Zillow’s ZG, -3.43% experiment in home-flipping has blown up in its face — and the company is blaming the “unpredictability” of home prices.
Now, some are asking about the reliability of the company’s so-called Zestimates, which provide an estimate of a home’s value based on a proprietary formula. Zillow says a Zestimate is published for around 100 million homes nationwide.
In announcing its latest quarterly earnings on Tuesday, Zillow Z, -2.12% confirmed that it will “wind down” its Zillow Offers division that focused on buying homes, refurbishing them and then selling them, hopefully for the profit. But the profit piece was missing.
“‘They had a vision where every Zestimate you would see on the website would be like a live bid.’”
During the most recent quarter, Zillow’s Homes segment, which includes Zillow Offers, recorded a nearly $422 million loss before taxes. That’s up from a roughly $76 million loss a year earlier. Reports have suggested that the company has already begun offloading some 7,000 homes, with a majority fetching a price below what Zillow paid for the properties.
“We’ve determined the unpredictability in forecasting home prices far exceeds what we anticipated and continuing to scale Zillow Offers would result in too much earnings and balance-sheet volatility,” Zillow Group co-founder and CEO Rich Barton said in the earnings release.
The company said closing its iBuying business segment will “take several quarters” and involve a 25% reduction in the company’s workforce.
In light of the challenges Zillow faced with its home-buying efforts, some people took to social media to call into question the accuracy of the Zestimate tool. “It’s really a toy,” said Mike DelPrete, a real estate analyst who tracks the iBuying sector. “It’s meant to drive people’s interest in property.”
The Zestimate, which first debuted in 2006, played a role in Zillow’s home-buying operations. “We leveraged the Zestimate in our Zillow Offers operations the same way we encourage the public to use it: as a starting point,” Zillow spokesperson Viet Shelton told MarketWatch in an email.
However, Zillow had more ambitious plans in mind for merging its Zestimate model with its iBuying division. In February, Zillow announced that the Zestimate would represent “an initial cash offer” for certain eligible homes across 20 cities, including Phoenix, Miami, Denver, Nashville, Houston and Los Angeles.
“Presenting the Zestimate as a cash offer to qualifying homes up front will save time, reduce friction and provide greater transparency — getting us closer to our vision of helping customers transact with the click of a button,” Zillow chief operating officer Jeremy Wacksman said in the announcement.
Consequently, to some, the recent turn of events suggest Zillow’s experiment failed.
“Zillow used to talk about how they had a vision where every Zestimate you would see on the website would be like a live bid, and you could just hit the bid and sell your home for that amount,” said D.A. Davidson analyst Tom White. “This move out of iBuying suggests that’s not going to happen.”
How accurate is the Zestimate?
Complaints about the Zestimate are nothing new. An unsuccessful 2017 class-action lawsuit against Zillow claimed that the company was misleading home buyers by publishing figures that were below what sellers were seeking for their homes. The judge ultimately dismissed the case, noting that the tool’s name “itself indicates that Zestimates are merely an estimate of the market value of a property.”
Zillow says that the Zestimate has a median error rate of 1.9% for homes that are on the market and 6.9% for homes that are off the market.
The accuracy varies significantly across markets. In Cincinnati, for instance, roughly 35% of Zestimates for off-market homes were within 5% of the eventual sales prices, and 82% were within 20% of the price. Comparatively, in Denver, 51% of Zestimates were within 5% of the sales price, and 94% were within 20% of the sales figure.
“According to Zillow, the Zestimate has a median error rate of 1.9% for homes that are on the market and 6.9% for homes that are off the market.”
In that sense, Zillow has likely succeeded with the Zestimate. Scrolling through the site’s listings has become something of a hobby for many, with Saturday Night Live even lampooning the habit. Finding out the value of homes in your area is likely part of the behavior’s allure.
There are certain circumstance where the valuation may be more trustworthy than others, though. For instance, with the “cookie-cutter type of homes,” Columbia University real-estate professor Tomasz Piskorski said, the model may be more effective since it can rely on “a lot of comparable transactions in the area.”
Zillow, itself, has recognized the need to improve the accuracy of its valuation model. In 2017, the company launched a contest in which participants sought to better the algorithm underpinning the Zestimate, with the winning prize going to a group of data scientists and engineers from three different countries. At the time, Zillow said the Zestimate was on average around $10,000 off of the actual sales price of a median-priced home, and that the information provided by the winning team would reduce that margin by around $1,300.
Challenges in the COVID-era housing market
The red-hot housing market in America since the summer of 2020 likely confounded Zillow’s efforts to appropriately price the homes it was purchasing.
The company’s spokesperson argued that the Zestimate was not at the crux of the pricing issues it faced, instead pointing to difficulties the company had in accurately forecasting “the future price of inventory three to six months out, in a market where there were larger and more rapid changes in home values than ever before.”
As some analysts argued, any real-time valuation will be limited in what it can tell you about the direction of prices in the future. “Valuation is inherently ephemeral,” said Michael Greene, co-founder and CEO of ResiShares, a residential real-estate investment company.
Greene used an analogy of a fast-moving train to describe the difficulty inherent to nailing down the price of a home in a competitive market like the one that’s been in place for over a year now.
“If a train is moving faster, knowing where the train is at any given point in time tells you much less about where it will be when you have to catch it,” he said.
“‘Valuation is inherently ephemeral.’”
Another factor were the structural changes that have occurred in the market — such as the sudden demand for larger homes in the suburbs — that made past data less reliable in predicting future prices, Piskorski said.
Zillow also had the disadvantage of being newer and perhaps having less reach in the markets it was operating in as a home flipper. Brokerages have better access to real-time transaction data, Jason S. Helfstein, head of internet research and managing director at Oppenheimer & Co., argued.
“Zillow wasn’t necessarily privy to that because they would see the value of a home that sold once it updated” in the multiple listing service operated by local Realtors associations, Helfstein said. There’s typically a lag in that MLS data becoming available.
Helfstein contrasted that with one of Zillow’s main iBuying competitors, Opendoor OPEN, -6.77%, which he said was “actually in a market doing transactions just like a brokerage firm.”
However, Helfstein contends that Zillow’s problems in iBuying had more to do with holding onto homes for too long and at a higher cost than expected, rather than because of deficiencies with the Zestimate.
Ultimately, Zillow’s troubles reflect how difficult it is to succeed at buying and selling homes as an investor. “IBuying is a difficult business model, and it was never expected to not be,” said Ygal Arounian, managing director of internet equity research, Wedbush Securities.