Automated home valuations are a growing trend on real estate websites. While Zillow’s Zestimates are clearly the standout leaders in this field, there are plenty of others. Eppraisal. Cyberhomes. Trulia has a one-city beta product in the works, while First American has had its own automated valuation model for quite some time. Even our own Geek Estate Blogger, Justin Britt, has a new Hawaii Life Value Score, taking a different angle with more of a comparative score valuation.
The vast majority of real estate professionals I encounter dismiss the value of these models, sometimes for very good reasons which I’ll discuss later. Technology enthusiasts like myself, however, find the same fascination with these valuations as the average consumer. It’s a brave new world in the real estate industry, and the amount of intellectual and financial horsepower being driven toward these online outlets of real estate data seems almost unlimited. In the end, though, my concern is with who really benefits from online automated real estate valuations?
Clear Benefit: Real Estate Consumer Websites
Zillow, Trulia, and their cohorts are in the business of differentiation. There are already tens-of-thousands of agent-based local websites delivering accurate, timely listing data for consumers. Recreating this same database on a national scale doesn’t add value for a local home buyer. To pull a consumer away from their comfortable neighborhood home, something shiny and new is required.
The national real estate websites are leveraging their superior finances and workforces to create tools that an average agent cannot. Automated valuations create a unique, interesting, and (potentially) financially valuable tool for consumers. When a home buyer begins using one real estate site to compare home values vs. home prices, he or she is likely to continue using that same site for a long period of time, to guarantee the consistency of the metric being used in the search. This creates a consumer who not only views a large number of pages within a site for comparison purposes, but does so repeatedly, day-after-day. Consumers will many times become quite loyal to the site, out of appreciation for its unique value, or even out of fear of making a mistake by not coming back and consulting this source of information.
This loyalty and traffic, generated by the unique content, creates a great amount of SEO and financial value through advertising. For a real estate portal with no associated brokerage actually selling homes, this is the number one goal. Profits come from advertising, and advertisers buy traffic visibility. Make no mistake, Zillow didn’t become one of the world’s most popular real estate sites because the forums had nice discussions. Zestimates were the sparkling new bauble leading the news media and the public through the door and allowing the company’s introduction to a massive audience.
Unintended Benefit: Real Estate Macro Market Statistics
Although I can’t guarantee this, after watching a company like Zillow for years, I don’t imagine that the founders ever thought that their market data would be used in the way it is today. The tools, in the beginning, were designed to compile large amounts of data and apply it to one home, creating one estimate of value at a time. The massive undertaking and collection of data sources, in the end, created an unexpectedly-accurate database for national real estate forecasting that is nearly unmatched in size.
Stan Humphries, Zillow’s chief economist, has been making a name for himself by beating Case-Shiller to the punch on housing statistics of late. The statistics are actually a very good thing for the overall real estate market, as they add another voice with a large body of credible data to back it up. The extra publicity doesn’t particularly hurt the company either, as Humphries and CEO Spencer Rascoff are regularly featured as national real estate figures based on their data.
Partial Benefit: Home Buyers
This point has been made by real estate agents ad nauseum, but I’ll repeat it because it is valuable. Automated home valuations have a small benefit to a buyer in their home search. In an introductory phase, buyers may want to see sales prices and values in a general area to begin to understand the area in which they’ll buy. Estimates may flesh out the neighborhood a bit more than simply viewing recently closed sales. The additional data could help a buyer understand that most homes in their preferred neighborhood are worth between $350,000 and $400,000.
That’s it. That is where the value ends for a home buyer. Estimates on individual properties are not worth the screen they’re displayed upon. Since Zillow is the most popular, and I know and like a bunch of people at the company, I’m going to pick on them in particular.
Nationally, Zillow’s estimates missed the sale prices of homes 45% of the time by more than 10% of the value. Zillow’s estimates missed the sale prices 25% of the time by more than 20% of the value. From these two statistics alone, it’s clear that you’d be foolish to make a home purchase decision based on an automated valuation.
This isn’t an opinion, it comes directly from Zillow’s statistics. The company itself points out that its estimates are not to be used as an appraisal, and is highly transparent about its own inaccuracies in the data for individual homes. Unfortunately, it seems that there are far more consumers relying on Zestimates than there are actually reading those disclaimers. Buyers should understand that once they’ve gotten past the initial phase of scouting out a neighborhood, using an automated valuation model isn’t going to provide accurate information on individual homes’ values.
Net Zero Benefit: Home Sellers
Home sellers register the most complaints with real estate professionals over automated home valuations, with good reason. While home buyers might miss a potentially good buy because they relied upon an online valuation, there are always other homes to buy. Sellers, on the other hand, have just one iron in the fire. If a seller lists his/her home for sale and an inaccurately-low value is being posted next to their house online, the seller is stuck explaining a botched mathematical calculation to buyers instead of promoting a home’s best features. There is no upside to a high valuation. Sellers have already named their price, and buyers won’t be increasing their offers because of the high valuation. While the number of potential buyers who may undervalue this home based on that valuation is unknown, it creates a great amount of anxiety for a home seller that they’re being unfairly judged by an outside source with little direct knowledge of the home.
All of the major real estate sites claim to have procedures for human input on their valuation models to adjust a poor valuation. This is very temporary band-aid, however. Human input will always be subjective. High valuations won’t be appealed. Listing agents and sellers will always appeal low valuations, even if their home is overpriced. In the end, human input altering the outcome of the model just reduces the intergrity of the model itself.
Automated Valuation Models Have a Long Way To Go Before Truly Benefitting Consumers
We’ve certainly seen some benefits to real estate statistics and greater market knowledge, and I enjoy watching another statistical powerhouse competing with the old guard for the most accurate industry projections. It’s fairly clear, though, that these automated valuations have benefitted their creators much more than individual home buyers and sellers. In the spirit of technology progression, I’m hoping that these algorithms are developed to the point that they become somewhat-accurate predictors of value. However, as much as I’m admirer of our ever-increasing ability to manipulate data for research purposes, I’ve sold far too many homes to believe that a computer model can compare one neighbor’s shiny granite counters to another’s crumbling laminate. The data is a nice tool for economists, but for the individual consumer it’s more entertainment than sound financial advice.