Zillow’s Zestimate tool is one of the company’s biggest pull-ins. The company is putting more behind the product with the announcement of major upgrades to the Zestimate home valuation model. With the latest enhancements, the Zestimate tool’s algorithm has been updated in order to be better connected to current market trends in order to offer even more accurate home price estimates to Zillow customers.
Zillow says that the update will improve the national median error rate a full percentage point up for more than 104 million off-market homes, hitting a median error rate of 6.9%.
Dr. Stan Humphries, Zillow Chief Analytics Officer created the Zestimate tool 25 years ago. He explained that this update goes along with Zillow’s commitment to its user base and offering the market’s best products and services. Humphries said:
"Since we introduced the Zestimate in 2006, we have never stopped innovating in order to provide consumers with the most accurate home valuations. The new architecture we're debuting today represents another significant step forward in our efforts to harness big data to create more certainty for consumers, which leads to better decisions."
Utilizing neural networks, the latest way machine learning operates, the Zestimate algorithm combines this with deeper property data history, like sales transactions, tax assessments, and public records, along with home location and square footage to offer consumers a more accurate price range for their home.
Neural networks are AI systems that imitate how humans think. They can map hundreds of millions of data points, draw connections among inputs, and predict an output based on the relationships formed from those connections. When it comes to the new Zestimate algorithm, the neural networks compare home information, location, housing market trends, and home values to offer a completely new and more accurate home estimate.
Zillow has already begun using its Zestimates tool as part of its initial cash offer service through its home buying program, Zillow Offers. Adding the neural network model to a national real estate dataset was an idea developed by the winners of the Zillow Prize. Jordan Meyer, one of the members of that winning team is now Zillow’s Senior Applied Scientist and works on home valuations for the Zillow Offers service.