Good morning,
Satya Nadella’s keynote at Microsoft Ignite yesterday was very much in-line with what I wrote about in Metaverses earlier this year; Microsoft wants to enable a whole bunch of them. I had originally planned on covering the keynote today, until I woke up to a very big surprise.
On to the update:
Zillow Shuts Down Offers
From the Wall Street Journal:
Real-estate firm Zillow Group Inc. is exiting from the home-flipping business, saying Tuesday that its algorithmic+ model to buy and sell homes rapidly doesn’t work as planned.
The firm’s termination of its tech-enabled home-flipping business, known as “iBuying,” follows Zillow’s announcement about two weeks ago that it was halting all new home purchases for the rest of the year. At the time, Zillow pointed to labor and supply shortages for its inability to renovate and flip houses fast enough.
In a statement Tuesday, Chief Executive Rich Barton said Zillow had failed to predict the pace of home-price appreciation accurately, marking an end to a venture the company once said could generate $20 billion a year. Instead, the company said it now plans to cut 25% of its workforce.
Despite Zillow’s home-buying freeze, and the news it was trying to offload its entire portfolio in one shot, this was still very surprising. To CEO Rich Barton’s credit though, I thought the company’s earnings call explained his thinking exceptionally well; I’m going to walk through Barton’s statements, but it’s worth a read in its entirety.
Where Zillow Went Wrong
The first and biggest problem was Zillow’s inability to accurately price homes. Looking back, the problem actually started to become apparent earlier this year, not because Zillow was paying too much, but because it was paying too little. CFO Allen Parker said on the Q1 2021 earnings call (emphasis mine):
Growth in Zillow Offers continued to reaccelerate in Q1. We reported home segment revenue of 704 million, which exceeded the high-end of our outlook with 1,965 home sales. Resale velocity was above our expectations. In Q1, we sold 128% of the beginning inventory of 1,531 homes was contributed to inventory decline at the end of Q1 to 1,422 homes. Purchases increased to 1,856 homes in the quarter from 1,789 homes purchased in Q4, but not quite at the pace we planned as we continued to work on refining our models to catch up with the rapid acceleration in home price appreciation.
So Zillow was paying too little; this is what Barton said on the Q2 earnings call:
As we discussed last quarter, we recognized that our Zillow Offers unit economics are above our plus or minus 200 basis point guardrails as we build and scale the business, a trend that continued in Q2 despite our efforts. We expect these unit economic trends to normalize over time as we iterate, continue to learn and as home price appreciation inevitably slows. We’ve been testing pricing elasticity in this hot housing market, and we saw rapid conversion gain throughout the quarter as we improved our offer strength. We believe these tests will serve as well across future market conditions as we strive to be a market maker across housing cycles.
In this case Zillow tweaked its model to be more aggressive, and it bought commensurately more homes. And then came last quarter, when the home price appreciation (HPA) slowdown Barton referred to arrived in Phoenix; Parker explained what happened on yesterday’s Q3 2021 earnings call:
As we came into the first half of the year and then into the second half and we made the decision to go ahead and grow, given our Q1 performance, we made a decision that we needed to increase our acquisition pace, and our forecasting process at that time was not keeping up with that rapid rate of HPA increase that was occurring. And as we went into the second half of the year, HPA still continues to increase and we believe that there’s still a strong market but the rate of increase declined significantly. So there was this volatility that Rich talked about that was way outside of the normal distribution you would expect. And that’s where I think our pricing forecasting accuracy was less than we had expected…we ended up with that volatility agreeing to prices that are going to be higher than what we expect to sell those houses for less selling costs.
The fatal flaw is clear in retrospect: if Zillow’s model was underpricing homes when the market was on the way up, it makes perfect sense that the same model was likely to overprice homes when the market was, if not on the way down, at least slowing dramatically; the fact that Zillow actually tweaked the model, not to make it right, but rather to encourage more aggressiveness on the upside, made the problem worse.
What Barton realized is that this made the entire business unworkable; from his prepared remarks:
When we decided to take a big swing on Zillow Offers 3.5 years ago, our aim was to become a market maker, not a market risk taker. And this was underpinned by the need to forecast the price of homes accurately three to six months into the future. We used historical data and countless simulations to test this belief. We set unit economic targets that required us to stay within plus or minus 200 basis points in breakeven, holding ourselves accountable to these levels publicly with you all.
Yet in our short tenure operating Zillow Offers, we’ve experienced a series of extraordinary events: a global pandemic, a temporary freezing of the housing market and then a supply/demand imbalance that led to a rise in home prices at a rate that was without precedent. We have been unable to accurately forecast future home prices at different times in both directions by much more than we modeled as possible, with Zillow Offers unit economics on a quarterly basis swinging from plus 576 basis points in Q2 to an expected minus 500 to minus 700 basis points in Q4.
Put simply, our observed error rate has been far more volatile than we ever expected possible and makes us look far more like a leveraged housing trader than the market maker we set out to be. We could blame this outsized volatility on exogenous black swan events, tweak our models based on what we’ve learned and press on. But based on our experience to-date, it would be naïve to assume unpredictable price forecasting and disruption events will not happen in the future.
Zillow Offers is certainly Barton’s baby — he made that clear in a Stratechery interview — but credit him for clear-eyed thinking on this point. Sure, Zillow could press forward, but if they don’t trust their model that basically turns the business into a role of the dice; that’s a particularly big problem given that Zillow is not only a public company, but a company that has a thriving Aggregator business in the Zillow Marketplace. For Barton that made the risk untenable; he said in the Q&A:
We’ve got these new assumptions that we’d be naïve not to assume will happen again in the future, we pump them into the model and the model cranks out a business that has a high likelihood, at some point, of putting the whole company at risk, not just the business, but in the more normal case, just causes a ton of volatility in earnings, which is not a great look for a public company. That’s basically what it boils down to.
Barton also acknowledged operational problems:
We have also experienced significant capacity and demand planning challenges, exacerbated by an admittedly difficult labor and supply chain environment. The combination of these factors has caused a meaningful backup in our processing of homes in the Zillow pipeline, which we announced two weeks ago. We judged future significant volume volatility to be a tough impediment to ramp a scaled operation, and any interruptions in the supply chain like we recently experienced will result in increased holding times, further increasing our exposure to volatility and lowering our return on equity.
I discussed Opendoor’s likely operational advantages two weeks ago; unlike Zillow, Opendoor was built from day one to turn over houses, and while the company doesn’t release earnings until next week, its public posture is that it remains open for business.
What is also worth noting is that Opendoor has two more advantages that come from being a startup laser-focused on home-buying: first, while Zillow started with a Zestimate tool that was about attracting customers to the top of the real estate funnel, and thus only ever had to be directionally correct, Opendoor knew from day one its entire fate as a business rested on its model’s accuracy; it’s definitely plausible to imagine its accuracy being much better as a result.
Second, Opendoor has a dramatically larger appetite for risk than Zillow does. Yes, it is a public company now, but it is a public company whose entire business is home-buying; investors know what they are getting into. And, by extension, Opendoor has no choice but to make the model work: they don’t have a profitable Marketplace business to fall back on.
Ultimately, though, the biggest problem with Zillow Offers is that it was always strategically flawed, and I for one should have seen it at the time.
Reflecting on Zillow, Aggregation, and Integration
When I saw the news I started mentally preparing myself for a bout of self-flagellation about the 2018 Article I wrote about Zillow Offers, Zillow, Aggregation, and Integration. Looking back, though, it holds up a bit better than I expected. The big thing I got right was Zillow’s first problem: risk and the problems that posed for the core business.
This is why last week’s news was such a surprise, to me anyways; granted, Zillow had been experimenting with facilitating sales to investors, but to fundamentally change your capital structure, margin profile, and compete with your customers in one fell swoop feels like something else entirely — and Wall Street agreed!…
Make no mistake, the business model is risky, but that is another way of saying the potential return is massive as well: truly becoming a market maker for an industry that does $900 billion worth of transactions every year has massive upside. And, by extension, massive downside for the status quo — which again, includes Zillow. That is one reason to act.
Even so, that might not have been enough for Zillow to make such a shift: remember, this is a public company accountable to shareholders, and sometimes doubling down on the core business is the most prudent course of action.
Still, the Article was ultimately a miss in two important regards. First, I didn’t really talk about the challenges in building an entirely new kind of company, and how a startup would have a big advantage relatively speaking. That’s a pretty obvious issue in retrospect, and one I should have identified up front.
The second one is more galling to me, because it should have been directly in my wheelhouse; that is the danger of Aggregators trying to vertically integrate. I wrote in a Daily Update about Google’s mistakes with Android:
For a few years there Android was the tail wagging the dog to Google’s significant detriment, particularly when it came to serving iOS users. Long-time readers know this is a persistent theme for me: horizontal service companies ought to be focused on serving the widest possible user base, and any hardware or platform efforts on their part should be in service of that goal (as opposed to vertical device companies whose services ought to be focused on differentiating their hardware and platform).
The challenge is that hardware-centric platforms tends to have “gravity” when it comes to company culture: the tangibility of a device is a sort of totem for everyone in the company, the division that made the platform expects everyone else to help sell it, and everyone else naturally wants to pitch in. We saw this for the longest time at Microsoft and the fact for years its applications and services worked better (or only) on Windows or Windows Mobile, but we also saw a similar dynamic emerge with Google and Android where Google’s services were purposely better on Android, ultimately costing Google dearly (when they lost Maps on the iPhone).
The great success of both Satya Nadella at Microsoft and Sundar Pichai at Google has been re-aligning the focus and priorities of their respective companies with their respective business models and strategic opportunities, and it makes me optimistic about the very long run.
Zillow Offers was the exact same category of mistake: the company sacrificed its horizontal position in the pursuit of vertical integration, ultimately making itself into a worse Aggregator. Barton admitted as such:
A final factor in this wind-down decision is that to-date, we have been able to serve only a limited number of customers. We’ve been able to convert only about 10% of the serious sellers who ask for a Zillow Offer, and we have tended to disappoint the roughly 90% who didn’t sell through us. Given our hard-earned position at the top of the seller funnel with 220 million-plus average monthly unique users and the popularity of the Zestimate, there are better, broader, less risky, more brand-aligned ways of enabling all of our customers who want to move.
Zillow the Aggregator encountered two problems with Offers: first, because the company felt compelled to push Offers, it was actually leaving most potential sellers with a bad taste in their mouth; this is a big problem given that an Aggregator’s advantage is the fact the end users like it and go there first. Second, the return, when properly priced for risk, of serving the fraction of customers that Zillow extended an offer to, didn’t outweigh whatever risk-free commissions Zillow might be able to earn by directing customers to other buying channels.
Again, this is absolutely something I should have foreseen, because it fits the model perfectly: an Aggregator is a horizontal business that ought to make investments to serve all of its customers; vertically integrating screws up those incentives, and that is a risk in its own right. I can see why I didn’t emphasize this point — the housing market is so large, and Zillow’s share was not only small but also served to preserve the warped U.S. real estate agent-driven model — but new opportunities entail opportunity costs, and I should have emphasized those more.
This does, of course, reduce Zillow’s long-term ceiling as a company (and increases Opendoor’s, which I have always been bullish about, along with OfferPad, another competitor in the space); Barton, though, isn’t just a founder, but also a fiduciary, and in that light this decision does make sense. Barton explained in the Q&A how he changed his mind:
Extraordinary volatility in our earnings and this big markdown on a relatively low volume of units, coupled with us running into internal and external kind of capacity and labor constraints basically brought into high relief and forced us to recognize the inherent risk that exists in the iBuying operation.
It’s almost as simple as that. I mean, we could solve our forecasting problems in most markets, for sure, and we can solve our operations problems. But what we can’t solve is what the model is going to tell us about how much capital we need to raise, deploy and risk in the future in order to achieve a scale that we think is necessary to offer a fair price to customers for their homes in a competitive way. We have been offering a fair price now while we’ve learned the business, but in order to get the scale economies, we have to get a lot bigger. So the score has changed a lot this quarter, but if we’re 10 times, 20 times the size we are now, it’s just doesn’t compute.
And look, that’s what happened. And luckily, our calculus is different. We built Zillow Offers on top of an incredibly strong core business that has grown tremendously while we’ve been taking a big swing on Zillow Offers. And it’s that strong core business that gives us such an enviable position in the upstream in the customer funnel. That’s what happened. It’s not actually all that complicated logically. It is complicated emotionally because of the people ramifications of our decision to wind it down. So that’s where it gets you.
I find this really admirable (and, to be clear, a restatement of the advantages Opendoor has by virtue of not having a business to protect). Again, Zillow Offers was Barton’s responsibility, so he deserves blame, but at the same time, that makes the fact he looked at the business logically instead of emotionally that much more impressive. Of course the best time to avoid making a strategic mistake is at the beginning; the next best time — and the far more difficult time, given sunk costs — is the moment you realize the decision was wrong.
This Update will be available as a podcast later today. To receive it in your podcast player, visit Stratechery.
The Stratechery Update is intended for a single recipient, but occasional forwarding is totally fine! If you would like to order multiple subscriptions for your team with a group discount (minimum 5), please contact me directly.
Thanks for being a subscriber, and have a great day!