Redfin launched a new estimator tool today, which the company says uses an unprecedented amount of data in order to deliver the industry’s most accurate prediction of how much a home is worth.
Redfin CEO Glenn Kelman noted the offering is similar to the Zestimate tool from Zillow, which has dominated this space since 2006, and others. But, unlike existing tools, he said Redfin has full access to massive amounts of home data from multiple listing services.
That means Redfin’s tool looks at things like whether or not a home has a view, if it’s on a quiet street, whether or not the basement has been finished. It then runs that data through geography-specific algorithms that weigh attributes according to what homebuyers in certain markets are looking for. In the end, the company says it’s able to predict how much a home is worth within a 2 percent margin of error on homes that are listed for sale, and 6 percent for homes that are not.
Zillow, meanwhile, says its median error across the entire country is 7.9 percent.
Zillow’s chief analytics officer Stan Humphries pointed out that the Zestimate tool tracks more than 100 million homes, compared to Redfin Estimate’s 40 million. If Zillow decreased its footprint to just the markets that are easiest to predict, he said the Zestimate tool would be “substantially more accurate.”
Humphries acknowledged that Redfin has easier access to MLS data because it’s classified as an online brokerage, meaning it employs real estate professionals and makes money when homes sell. Zillow, meanwhile, simply lists homes and earns revenue from things like ads. But Humphries said Zillow has been compiling its data market by market since 2006, and now it has “all the MLS data we want.”
My brother recently bought a house in the Atlanta suburbs for $305,000, so I thought I would put the estimators to the test. While Redfin’s tool predicted he would have spent $296,000 (3 percent error), Zillow pegged the value at $257,863 (15 percent error).
Redfin’s tool is brand new, so it hasn’t been fully vetted quite yet. But at least coming out of the gate, it seems like Zillow is about to face some serious competition in the home valuation business. Both tools are free, but they serve as a way to pull valuable homebuyers into the sites.
“Once we knew we could use data from the MLS, we knew we had a significant advantage over anyone else trying to value homes,” Kelman said. “So we just took that to the bank and decided that we could probably process more data using the cloud, we could access more data via the MLS, we could build a second generation estimate that had a lower error rate than any we have seen before.”
