Alan Dalton recently suggested http://www.youtube.com/watch?v=0DNtYtM_gX0 that Zillow’s Zestimate’s are really nothing more than a modern day parlour game. Fair enough. There are many of us who agree whole heartedly with Mr. Dalton. Some of you out there may disagree. My default, like many of you, is to make a gut call on a lot of controversial issues. Zillow’s Zestimates seem a little too good to be true to me–my gut said they were bogus. But as someone who spent many years teaching and studying the Scientific Method (pre-med in college, former Biology and Genetics teacher) I decided that the only way forward was to take a look at the real estate numbers in my corner of the world: Boulder, Colorado.
For the month of March, 2007 there were 11 sales in Boulder’s 80301 zip code (just one of 4 zip codes in the area…I’ll follow up with the others). I’d like to examine how well Zillow did with their Zestimates vs. how well the Boulder Realtor(s) who sold the properties did with their pricing.
Here’s a very simple spreadsheet with the relevant information:
There are only 10 properties here because I was not able to get a Zestimate for the 11th. The 11th property was 4525 Robinson Place, Boulder, CO 80301. This was a listing presented by Karen Bernardi, the uncontested queen of Colorado real estate. The reason I couldn’t use it is because on March 29th, when I pulled these numbers, the Zestimate reflected the current public recording of this sale. I’m sure there is a way to go back and see what the Zestimate was right before this sale (please let me know if you know how to do this), but I don’t know how to do that, so I tossed this one out.
The results are:
Zillow was within 5% of the sales price 70% of the time. The Zestimates were within 10% of the sales price 80% of the time.
The Realtors scored within 5% of the sales price 80% of the time, and were within 10% of the sales price 100% of the time.
Many of you may think that this is an exercise in redundancy, as Zillow keeps it’s own internal statistics. The scientist in me realizes that a sample set of 10 data points is nothing more than anecdotal evidence. But still, you’ve got to start somewhere.
I’ll continue to keep track of these numbers, but it seems from this anecdotal representation that Zillow is doing what they claim: they are offering the consumer a starting point for home valuation. http://www.youtube.com/watch?v=THg7BVqPRuc

Comments (10)Subscribe to CommentsComment