Do I need an Estimate or Zestimate?

If you have not already noticed, Zillow was on the cover of Fortune Magazine. Lately, Zillow seems to be the worlds most effective PR machine. There is also a very different article in the Wall Street Journal today that sites Zillow’s inaccuracies. I recently had a run in with an inaccurate Zillow Zestimate that I thought would be fun to share.

Above are two Zestimates I performed using Zillow on February 13, 2007. The two properties are right next door to one another. 4347 Harbord was sold in April of 2004 for $1,152,000. 4355 Harbord was on the Market and recently expired going unsold. I have physically seen the inside of both properties.

First, I can tell you that the Zestimate on 4355 Harbord is inaccurate because there are approximately 172 livable square feet that are not accounted for in the public records and thus they are not accounted for on Zillow’s site. The real square footage of the property is approximately 1,900 square feet. This has long been one of Zillows major problems that they will have to fix in order to make their Zestimates more accurate. Much of the public records in the Bay Area are incorrect. Zillow’s Zestimates rely on this data. While I was visiting the property I also read the disclosures. The house was pretty clean, but don’t ask Zillow about disclosures. Their Zestimates do not take property inspection reports and seller disclosures in to account. The moral of this story is that if you put garbage in to an equation you get inaccurate garbage out. Zillow has attempted to ameliorate this issue, by allowing users to alter their home’s data on their site. Would you trust a seller to accurately post their data? If you do then my 10,000 square foot home just went up for sale!

Second, if the square footage of these two homes is now almost identical, how can there be a $600,000+ difference in the Zestimates. My guess is that Zillow is weighting the more recent sales data of 4347 Harbord that sold for $1,152,000 in 2004. It is important to note that Zillow’s Zestimate is a complicated algorithm. An equation that is not public knowledge might I add. No bank would accept the validity of an appraisal if they could not follow the logic behind the report. Yet, Zillow is feeding the public data that is faulty all the while leaving their evaluation methods shrouded in mystery.

Third, Zillow is claiming that they can address a home like a stock. Check out the thirty day change above. How could one home go up by $110,201 in the last 30 days while the other one goes down by $30,782. For heaven sakes the two properties are right next door to one another. When we value stocks in real time like on Etrade, it is based on actual sales that are occurring every second constantly altering the value of the stock. Since most homes are not even sold from one year to another trying to evaluate price change every thirty days is crazy. Although viewing houses like stocks may be intriguing it is just not accurate. I will be the first to tell you your house is not an ATM and it is not a stock.

I am bringing this to your attention, because I viewed 4355 Harbord. I viewed the property disclosures. Their was nothing dramatic about the disclosures (another important piece of the value equation that Zillow knows nothing about). The property was listed at $930,000 and has now expired going unsold. If you are a seller you may want to visit Zillow’s site before listing your home and while you are at it check out my previous article, Zillow - How Inacurate Data Could Harm Your Homes Value.

My personal opinion is that if a company puts an exact price on a home that is so exact as to be quoted to the dollar (ie. the Zestimate on 4355 Harbord of $811,255) then they are influencing the market to believe that the stated price is the exact price. Even if they post a range of values below. If a company estimates a 30 day trend, they make themselves look like an exert who has been tracking value over time. Even if their original estimate is completely inaccurate. Don’t get me wrong, Zillow is interesting and it is good for estimating general price data. I recently used Zillow’s heat maps to help a client who was moving out of the area and needed a general sense of neighborhood pricing in a city out side of California. It is good at generalities, just don’t rely on it for specifics.