Zillow Misses the Mark Completely in Utah

zillowdarts.jpg 

It’s no secret within the real estate industry that Zillow, and it’s Zestimates are not accurate, and really- they aren’t even supposed to be. Accuracy doesn’t even fit into Zillow’s core objectives. Their business model is not about being accurate at all- its about selling ads. But Zillow does create a real problem for many people who believe the ridiculous Zillow price estimates.

Zillow’s business model is simple;

1- Assign a price estimate for every home

2- Generate interest by showing how “fun” it is to see what it says other people’s homes are worth and satelite images of celebrity houses

3- Sell advertising on the website

Nowhere in that business model is accuracy. They want to be accurate enough that people will go to the site, but that’s about it. The state of Arizona’s lawsuit against Zillow only feeds them more exposure. And I’m a supporter of Zillow’s. I have been from the beginning and I advertise with them. But it’s important that home buyers and sellers realize that Zillow doesn’t know anything about real estate. It’s not a real estate company- and that’s the thing that most people don’t realize.

It’s a technology company disguised as a media company disguised as a real estate company.

The people who own and run Zillow have no experience in real estate, and most of the people working for Zillow have no experience in the real estate industry. They are computer programmers, PR and media people.

Apart from how ridiculous it is to actually believe that a computer program can determine the value of something that is as much about feel and flow as it is square feet and bedrooms, the idea that you could take an algorithm to determine values without even seeing what you are evaluating is absurd.

I haven’t put enough thought into the lawsuit the state of Arizona has filed against them to have an opinion, but I do think it’s a shame that consumers would be so fooled by the website that I do believe they need to make their disclaimers more obvious so people realize that zillow is simply a fun place to check out your neighbors house.

It’s like trying to put a value on a person based on a pre-set group of data points and an algorithym that places value to different attributes you possess. Sure, it might be fun and entertaining, but would you actually believe the values?

Sad thing is, many people would…

So let’s take a look at some actual examples of how accurate Zillow is;

First, let’s look at my house…

zillow-richway.jpg

The first thing that I notice is the irrational price fluctuations, up and down and up and down. There was a day in November that it showed my home went down in value $20,000 in one day. Then it went up again over $20,000 in less than a month. And it shows that since January my house has gone down about $40,000. It shows that the house has 4200 SF, 6 bedrooms and 3.5 bathrooms.

The reality is my home has about 4800 SF, 7 bedrooms and 4.5 bathrooms and some of the best city and mountain views I’ve ever seen. Zillow says my home is worth about $420,000 today, but my home was just appraised and could sell today for almost $600,000 and has been appreciating every month for the last year and a half. The graph should look like this;

zillow-richway1.png

Now let’s look at some more homes to see how Zillow has done with them. I pulled up ten homes in my area that recently sold, and compared them with Zillow’s Zestimates to see how close the website was to actual market value and how accurate the site is with home information. This was not complicated. I simply went to the MLS and pulled homes that have sold in the last 60 days or less within a few blocks of my home. Let’s look at these ten homes.

Here is the MLS sheet showing the home information

mlscomps1-1.png

mlscomp2-2.png

And Zillow’s Zestimate’s on each;

zillowcomp1.png

Zillow says $467,000-  2490 SF- 3 Bed – 3 Bath

Actual sold $480,000- 3771 SF- 6 Bed- 4 Bath

Difference= -$12,989 (2.7%) Not bad, but shows value has dropped recently?

zillowcomp2.png

Zillow Says $618,346 – 1032 SF – 3 Bed – 2 Bath

Actual Sold $499,000 – 2682 SF – 5 Bed – 4 Bath

Difference= +$119,346 (23.92%) Home went up over $400,000 in one month?

zillowcomp3.png
Zillow Says $444,173 – 3209 SF – 6 Bed – 3 Bath

Actual Sold $539,000 – 4466 SF – 6 Bed – 3 Bath

Difference= -$94,827 (17.59%) Home took a $50,000 dive?

zillowcomp4.png

Zillow Says- $552,257 – 3784 SF – 3 Bed – 3 Bath

Actual Sold $525,000 – 3754 SF – 5 Bed – 3 Bath

Difference= -$27,257 (5.19%) Look at the value jump up and down

zillowcomp5.png

Zillow Says- $494,478 – 3135 SF – 6 Bed – 4 Bath

Actual Sold $579,900 – 4092 SF – 7 Bed- 4 Bath

Difference= -$85,422 (14.73%) Again showing price drop

zillowcomp6.png

Zillow Says- $525,476 – 3834 SF – 4 Bed – 3 Bath

Actual Sold $600,000 – 3976 SF – 6 Bed- 3 Bath

Difference= -$74,524 (12.42%) Price going up and down 5 times in 1 year

zillowcomp7.png

Zillow Says- $469,925 – 3834 SF – 4 Bed – 3 Bath

Actual Sold $650,000 – 5164 SF – 5 Bed – 4 Bath

Difference= -$180,075 (27.70%) Way under value- not even close

zillowcomp8.png

Zillow Says- $633,693 – 3299 SF – 5 Bed – 3 Bath

Actual Sold $805,000 – 5300 SF – 7 Bed – 4 Bath

Difference= -$171,307 (21.28%) Major price fluctuations

zillowcomp9.png

Zillow Says- $959,341 – 6744 SF – 6 Bed – 4 Bath

Actual Sold $985,000 – 7511 SF – 6 Bed – 5 Bath

Difference= -$25,659 (2.60%) Went up $220,000 in a month

zillowcomp10.png

Zillow Says- $1,091,496 – 6350 SF – 4 Bed – 4 Bath

Actual Sold $1,825,000 – 6550 SF – 5 Bed – 5 Bath

Difference= -$733, 504 (40.19%) At time of sale Zillow was off over $1,000,000!

For all ten homes Zillow was off an average of $152,491 (20.36%)and what is really telling about the Zillow charts is how the home values have gone up and down so dramatically, while the area has actually had a steady increase in values over the last 12 months. Maybe Zillow bases it’s algorithms off the stock market? Whatever they are using in Utah, it’s not even close. Not one home had the home information correct and 9 of the homes were off by over $25,000. Four were off by over $100,000 and Zillow was off by an average of over 20%.

Zillow is a fun site to play on, but when it comes to the business of real estate Zillow needs to put it’s little toy away and let the grown-ups do the work.

 

23 thoughts on “Zillow Misses the Mark Completely in Utah

  1. Hi Greg,

    Salt Lake County has incorrect data for your home. They’re short 400 sq.ft. and a few rooms. Zestimate values are calculated using the public records; your house’s Zestimate value will definitely not be accurate. Thanks for publishing your corrected home facts on Zillow — we are working towards being able to drive Zestimate values off owner provided data.

    Actually, the government records seem incorrect for every one of the homes above. It may be a good thing for the owners’ property taxes but it obviously plays havoc with Zestimate accuracy. I’ve seen bad government records but it seldom applies to the majority of the properties in a county as these examples would imply. Did you randomly choose this set of homes?

    Note that the Zestimate was closest to the sale price (5%) in the single example for which the government records are almost accurate.

    As you know, Utah is one of 12 non-disclosure states and so sales transactions are not public records. Inaccurate home facts aside, our accuracy in Utah is limited by lack of sales data. Our median error in Utah is 9.2%. We’ll also look into opportunities to have our community tell us about sales. But remember that the Zestimates change when we learn more about the house. Sometimes that data comes in chunks which can cause weird graphs. For better or worse, the graphs reflect the historic Zestimates, not a back-ward looking estimate of what the home was worth.

    FYI – if you read all that, thanks. And remember, we’ve got a few more toys on Zillow;
    1) EZ Ads for Blue Roof targeted in a few Salt Lake Zip codes.
    2) If public facts are a problem in Salt Lake; publish a (free) profile on Zillow & offer to help Salt Lake sellers edit their home facts. Free leads!
    3) Publish your clients’ listings on Zillow (it’s free). When you do, their sale price replaces the (incorrect?) Zestimate as the most prominent info on the screen.

  2. Zillow still stinks. Once in a while, they get it right. But even a broken clock is right twice a day.

    If a home seller ever wants accuracy, they get a good agent to help them.

  3. The better question becomes, where does Zillow hit the mark? Texas? Vegas? Bogata? Anywhere Town, USA? Jamaica?

    I like Greg’s synopsis of Zillow’s business model and if anything, I feel sorry for them. They’re exactly like a consumer who recently purchased a plasma screen TV thinking that their viewing experience was going to increase drastically, only to find out that after plugging in and turning on, the picture looked more pixilated that then 17 inch lunch box predecessor that’s collecting dust in the garage. Up conversion my friends, up conversion.

    If you want plasma viewing holy grail 1080p clarity, you have to have the content that provides it (i.e. HD Cable box, HD DVD player, HD DVD). Zillow’s problem is that they built this killer app to view houses but they’re setup on the equivalent of TV antenna (aka crap feed). This feed will not change because local MLS boards control it and you know what those people think of technology… Fear (insert blinking .swf file here reading “FEAR FEAR FEAR”).

    So anyway, that’s a major front for Zillow to tackle, good luck. Getting Rosie O’Donell on the cover of a SI swimsuit issue might be an easier “up conversion” task, but who knows, adobe’s new creative suite sounds pretty good.

    Nevertheless I see Zillow’s major issue being the actual advertising model itself. I don’t see this being successful in this business due to language barrier. CPM, ROS, PPC, CPC and all the other dumb internet advertising acronym’s on the frontend, let alone conversion, A/B testing, complicated landing pages backend, is a language that nobody speaks around here. Even the most tech savvy brokerages thinks arbitrage as a process that happens during a divorce proceeding.

    At the end of the day we can build all the freeway’s of traffic we want, but until someone (data liberators “zillow, trulia”) figures out that freeways without off ramps are like fat kids without cake, I don’t see any Travelocity-esque success stories.

    [Parting shots: Blackberry’s are not meant to write this much text, Google’s acquisition of Double-click is a sign of the times, and I believe that I’m the only REALTOR that has ever attending a techcrunch event]

    Kind regards, -Derek Rey

    P.s. Greg, don’t say I never blog because I just blogged on your blog. Cheers

  4. Derek- it’s about time you show your face again- as always you make some great points. A few things…

    MLS’s may be going away, or at the very lest- they will evolve because they will have no choice. I see this beginning to happen.

    The DoubleClick acquisition fits right into where Google has been headed for four years now- where Google (or someone) needs to go next is a seamless integration of all media into an easy AdWords-type product.

    I attended the Seattle TechCrunch party in May ’06

    PS- I like the new teaser you’ve got up for your coming soon site- Going to be a cool template site, or what?

  5. Zillow is like most out-of-state AVM vendors when it comes to sale data – it isn’t accurate because all homes in Utah sell for “$10 or other good and valuable consideration” according to the county recorder. But thank goodness for the fact that loan amounts are recorded. You can divide all these loan amounts by 75% and – wella – you have a sale price! Whoops, I forgot to tell you that the loan was actually 50% or 95%. That might have a large affect on value. But we can ignore that. That’s why free appraisals are worth what you pay for them. There is one exception that I know of. There is a company in Riverton named Bedrock Appraisal that we have used to substantiate a value before ordering a full appraisal when doing investment properties. They have a certified non-inspection appraisal product that they will sell you for $40 on a URAR form with full property data and gridded comps. Since comp searches are illegal in Utah anymore, its the most cost effective way of getting a pretty good idea of the property value, and you can verify the data. I’ve seen as many as 15 actual sales from the MLS gridded in their report. We used to pull the comps ourselves, but its cheaper for us to pay a small fee and have the foot work done for us. Then if the value seems to be there, we go ahead and get a full appraisal done through the lender or they’ll credit their fee toward an upgraded order. Bottom line, don’t rely on vendors outside of Utah that can’t get access to MLS data. Remember the old adage – garbage in, garage out!

  6. Of course Zillow’s estimates are not accurate. A homevaluation machine can not accurately determine the price of a home. It can only come close, and in some instances not at all.

    Zillow, however, stives to give the impression that their Zestimates are accurate, calling themselves the Kelly’s BLue Book of Home Valuations (while trying to convince the Arizon board of appraisers that they are not providing appraisals with out a license).

    At HomeGain we also provide an instant free homevaluation-but we provide a range and also show the comparable recent home sales.

    try it at http://www.homegain.com

    We believe that that an automated homevaluation can only give the consumer an indication of a home’s worth.

    We believe there are things that an automated homevaluation do not and can not take into account, including the condition of the home and the role that a Realtor plays in setting and negotiating the price of the home. Realtors know their local market and are local experts on the housing market.

    For that reason Homegain encourages visitors to try our free home valuation tool. After receiving your free home valuation estimate we encourage visitors to either contact the featured real estate agent whose photo and contact details are presented along side each instant homevaluation that we render or to participate in homegain’s agent evaluator program in order to select the best realtor to help consumers to get the best price possible for their homes. http://www.agentevaluator.com

    BTW-HomeGain created the first instant home valuation tool seven years ago and just last week revivied the tool.

  7. Zillow is inaccurate in Utah because Utah is still (foolishly) a non-disclosure state.

    Realtors make sure their power in the Utah legislature isn’t challenged so this status will be very difficult to change. As long as the Utah Board of Realtors controls most of the market data it will be difficult for anyone not associated with their monopolistic organization to provide an accurate estimate of value of any property in Utah. Your criticism of Zillow seems a little hollow considering you as a Realtor are helping maintain an environment with as little publicly available market data as possible.

  8. Jeremy,

    The fact that Utah is a non-disclosure state was stated in my blog post, in David (from Zillow)’s comments, and in about three or four of my previous posts.

    If you think I am in favor of “maintaining an environment with as little publicly available market data as possible” then you obvioulsy have not read through my blog or looked at the real estate website that this blog is named after at all.

    You may want to do a little research before you try calling people out because you just sound foolish saying those things to the person who has probably been the most vocal in the entire state of Utah about changing the “environment” you speak of.

  9. Greg,

    David’s comment does make a point about Utah being a non-disclosure state but I didn’t see anything about that in your post.

    You did do some posts about Utah’s non-disclosure status and its affect on Zillow in January but according to your site’s search functionality those are the only posts on this blog with the phrase “non-disclosure” in the body.

    I have only been reading your site for the past couple of months and I really do admire your work. Your efforts to provide more market information for your clients and anyone else who checks this site should be commended and I’m sorry that I foolishly said you personally were helping “maintain an environment with as little publicly available market data as possible”. Utah’s non-disclosure status is a hobby horse of mine and I let my negative opinion of that policy overcome in my mind the obvious fact that your website itself is proof of your willingness to provide market data that very few other realtors are able/willing to provide.

  10. Greg –

    Is this a random sample? What is your experience with the general accuracy of the assessor’s data in Salt Lake?

    Louis –

    Kelly Blue Book does not tell you the exact value that your car will fetch. It’s actually a great analog for Zillow and one which clearly suggests that Zillow is transparent when it comes to Zestimate accuracy. In fact, we measure and publish Zestimate accuracy on Zillow.

  11. Greg you always seem to tell it like it is and I like that about you. Zillow is way off in my neighborhood too. Most of the homes are very similar in square footage and lot size, yet Zillow has homes priced over $50,000 different right next to each other. What sense does that make?

  12. Its the height of arrogance or stupidity to think that an algorithm can precisely determine the price of a home based on prior home sales and the data a few homeowners chose to add to mix.

    Alchemy is a more valid science that Zestimating the value of a home.

    Homegain created the original instant home valuation tool almost a decade ago.

    While it is a clever innovation (for the 20th century), listening to some journalists and paid Zillow bloggers, they would have us think that Zillow wasn’t just pulling data from public records and putting pins in maps to plot the results but rather changing the way the entire world thinks and curing cancer at the same time.

    A house is only worth what you can get for it and a Realtor helps you get your price. Trying to value homes precisely through old market data is not the same as the minute by minute market values provided by Nasdaq with respect to heavily traded stocks.

    Because of the transparency of the stock market you don’t need a stock broker to get you a better deal on Yahoo or Microsoft stock.

    But often to get the best deal on your home you need a Realtor.

    HomeGain’s service is designed to help consumers who are interested in working with Realtors find one.

    Zillow can continue with their Zalchemy while Homegain focuses on matching consumers to realtors.

  13. Yawn.

    “The people who own and run Zillow have no experience in real estate, and most of the people working for Zillow have no experience in the real estate industry.”

    Which is why Zillow came up with what is a pretty decent service and the real estate industry did not.

  14. pwb,

    Maybe you just want to say something against the real estate industry, but you could probably find something that makes more sense to slam the industry for (lack of information from the MLS’s, commissions, etc).

    Being off by over 20% isn’t “pretty decent” if you’re looking for a valuation service. If it’s just entertainment you want (which is what Zillow is good at) go to YouTube- you don’t the real estate industry for that.

    As far as searching for homes I feel pretty confident putting BlueRoof.com against Zillow any day- and the comps on BlueRoof are actually relevant.

  15. Greg –

    Most Zestimates are not 20% off – far from it. Even in Utah. More than 4 million monthly visitors to Zillow seem to agree that it’s a good starting point. Don’t forget that there is much more to Zillow than merely Zestimate values — it’s a great place to advertise Blueroof’s local services.

    Since you won’t answer my questions, I suppose I must assume that you cherry-picked these examples. That’s probably good news for Salt Lake tax payers – I’d hate to think that all the County’s property records were generally this bad. The primary reason for the horribly incorrect Zestimates above is that the public records are also horribly incorrect.

    If you inform the County about the changes they need to make, those will be reflected on Zillow – usually after the next tax roll is published. In the meantime, you can publish corrected facts for your house on Zillow. You can also use those updates to get a refined estimate using the “My Estimator” on Zillow.

    Louis –

    Zillow does not pay bloggers. That is ridiculous. The rest of your comment also makes no sense — no-one said an AVM can precisely determine the price of a home. Where do you get this stuff?

  16. David,

    I didn’t cherry-pick examples to use. As I stated in the post- I took my house and the ten homes that most recently sold around my home. I figured that was the easiest way to select a group of homes- and by using my home I could give an example where I had intimate knowledge of the value.

    That’s pretty simple. My house plus the ten most recent sales- no cherry-picking. I understand how you don’t like the results I found, but the results are a result of Zillow’s inability to accurate judge Utah’s values, partly because of Utah’s non-disclosure status and partly because you cannot place a value on a home you haven’t seen.

    I realize and have acknowledged that Zillow is a “starting point”. Everyone who has knows much about Zillow knows that’s your motto because it is said in every conversation and blog comment left.

    I have also stated nunmerous times that I like Zillow, I advertise on Zillow, and I think it’s a fun site.

    This post is not about if it’s a fun site- it’s about the accuracy in Utah. Zillow is not accurate in Utah. Frankly, I’ve had conversations with many clients about what they found on Zillow and it’s been a pain in the ass trying to explain and educate over and over again why a website that is so well known is so damn off. Many people think that because Zillow gets a lot of press and traffic that it’s accurate and that presents a problem when a home is worth $450,000 but the seller wants to sell for $500,000 because that’s what Zillow says it’s worth.

    Zillow has it’s place and you are defending your company, which I understand. I also realize that there are other things on Zillow- but the site was built around, and it’s buzz is about, it’s home value estimations. And that is what this post addresses.

  17. Zestimates are just that. They are not proper, helpful or accurate. They are completely worthless. Values are always changing, and home sellers need a professional’s guidance and consultation. Even FSBO’s have 2-3 agents to their home to help them price it.

  18. zestimates are worse than worthless, they’re dangerous to your financial well-being. what’s the point of having a ‘starting point’ that is so inaccurate in so many respects.
    they call that roulette

  19. I sold a home in Cottonwood Heights six months ago. The Zestimate at the time was high IMO, considering that the home was in serious need of updating and remodeling, so my selling price without an agent was at the low end of the Zestimate. Based on tax records and builders specs, the home has 3,200 square feet. That part of the Zillow info was accurate, but the home is now listed for the “flipper” by a Realtor as having over 3,800 square feet according to his “measurements.” No exterior walls were extended or floors added, so considering how my home was expanded for resale, I would advise potental buyers to have ALL realtor supplied interior square footage dimensions checked for accuracy before making an offer.

  20. As an appraiser, I think Zillow is contributing to the unrealistic value expectations I’m encountering with internet-savvy sellers. We sure don’t need any more bad data out there than we already have.

  21. One thing I noticed in your examples … with one exception, Zillow’s prices were all BELOW the actual sales price; which leads me to believe that if Zillow says that my home is worth $200K, it is actually worth more than that.

    However, it’s been 3 years since your post, and market values have now dropped in Salt Lake City … perhaps an updated post showing current accuracy is in order?

Leave a reply to Louis Cammarosano Cancel reply