regionals: in many of the top markets, there are large regional brokers who have substantial market share comparable to or better than the national brands. This is a highly competitive industry and it was only a matter of time until these large companies demanded their own deals and the following companies seem to have done so to various degrees: Howard Hanna, Weichert, Long & Foster, Rodeo Realty, Fox & Roach, etc. They all have large market share of the listings in their core markets, that is the key and their listings are largely protected. 

4. RE/MAX and Berkshire Hathaway Homeservices are the number 3 and 4 largest national brands, respectively. RE/MAX is newly public and still controlled by the Founder/CEO, and Berkshire recently consolidated Prudential’s residential brokerage business and is of course part of BRK. In many markets RE/MAX has 10%+ share and BRK a wider range with 5-15%+. Interestingly, neither of these companies seem to have their own blocking deals (yet), however RMAX has publicly stated that they are well aware of the deals that their competitors have. They also realize that not having protection on their listings puts them at a competitive disadvantage to Realogy, Keller and the large regionals as any agent that works for a brand with a protection deal can buy leads on RMAX listings, but RMAX agents cannot buy leads on any listings protected by their competitor’s brands (this factor could also lead to a broker without a blocking deal having agent recruiting/retention issues). RMAX and BRK are both large national brokers as well as public companies and if smaller regionals can cut a deal, there is no reason why they can’t and won’t. Go to minute 5:15 and watch until 6:30 in this video and see what the CEO/founder of RMAX says to the CEO of Zillow at a RMAX convention; “As long as we are partners in success we are happy, if we become competitors, you don’t get any more of our listings” www.youtube.com/watch?v=BEwuf7BcXds

Zillow’s entire business model is reliant on these brokers who not only don’t like them, but who own and control the listings data content without which the company couldn’t exist. This is the flaw in their model and it has been exposed by these deals. Today it is estimated based on third party surveys done that up to 40-50%+ of the inventory in the best markets is now blocked and unavailable for sale (RLGY + Keller + Regionals) and if/when RMAX and BRK catch up that number can go up by another 10-20%. We will get into the sell side numbers and expectations below but nobody’s numbers account for 50%+ of the Total Addressable Market (TAM) being unavailable for sale. 

Taking it a step further, it is unlikely that the many of those 62,000 Premier Agents buying leads/impressions from Zillow realize that up to half of their markets are protected from lead purchase (and that half includes listings from some of the best brands, which tend to have the most desirable, higher priced homes which would produce the most valuable leads). Zillow’s advertising is driving a lot of traffic and impressions but these impressions can’t be served on protected listings --- they are forced to be repeatedly served only on unprotected listings. This mathematically has to materially affect lead quality, conversion rates and ROI for Premier Agents – ROI is the key to the Zillow sales pitch and its degradation should lead to pressure on ARPA and increased churn.

Evidence of all of these deals is researchable and verifiable right on Zillow’s website. There have been studies done estimating how much of the market is protected and it is pretty easy and not that time consuming to sample zip codes, see which brands are and aren’t blocked, and count how many listings come up without those three agents in the box or with no box at all. 

News Corp (NWSA) / Realtor.com
Realtor.com is another national aggregated listings website that is a close #3 behind Trulia and growing in terms of traffic count with 37 million monthly unique visitors (uniques). Unlike Zillow, Realtor.com has played nice with the brokers and the industry, and while part of their business is selling leads, the key difference is that competing agent faces have never been allowed to show up on seller listings. In addition, Realtor.com has strong ties to the large trade group NAR, and most importantly, has a key advantage as they enjoy nearly 100% accurate and real-time listings data from their industry partners via direct feed. 

Realtor.com was owned by a public company called MOVE and there are many reasons why historically the platform has not been optimized, including subscale marketing resources in the advertising battle with Z+TRLA. However, this has all changed now that MOVE was recently acquired by NWSA. https://newscorpcom.files.wordpress.com/2014/10/move-acquisition.pdf

There are several implications of this deal. First, realtor.com for the first time has access to not only competitive marketing dollar resources given NWSA size (close to $1B of EBITDA and $2B+ of idle cash) but also has access to the media expertise and cross-promotion exposure that NWSA has committed to drive traffic and awareness to realtor.com (NWSA’s platform includes the Wall Street Journal, WSJ.com, Marketwatch, Barron’s, NY Post, News America Marketing, etc. which in aggregate have >700 million monthly page views). On the conference call describing the acquisition, News Corp management outlined the strategy to “turbocharge” the growth at realtor.com via cross marketing of their media platforms. 

Second, as mentioned above, one of the sources of the listing information for Zillow is a syndication company called Listhub; Zillow and Trulia use this syndicator to get the listing information when they don’t have feeds from either the broker or the local MLS. It is estimated that between 20-40% of all listings on Zillow and Trulia come from Listhub. The problem is that Listhub is owned by MOVE (now by NWSA), Zillow’s listings feed deal expires in April of 2015 and NWSA just terminated the feed to Trulia based on the change of control (the deal was set to expire in 2016, Zillow has sued to keep the feed disputing the change of control and lost that battle). The upshot is that there will be a period of time starting shortly when Z+TRLA’s sites will be missing 20-40% of all listings in the country until the companies can backfill the data by going to each broker and MLS one by one and cutting bespoke deals (the leverage again here is with the brokers and many have gotten preferred “top slot” Listing/Seller Agent placement on their listings in return which decreases the value to the other 3 paying agents in the box that show up below them – now in many instances four agents show up in the box rather than 3 and the Listing/Selling agent is always on top). During this time period, Z+TRLA are vulnerable as Realtor.com will have full listings coverage via their direct feeds and the marketing heft to tell the world about it. 

Lastly, having Realtor.com in stronger and friendlier permanent hands at NWSA has created a completely credible alternative to Z+TRLA and gives the brokers more leverage now when negotiating with Z+TRLA in the future. Realtor.com is and will remain industry friendly and is highly focused on staying this way; this was evident when Rupert Murdoch himself came to a recent real estate industry conference to communicate this to the nation’s top brokers and agents. The CEOs of Z+TRLA addressed the same crowd and it was telling when their presentations barely got a clap vs. at the end of Murdoch’s speech, he received a standing ovation. This plus his whole presentation can be seen here: www.inman.com/2015/02/26/what-the-hell-does-zillow-mean-watch-the-video/

More Issues / Flags
1. Agent / Total Addressable Market (TAM) limitations: (i) while there are more than 1 million agent members of NAR, only 300,000 or so have real estate as their full time job and Z+TRLA already capture >1/3, furthermore, there are only 5-6M homes sold per year, a finite number of deals; (ii) internet leads have long harvest times plus low conversion rates and the ROI are controversial and unproven stats that the company won’t disclose, (iii) most agents can’t afford to pay $360 per month/$4,300 per year nor do they have the bandwidth to close 5 more homes, agents are independent contractors and a good agent who closes 12 homes per year grosses ~$55K from which they have to pay all expenses including marketing (the average agent spends 10% of their gross on marketing), (iv) there are Zillow “superagents” who spend a ton per month and have figured out how to monetize internet leads better (CRMs, phone banks, teams, etc.) and Zillow is targeting selling them more and more impressions, however there can really only be one superagent in a market and none of this matters if there is limited impression inventory to buy thanks to the broker deals, (v) 2/3 of all people find their agents via referral which makes sense given it is the largest purchase most people ever make, this relegates the remaining 1/3 to other sources of which internet is only one, (vi) part of the bull case is that at the real estate industry as a whole (brands, franchisors, brokers, agents, etc.) spend $12-$14B on marketing per year across media and a big chunk of this will continue to go digital -- the industry stopped using newspaper and print a long time ago and this report suggests the shift has already occurred: www.borrellassociates.com/industry-papers/papers

2. Potential RESPA (Real Estate Settlement Procedures Act) violations: Zillow has a program called PALs which is a co-marketing program that allows preferred mortgage providers to show up on an agent’s listings/page on Zillow and in return the mortgage provider contributes to the agent’s monthly spend (i.e. towards the $360 ARPA); the PALs program started 18 months ago and has ramped to an estimated 1/4 of all Premier Agents having a mortgage contributor (who can pay/contribute up to 50-90% of an agent’s spend) --- this has therefore been a material driver of ARPA itself. The issue is that referral payments for real estate leads like this are potentially illegal and the PALs program may be in violation of RESPA rules; TRLA included this as a new risk factor in their 10K, Z did not; if the PALs program is halted it could be a large hit to ARPA; here is an article discussing a recent whistleblower: www.inman.com/2014/12/01/lenders-said-to-be-bending-rules-of-zillows-co-marketing-program/

3. Insider selling: senior management and founders are very large sellers of their own stock; this is a very promotional management team who is on the road constantly discussing the huge upside in the stock, the TAM and the transformational TRLA acquisition --- all while consistently selling stock (the CEO owns little/no stock, has lots of options and is consistently selling under 10b5-1)

4. Project Upstream: last but not least as this may be the most important item and the broker’s nuclear option. Upstream is a new co-op that was created by the top national and regional brokers that represent a reported 80%+ of all listings in the U.S. The concept is that the listings data is messy, unstandardized, and disaggregated amongst the legacy MLS systems and the brokers and therefore it makes sense to have one industry utility where all listings would be inputted directly into one system under a common and shareable format – this system is Upstream. If/once this system is put in place, all listings distribution would come from Upstream as the sole source and will be syndicated out from Upstream to both (i) all of the MLSs as well as to (ii) any portals like Zillow who want access to the feed. The rub is that the listings will now all be in aggregated form and the dissemination of the content will be centrally controlled by the individual brokers who again legally own the listings. It is plausible that Upstream will thus give all brokers involved the leverage over Z, TRLA, Realtor.com and any others who want/need the listings data --- and in return for licensing the data, the terms of usage could conceivably state that all listings are to be protected from competing agent lead sale. The CEO of RMAX has publicly discussed Upstream, Inman news has written extensively on it (see below) and Gary Keller presented the merits of Upstream to all Keller Williams brokers and agents at their recent annual convention. www.inman.com/2015/02/09/gary-keller-delivers-150-minute-99-slide-tome-to-10000-kw-agents-in-orlando/ The Upstream company is formed, funded and currently in search for a CEO. Per Inman news on 2/16: “The Upstream project, which promises big franchises and brokers more control over their data, will get more momentum. This could ultimately lead to no ads on listings, which would hurt both Zillow and Move, as some version of this ad unit represents their most lucrative revenue source.” www.inman.com/2015/02/16/what-to-expect-now-that-zillow-and-trulia-are-one/;www.inman.com/2015/03/02/mysterious-upstream-initiative-wont-be-consumer-facing/

Some Bull Case Arguments and the Rebuttals
1. Zillow will renegotiate the broker deals now that they own Trulia and are bigger and the clear traffic leader. Rebuttal: I believe that the TRLA deal was a defensive move because of these broker deals; Zillow for years was winning the advertising, traffic, agent count and revenue war; then out of nowhere, they turn around and pay $2.5 billion for TRLA giving away 1/3 of the company in a stock deal. Owning TRLA doesn’t change the fact that the broker owns the listing and Zillow already lost that negotiation with RLGY and Keller plus these deals backfired on Zillow once the rest of the industry found out about them. It seems highly unlikely that RLGY would agree to pay more and Keller would agree to pay ANYTHING to keep their listings protected on Zillow. If they both pulled, Zillow would be missing 30%+ of all listings and the rest of the large brokers could be emboldened to follow and pull too. Realtor.com is now a credible alternative with 37 million uniques, deep pockets and close to 100% listings coverage; over time the consumer will find the site with the best listings coverage to do their searches, not the one with the most commercials and advertising. With 50-80% listings coverage, Zillow could be viewed as useless to a consumer. 

2. REA and Rightmove are the best public comps to Z in Australia and the UK, they are big companies with big margins and this is the playbook for Zillow. Rebuttal: This is a longer topic but the key here is that those markets are completely structurally different than the US market and they are not relevant comparables (these markets have no MLS/IDX system, agents control the listings, not the brokers (who own them in the U.S.), listings in the U.S. are freely available to any site that wants them unlike abroad, etc.). 

3. The Listhub problem is not a big deal because Zillow is going direct and getting MLS feeds and that is how they will get their listings data. Rebuttal: the MLSs are nothing more than aggregators of the listings data that they get from the brokers/agent input; the MLSs do not own the listings and simply provide a pipe for the aggregated listings feeds. The access to that pipe is controlled by the owner of the listing (the broker) who can choose to opt in or out and who to share those listings with. Furthermore, the big brokers (as explained above) send the direct feeds they control to Zillow anyway trumping the duplicative MLS feed. Therefore, all of these direct deals with the MLS that Zillow is touting are borderline meaningless. Lastly, if Upstream happens, all of the MLSs will have to rely on Upstream for the broker’s listings feed. 

4. 100% of RLGY and Keller listings are not protected, why is that? The reason is that there are a small subset of listings from each system that get to Zillow not directly from the parent broker feed (this self corrects with Upstream) but rather through Listhub (this self corrects once the deal expires in April) or a smaller MLS (also self corrects with Upstream). The reason this occurs at Realogy is that there are a few franchises that don’t pay to participate in the protection program and the reason this occurs at Keller Williams is that there are a few franchises that that don’t participate in the direct feed. 

5. Zillow has other businesses that they can grow. Rebuttal: the Premier Agent lead business is ~70% of revenue, is highly challenged and could go away if trends continue. The company has stumbled in the mortgage business which is subscale and hasn’t grown and the rental business has no scale and is a structurally different market with major competitors who are well ahead of Zillow. This leaves Display advertising and I don’t think people are paying $7.6 billion and huge multiples for Zillow to display Verizon ads. 

Numbers
Zillow standalone finished 2014 with $325 million of revenue comprised of $239M Real Estate + $28M Mortgage + $59M Display. The Real Estate segment is the Premier Agent lead business representing 73% of revenues. Notably, the Realogy deal (per Zillow) is not included in this segment but rather for some reason in Display (i.e. the $15M mentioned above paid by Realogy agents to Z annually is in the $59M line item – maybe it is left in this segment so it doesn’t skew down the all-important ARPA stat?). Taking out COGS of $29M, sales and marketing $168M, tech & development $86M and $66M G&A, EBIT was negative ($23M). To get to the company’s “Adjusted EBITDA” of $50M, you have to add back $36M of D&A, $3M of other and a whopping $34M of stock comp expense. At year end, the three most important metrics were: (i) Premier Agent count was 62,305, (ii) 4Q ARPA was $359 and (iii) 4Q average monthly unique users were 76.7 million. 

With respect to 4Q earnings, much like the 3Q, the company missed on both agent count as well as EBITDA with higher ARPA providing a partial offset. Starting with agent count: if you go back a year ago to the beginning of 2014, the story was all about agent count growth and the Street numbers had them ending with >70,000 Premier Agents (vs. the 62,305 they actually ended with). Agent count growth has slowed materially and the graph can be seen in any sell side report. Once this sub growth start slowing around the 3Q call, management’s story abruptly changed and they said Z was no longer as focused on agent count but rather on selling more impressions to their best agents to drive ARPA (the 4Q agent adds were the smallest since 4Q11 and on the most recent call after the 4Q14 miss, ongoing agent count growth was guided to be “modest”). 

Like agent count growth, monthly uniques ramped in 2014 as they company spent ~$80 million on advertising, but the growth rates have also since slowed. Uniques are arguably the most important metric as they show consumer interest and also of course drive more impressions to sell. This is also a number used by the Z salesforce when they sell their services to Premier Agents. However, there were two recent strange releases put out about this metric. First on January 7th, they put out their December number with a footnote that stated: “For December 2014, the reported monthly unique user metric was estimated by Zillow based on historical trends by calculating the percentage change in monthly unique users from November 2013 to December 2013 and multiplying that percentage change by the reported November 2014 monthly unique users. Zillow transitioned to an upgraded version of the Google Analytics measurement service, Universal Analytics, in the month of December 2014 on both its mobile application and website platforms. As a result, Zillow is not able to provide an accurate count of the monthly unique users as reported by the service for December 2014.” 

Then, after the close on Friday February 7th, an 8K was released stating without further explanation: “Beginning in March 2015, Zillow will no longer release its monthly unique user metrics on the Zillow IR Page. Zillow Group does not expect to report monthly unique user metrics on its Investor Relations page.” So inexplicably, they basically made up the traffic/uniques number in December, somehow gave a reportable number in January and then decided in February to no longer release their most important metric anymore? Why can’t an internet company track their uniques and why wouldn’t they release a number that has been very strong unless that number was either wrong or it’s going the other way?

Lastly, that brings us to ARPA which is again is a function of price/mix and volume (number of impressions delivered). ARPA has been growing nicely, has made up for slower agent count in the reported numbers. Deconstructing ARPA, the company has taken some price and we also know that mix has been positively affected by the PALs program described above. The other part of the mix are the effect of the company selling more impressions to its best superagents, who according to the company, some spend $5-$10K per month for leads for their teams of agents. 

On the volume side of the equation, uniques are up a lot year-over-year and therefore impressions delivered (volume) has likely been the main driver of ARPA growth. However, uniques growth is decelerating and even if the company’s reported ~77 million monthly uniques is accurate, there are only so many adults in the country and a smaller subset looking at homes and this number has to plateau at some level. 

So the mix and volume tailwinds of ARPA are slowing which leaves the company vulnerable to any deceleration in traffic/servable impressions and with price as its last lever. Price can only be driven up if conversion rates and the ROI improves and with less and less inventory available and more and more impressions being forced into the remaining unprotected listings (thanks again to the broker deals), price power from here seems like a tough bet; those superagents may not spend $5-10K per month anymore once they realize they are only buying leads on a subset of their market and that subset is getting smaller.

Briefly on TRLA, their numbers have been softer for longer and the story has been similar with decelerating agent count and traffic and weak 4Q print. Interestingly, after going on an advertising binge all year, in the 4Q, TRLA decided to spend almost nothing which led to a low quality headline EBITDA beat. In 2014, TRLA had $252M of revenue and $30M of Adjusted EBITDA; after taking out $45M of stock comp, EBITDA was negative ($15M). Also notable is that a sizable chunk of TRLA’s revenue comes from a low margin CRM tool that they overpaid for called Market Leader, which Z management has alluded to as being non-core.

Valuation
Putting the two companies together, there are ~70 million fully diluted shares outstanding per the company; at $115 per share and with ~$450M of net cash, the total enterprise value (TEV) is ~$7.6 billion. In 2014 the combined pro forma revenue was $578M, Adjusted EBITDA was $80M and after deducting stock comp, EBITDA was ~zero. 

The bull case on TRLA synergies is (i) $100M of cost cuts plus (ii) $50M of revenue synergies by 2016. On the cost side, they already let 350 TRLA employees go; that is worth ~$50M and it is unclear where the other $50M comes from as they are increasing advertising spend this year on a combined basis and for the time being, keeping both brands/sites alive. On the revenue side, the synergies are to be sourced from moving TRLA from a “share of voice” model to impression-based and it is likely true that TRLA was undermonetized in this respect; however, this assumes they can keep TRLA’s agent count and sales force in place during an integration and full business model transition. 

The Street is of course giving them full credit for these numbers plus much more and 2015 and 2016 consensus revenues are expected to grow to $750M and $1,010M, respectively and Adjusted EBITDA to $150M and $300M, respectively (both numbers before $70-$75M of stock comp expense). This implies that revenues close to double and Adjusted EBITDA is up almost 4x and with 30%+ margins, despite all of the fundamental business risks set out herein + a slowdown in agent adds, traffic/uniques and some structural challenges to ARPA. On these Street numbers, on TEV/Revenue, the stock trades at 13x, 10x and 8x respectively for 2014, 2015, 2016 and on TEV/EBITDA it is 95x, 51x and 25x, respectively. Deducting stock comp, the 2015 and 2016 EBITDA multiples go to 95x and 34x. Metrics like EPS and Free Cash Flow don’t seem to matter. 

In order to justify the average price target of $130-$150, the sell side is placing an arbitrary 30-35x EBITDA multiple on their 2016 number of ~$300M. Furthermore, many of the analysts actually project out to 2020 and beyond, assume Z achieves a huge share of the supposed $12 billion TAM, a 200,000-300,000 agent count on the platform and >$1 billion of EBITDA. Even in these crazy growth scenarios plus placing double digit terminal multiples on >$1B of EBITDA and discounting it back at only 9-11%, the present value of their DCFs only imply a stock again in the $130-$150 range or up only 15-30% from today’s price. (these same models fail to take into account that 50% of the inventory needed to hit these numbers is unavailable for sale, again thanks to the broker deals).

Conclusion and Catalysts
This is a momentum internet stock trading at a huge valuation, with a flawed business model, material structural risk to ~70% of its revenue, an inventory shortage, a problem with its industry partners who are the suppliers of their content, and a new aggressive and well-funded competitor in NWSA. I think part of the reason the opportunity exists is that it is very complicated, the stock has worked to date and many on the internet side don’t appreciate the details of the real estate side. Catalysts include: (i) REMAX and/or BRK doing a protection deal, (ii) one or more of the large brokers pull listings, (iii) Upstream launches, (iv) RESPA violation materializes, (v) traffic slows, (vi) Realtor.com gains share due to their marketing and other advantages during and after the Listhub backfill period, (vii) agent adds stall, (viii) ARPU slows. 

In the Street’s bull case multiples and DCFs stated above, downside to the short would be $130-$150 or 15-30%. It is unlikely that Z would get bought by another company given the size, valuation, business model flaw/risk and the fact that they own little/nothing proprietary and rely on others for their traffic-driving content. 

Conversely, in the base case situation we have today where up to half of the inventory in many markets is not for sale, the effect this has on TAM, price power, agent attrition and ultimately earnings power materializes and the Street’s blue sky projections fail. Even assuming the business model continues and they achieve real synergies, at $200M of EBITDA at 20x, the stock price would be in the low $60s, down 40-50%.

In the real bear case where inventory protection goes from the current 40-50% to 60%+ with a RMAX and/or BRK deal or to 80%+ with Upstream --- the entire Premier Agent/lead generation current business model faces existential risk. If this happens, it is unclear how the company ever makes money and where the stock would find a bottom.

Catalyst

Catalysts include: (i) REMAX and/or BRK doing a protection deal, (ii) one or more of the large brokers pull listings, (iii) Upstream launches, (iv) RESPA violation materializes, (v) traffic slows, (vi) Realtor.com gains share due to their marketing and other advantages during and after the Listhub backfill period, (vii) agent adds stall, (viii) ARPU slows.