It's been several years since I did a Zestimate analysis. I am working with a buyer who seems to put a lot of trust in Zestimates in determining what price to offer on a home. Of course, buyers will always trust a Zestimate instead of any data you put in front of them to the contrary.
So I thought I would see for myself if Zestimates hav gotten any better in the area that I'm working with my buyer.
I've heard experts say that a home is only worth as much as someone is willing to pay. Therefore, I took all of the sales over the past 30 days in the Dunwoody/Chamblee area and compared the actual sales price to the current Zestimate for each property.
There were a total of 74 sales, single family homes. I took out the seller paid costs and only compared the net sales price to the Zestimate.
Zestimates differed from the net sales price by -32% to +45%.
I would say that to have a Zestimate be worth anything, it would have to be within +/-5% most of the time. This would give a range for a $400,000 home to have a Zestimate between $380,000 to $420,000.
Of the 74 sales, only 19 (25%) fell within this +/-5% price range.
That means that 75% of the Zestimates are more than 5% off from what the actual sales price was.
To me, that is not good enough to be of use as a serious valuation tool.