Zillow's claim is that in the state of Virginia, their "Zestimate" is within 1.4% 50% of the time. The other 50% of the time it is off by more than that. That claim is unimpressive, to say the least. It's akin to bragging about your grade after you cheated on the test. Zillow takes the listing data (a real estate agent's informed estimate of sales price) and adjusts up to the minute their "estimate" based on that.
However, they also claim that their off-market accuracy in Virginia is within 5.5% 50% of the time. This begs the question, how do they claim that? If the property never goes to market, how can they claim accuracy?
A "Zestimate" - has no accountability, recognized methodology, or credibility.
An appraiser - is accountable to the State Board of Appraisers, is required to use recognized methods and is the only qualified professional to be able to determine market value, by law.
So, I decided to take a random sample of off-market refinance appraisals and see how Zillow stacked up in estimating their market value. The Pie chart below shows the percentage of times the Zestimate was within a 7% range (ie. "7%" is the 0-7% error range, "-7%" is the 0- -7% range, etc)
As it turns out, Zillow's claim of being within 7% 50% of the time... is approximately 100% wrong for Central Virginia.
The sample showed that 52% of Zillow's zestimates were within 14% of the market value of the property. Approximately 100% more than their claim. 48% of the sample was wrong by more than 14%.
Given that the most recent results of Zillow buying properties to flip them is resulting in a loss of 6% to the company, I think its fair to ask...
Would you trust those numbers with your most valuable asset?