There is always room for growth.
Zillow is a website well-known for offering the public home-values and other real estate information for specific areas. The company, noticing it’s margin for error, decided to look into improving the analytics behind their home-values.
Between 2006 and 2017, Zillow’s error rate has gradually decreased from 14% to 5.7%, according to the Florida Association of Realtors.
The estimates published on Zillow for homes would often provide an unrealistic expectation of median prices for home sellers. The data scientists behind the “Zestimate” created a world-wide competition a year and a half ago, to see who could best improve the analytics behind the Zestimate.
This past week, three men have been awarded $1,000,000 for their work in improving the Zillow algorithm. The process of improving the algorithm included looking at hundreds and hundreds of differing data points that could affect and influence the possible price of a home.
According to the FAR, “homes nationally sell on average for about 2% less than the list price set by brokers… The median error rate of 4.5% nationally means half of home values are wrong by more than 4.5 percent.”
With more research on the various data points, the algorithms should continue to shorten the gap.
For more information, click here for the Florida Association of Realtors article on the improving Zillow algorithm.