There's More to Buying and Selling Than the Click of a Button on Spreadsheets: Zillow

The real estate industry news has lately had quite a few stories about Zillow's "pause" in buying homes.  As the days went by, the "pause" was just another word for Zillow's decision to stop buying homes--why? because it made too many bad decisions about pricing, the market, and a third reason, what I consider to be its faulty reliance on its own automated valuation tool.  Zillow only recently became a real estate brokerage (in 2021) after many years of being a third party data aggregator, showing listings online in the way many companies did by obtaining permission from the various MLS's across the country, and by then offering advertising of these listings to the Realtors who listed them in the first place (for a price of course)  so that prospective buyers and sellers would have another portal by which to find listings and possibly a future agent.  You've probably heard of "Zestimates", their early automated valuation tool which was typically inaccurate, for various reasons.  One of the major reasons was that, as a non-broker, Zillow was not using the full information gained by subscribers to the Realtor MLS, but drawing from the "bare bones" information of the public property tax records.  Value of a property can be in the subjective eyes of the beholder, aside from the schooled requirements and standards used in the professional appraisal industry, and this intuitive factor was quite elusive to Zillow's valuation tools.  

And it continued along this same path after becoming a licensed brokerage (something it used to claim would never happen) and Zillow proceeded to buy, buy, and buy more properties with the intention of "buy, fix and sell".  Until in recent weeks it hit the "pause" button, and now its leadership says:

“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,” CEO Rich Barton said in a statement.

Barton said that Zillow will instead focus on “creating an integrated and digital real estate transaction that solves the pain points of buyers and sellers while serving a wider audience.” (Quote from Realtor Magazine, Nov. 3, 2021, their sources from “Zillow Stock Dives After Analyst Highlights Two-Thirds of Homes Bought Are Underwater,” MarketWatch (Nov. 1, 2021); “Zillow Quits iBuying, Will Lay Off 25% of Staff,” The Real Deal (Nov. 2, 2021) and “Zillow Quits Home-Flipping Business, Cites Inability to Forecast Prices,” The Wall Street Journal (Nov. 2, 2021) [Log-in required.])

Ultimately, I think that the reference to "the pain points of buyers and sellers" is an acknowledgement that the purchase and sale of homes is indeed a human process, one that is full of an unfolding series of steps connected to what can be quite emotional associations and psychological
factors whether one is acquiring a home, or leaving it after many years.  

Make no mistake, automated valuation tools are widely used, not just by Zillow, and have been for a number of years.  But some can be much more accurate, especially when used in conjunction with market knowledge that comes from experience, "feet on the ground", and solid professional knowledge connected to working people.

Julia Huntsman, REALTOR, Broker | www.juliahuntsman.com | 562-896-2609 | California Lic. #01188996

How Much Can You Rely on Automated Home Value Widgets?

Have you looked for a new car lately?  If you're like me, you checked out everything you could find on the internet before you ever went to a dealership for a test drive. 

The websites have interior and exterior 360 degree photos, closeups, zoom capability, video, and even the actual car price, which we equate with value. That's something we love, knowing the price, not the one we have to haggle over. You get everything on the internet except the new car smell--minus one more thing, the actual driving experience.

It's the same with house listings, there's so much available information on numerous sites, and the photos often make them look so attractive that you might think you want that one .  But unlike a car, a house price reflects many more "moving parts", condition, location, upgrades, additions, deferred maintenance, permits or lack thereof, remodel, the immediate surroundings, earthquake zones, flood zones, and much more.  And like buying a car, you really have to be there to see for yourself. The asking price could be very different from value when all is said and done with negotiations and the appraisal.

So what does this mean for AVMs (automated valuation models) such as Zillow and its "Zestimates", and other valuation widgets found on many home search sites?  It means that they are tools, rather general tools, but like driving the car and seeing the house, you have to be there. AVMs can't judge the condition of the house, or know how many prior water damage claims were submitted on it, or check the unpermitted rooms, or see if there are title defects which will prevent mortgage financing.  Zillow values are calculated on public records (strictly data oriented) and user submitted data points (selective pool of information).  By the time a given property is negotiated through buyer/seller agreements, the value may be off by as much as 30% from a Zillow estimate.   Zillow's CEO recently sold his house:

     To see a Zestimate at work, consider the fact that in July of this year Zillow Group CEO Spencer Rascoff listed his four-bedroom home in Seattle, Washington, for $1.295 million. At the time, the Zillow Zestimate valued the home at about $1.39 million. That’s over 7% higher than the list price, but within the Zillow median margin of error.
     Gordon Stephenson, the listing agent for Rascoff, told industry publication Inman that Zillow probably overestimated the value of Rascoff’s home because “its algorithm might not have accounted for the home’s unique floor plan.”  Christian Science Monitor, November 14, 2015.

When an appraiser performs his/her job during escrow, an AVM is not a part of the process. When a Realtor helps a seller or buyer established listing or offer price, knowledge of appraisal parameters may be used along with knowledge of recent sales and individualized comparisons to other properties. 
Algorithms and weights assigned to data can be educational up to a point, depending on what properties get caught in the AVM's net, but when a buyer sees multiple houses combined with knowledge of area sales, the reasons for the final result become known. 
So automated robotic prices are fun to look at and may help educate on overall price range for a given area, but they are not a substitute for a complete home price determination.

See the complete Christian Science Monitor article.