The Accuracy of a Zillow Zestimate

[photopress:2faced000.jpg,thumb,alignright] Much has been written about the accuracy, or lack thereof, of the Zillow Zestimate of a home’s value. What one must remember is that a property can sell at the low point or the high point of its Zestimated range.

I don’t pay much attention to the articles written on bubbles bursting and what kind of market we are in, because I always know what kind of market I am in. I feel it in my bones, the same way an old person can tell that it’s going to rain before the weatherman predicts it. I know just how far I can push a price in either direction, depending on market conditions and who I represent in the transaction.

Every agent wears two hats and is two-faced, because a home’s “value” has to be higher when I represent a seller and lower when I represent a buyer, and it is my “job” to “make it so”. The Zillow range of value represents my best hope for my buyer client at the low end of the range, and my highest hope for my seller client at the high end of that range. I have yet to meet an agent in the Country who can jump back and forth over that line as well as I do. I guess that makes me two-faced, but being very good at being two-faced has always been my forte.

When I represent a seller I try to get the seller to give me the key to his house for a couple of days and go away and give me “carte blanche”. Mainly because I look very odd when I am “doing my seller thing”, somewhat like “MONK” at a crime scene. I keep going out to the street and walking up to the house, at various paces from all directions, emulating a buyer getting out of the car from every possible available parking spot. The neighbors must think I’m looney.

I trim trees and bushes based on the angle of the “walk up” and what I can see and what I can’t see. “Good, that bush blocks that window frame that needs painting…bad, that tree is blocking the main feature, the rounded brick archway…then I trim the tree to “accentuate the positive” until I can see the brick archway from the position of the buyer driving by or getting out of the car, and leave the bush overgrown to “de-emphasize the negative”.

I walk into the front door 25 or more times and change things, until I remove any negative influence in my sight pettern (which is eye level side to side, without looking up or down). I always tell agents, “if you are standing in one place when you are staging a home you are “decorating” and not “staging”. Walk through and walk fast. Remove negative influence or distract the eye away from the negative with a bright vase or photo in the opposite place from the negative. If you can’t eradicate the negative, draw the eye toward the positive. That is staging, and that is why the agent has to do it themselves and not hire landscapers and decorators.

Staging is about the real estate, and a real estate professional must be in charge of what will and will not be done, to enhance the sale price.

Conversely, putting on my other hat, I take out my other face when evaluating homes with buyers. We both step into the house, and usually I walk one way and they walk the other, and I see things quite differently than they do.

They say they “love” it and I say “Oh, my God WHY?”. I get them to focus and point to what they like. Sometimes by forcing them to tell me what they like, it turns out to be a painting or a piece of furniture. I say great, let’s find that painting to put in your new house, but for now let’s go back and try this again and look at the “real estate” of this place. Sometimes, if it is vacant, I actually have to move the staging so they can see what I see.

To achieve the lowest possible price for my buyer, I either have to find a seller who has “left money on the table” or I have to find a property that is overpriced. If a house could have sold for $510,000 or $515,000, but the seller priced it at $519,000 and staged it incorrectly, I can usually get it for $500,000. That’s a standard best case scenario. If a property comes on at less than fair market value, which happens on occasion, I can usually swoop in and modify terms, to grab it while the vultures are still hovering.

Often people comment on the Zillow Zestimate wondering “Where exactly is the value of this house? Is it closer to the high end of the range or the low end of the range?” The answer is it is the agent’s job to pull “the value” in the direction of their client. When I represent the seller, I have to DO something before I hit that “live on the mls” button that makes it go higher. When I am representing a buyer I have to DO something to force it back in the other direction.

I pretend that all my clients are Captain Kirk, who command me to “make it so #2” πŸ™‚

42 thoughts on “The Accuracy of a Zillow Zestimate

  1. You may be a good real estate agent, but you do a poor impression of a Star Trek fan. “Make it so” and referring to the first officer as “Number Two” is what Captain Picard (of Star Trek: The Next Generation) is known for. Not Captain Kirk.

    Yes, I’m a complete geek and I couldn’t let that just slide.

  2. You may be a good real estate agent, but you do a poor impression of a Star Trek fan. “Make it so” and referring to the first officer as “Number Two” is what Captain Picard (of Star Trek: The Next Generation) is known for. Not Captain Kirk.

    Yes, I’m a complete geek and I couldn’t let that just slide.

  3. It was Picard in my head, but Kirk came out of my “mouth”…talk about fantasizing. I could have written several episodes with me, Picard and “Number Two” that would have put them in the “Hall of Shame” category πŸ™‚

    The whole “make it so” analogy came from my brother’s definition of a “real estate agent” that would be “worth it”. He said, I want to point to a house and say I want “that one”. I want the agent to tell me if I shouldn’t want “that one”, and show me which one is better than “that one”.

    When the agent and I agree on the appropriate “that one”, I don’t want to talk to the agent again EVER. I just want to pack up my stuff and go to my new house and have the agent 1) get the house for me and 2) go back and sell the house I left. He’s a singer πŸ™‚

    Actually, one of my current transaction(s) is working out exactly like that…but I wouldn’t have “operated” in the manner I do without the benefit of my brother’s “make it so” definition of an agent conveyed to me many years ago.

    Thanks David G…David Gerard that is, not Mr. Zillow πŸ™‚ I always do a double take when David G. comments. Have to make sure it’s not my brother pulling my leg before responding.

  4. Pingback: BloodhoundBlog - The weblog of BloodhoundRealty.com, an exceptional-service residential real estate brokerage in Metropolitan Phoenix, Arizona.

  5. Zillow is consistently wrong.

    Look, I think it’s a wonderful idea to create better online tools for sellers and I like the scope of their vision, but I think my golden retriever could do better at valuation.

    Here’s the latest. We just listed this 1,200 sqft house at $319,900. Dallice and I not only live in this neighborhood (2 streets over) but I’m the secretary of the neighborhood association. We also spent 4+ hours developing our opinion of value. You could say I’m familiar with the comps.

    Zillow thinks it has only 792sqft and is valued at $339,087. Off by $20,000, more than the total commission to both sides of the deal. And it’s not the first time.

    What a joke.

    p.s. I love that extra $87 bucks they throw in at the end, wrongly suggesting Zillow’s accuracy.

  6. Zillow is consistently wrong.

    Look, I think it’s a wonderful idea to create better online tools for sellers and I like the scope of their vision, but I think my golden retriever could do better at valuation.

    Here’s the latest. We just listed this 1,200 sqft house at $319,900. Dallice and I not only live in this neighborhood (2 streets over) but I’m the secretary of the neighborhood association. We also spent 4+ hours developing our opinion of value. You could say I’m familiar with the comps.

    Zillow thinks it has only 792sqft and is valued at $339,087. Off by $20,000, more than the total commission to both sides of the deal. And it’s not the first time.

    What a joke.

    p.s. I love that extra $87 bucks they throw in at the end, wrongly suggesting Zillow’s accuracy.

  7. It’s true. Zestimates are not pinpoint valuation of one’s home.

    I’ve found it really strange that so many folks are wrapped up in the idea of Zillow’s accuracy. I think most people know to contact their local agent or appraiser for better and authentic valuations.

    Perhaps I’m on my own island of thought here, but the impact of placing sold property information online speaks for itself. I would guess it ended up costing CBBain and JLS, among others, a ton a money to follow along and include sold property data. Never mind the cost of losing mind-share or coveted online eyes.

    Did JLS, Windermere or CBBain or others have “must include sold data” on their to-do list for web development before Zillow hatched? I don’t know that they did or would—may not have gone over too well with their constituents.

  8. It’s true. Zestimates are not pinpoint valuation of one’s home.

    I’ve found it really strange that so many folks are wrapped up in the idea of Zillow’s accuracy. I think most people know to contact their local agent or appraiser for better and authentic valuations.

    Perhaps I’m on my own island of thought here, but the impact of placing sold property information online speaks for itself. I would guess it ended up costing CBBain and JLS, among others, a ton a money to follow along and include sold property data. Never mind the cost of losing mind-share or coveted online eyes.

    Did JLS, Windermere or CBBain or others have “must include sold data” on their to-do list for web development before Zillow hatched? I don’t know that they did or would—may not have gone over too well with their constituents.

  9. Zillow provides valid SOLD data to the public. Zillow’s “value range” is very useful. These are good things in my book. People do seem to get hung up on that “Zestimate” which is amusing, since the Zestimate is never a nice, round number, anyway.

    I’m guessing the Zestimate is generated by a formula that first computes the price per square foot, then applies a little calculation such as “if FIREPLACE field is Y, add XYZ; if VIEW field is N, subtract XYZ”.

    Brilliant. Absolutely brilliant. I wish I’d thought of it. πŸ™‚

  10. Zillow provides valid SOLD data to the public. Zillow’s “value range” is very useful. These are good things in my book. People do seem to get hung up on that “Zestimate” which is amusing, since the Zestimate is never a nice, round number, anyway.

    I’m guessing the Zestimate is generated by a formula that first computes the price per square foot, then applies a little calculation such as “if FIREPLACE field is Y, add XYZ; if VIEW field is N, subtract XYZ”.

    Brilliant. Absolutely brilliant. I wish I’d thought of it. πŸ™‚

  11. Just FYI – You can use the “Zestimator” to adjust once you know the actual details of the home. I find that even with some basic adjustments, zillow’s not that bad. For example, zillow initially estimates my house at about $300k less than it recently appraised at. However, if I add in the bathroom and bedroom and finished basement we have that zillow (and King Count, ahem) don’t know about, it’s actually just slightly over the appraisal. And I’ve had some independent confirmation (couple different agent friends) confirm the range is about correct (they were a little more optimistic, but no one was wildly different).

    I wouldn’t use it as the last word in anything, but it is a useful/fun tool. Then again, it will be interesting to see the fallout if buyers start saying “if zillow says the house is worth $x, I’m not paying more than $x” – which would then get sellers really POed that zillow puts up numbers that don’t accurately reflect their home’s details. I’d hope that buyer’s agents in those cases start saying “you need least factor in that this is a computer program that hasn’t seen this house. Or at least make sure you use the zEstimator to include all the house’s actual details before you rely on the range”.

  12. If you asked me to define why Zillow works, it is not because their estimates are “right”, but rather they are “good enough” for many situations. Sometimes just showing up to the game is the most powerful way to have an impact!

  13. It humors me whenever agents defense to Zillow is to claim Zillow is inaccurate. But, then, what’s accurate. Have 5 agents do price opinions and you’ll likely get 5 different values. They all can’t be right ( & accurate). Pricing valuations are nothing more than someone/something’s opinion.

  14. Osman said:
    “We just listed this 1,200 sqft house at $319,900… Zillow thinks it has only 792sqft and is valued at $339,087. Off by $20,000…”

    So Zillow value estimate is within 6% of Osman’s careful professional evaluation. IMO, that’s actually pretty good for an automated Web tool that’s trying to project values based on limited information. Are valuations from professional to professional always as consistant?

    Of course, not all Zestimates are as accurate… and the sq.ft. value in the above example is pretty off…

    I’m just wondering… Would it be better in any sense if Zillow (or any similar tool) presented more prominent disclaimers/indicators of the expected error rates of their values — or perhaps presented the values as +/-7% or so ranges, such as “$315,351 to $362,823”, instead of a single exact-sounding price value?

    Or would Zillow be too concerned about not being taken as seriously if they present less-than-exact-looking values?

    Or is the whole effort over producing a tool like this more about making a splash and buzz over a cool new feature than about making that feature as good and usable to the consumer as possible?

    Just wondering…

    Oh, and Picard called his first officer “Number One”, not “Number Two”. πŸ˜‰

  15. The 20,000 that Zillow is off by falsely inflate sa seller’s impression of what her house is worth. It may only be six percent, but as Ardell’s latest post right points out, a small difference exists between a successful transaction and dead wood.

    In this kind of market, it’s a critical mistake. The homes that are selling are doing so much faster than last year. Meanwhile, the ones that are not selling are sitting for much longer.

    Mispricing is probably the worst mistake you can make.

  16. The 20,000 that Zillow is off by falsely inflate sa seller’s impression of what her house is worth. It may only be six percent, but as Ardell’s latest post right points out, a small difference exists between a successful transaction and dead wood.

    In this kind of market, it’s a critical mistake. The homes that are selling are doing so much faster than last year. Meanwhile, the ones that are not selling are sitting for much longer.

    Mispricing is probably the worst mistake you can make.

  17. Osman is right on. Last year this was not as critical. This year, and this time of year, there is little room for mistakes and you must evaluate your position every 10 days, not every 30.

    In January you can make every mistake in the book and use the 30, 60 90 rule and be right on by April 15 and do just fine. In August, you have to be right on by September 15th, so fine tuning must be much quicker at 10/10/10 vs. 30/60/90 with “out the gate” being more critical.

    I’m talking shorthand here…hopefully most of you are tuned into “Ardellspeak”. Feel free to ask questions if I’m going too fast.

  18. Thogek,

    The “whole effort over producing a tool like this” is about getting seller leads to the table vs. same old, same old, buyer leads. People look at the value of homes are OWNERS, most internet buyer leads, heretofore, or at least a higher percentage of them, are first time buyer, currently renters.

    “King of the Hill” will be the one that generates the most bona fide seller leads, and you don’t get them by offering “A free market analysis of what your home is worth”, as all too many agents know.

  19. No argument from me about the fact that a $20k/6% difference can be critical to the business of selling a house. But (IMO) anyone depending on a public automated tool such as Zillow for anything much more accurate than that deserves what they get. 6% variation is pretty good for an automated online tool like that (as they go), but a more consistant and precise valuation requires human attention, which Zillow’s otherwise very interesting tool cannot provide.

    Of course, this doesn’t help the professional whose client thinks Zestimates are gospelish, etc…

    My only point is that Zillow does pretty well for what it is, but that there seems to be a lot of mixed [mis]perception over what it actually is (perhaps not helped by Zillow’s own fanfare). My musings above were toward what it would take to clarify this inherent inaccuracy to Zillow’s users without necessarily killing the tools interesting consumer appeal. It is clearly not a professional property valuation service, and should not be used as such by anyone who cares about the success of buying/selling a home. Sorry if I was less than clear about that point.

  20. My experience has been that no one takes a Zillow Zestimate as “gospelish” and no one takes the agent’s CMA value as “gospelish” either.

    Whether or not Zillow is accurate has to do with the “subject property” and the variance of value at Zillow can be the same if three professionals come in with three different values, as is often the case.

    To say that “professionals” are always “spot on”, or are always accurate and more accurate than Zillow is not realistic either. Never say never or always πŸ™‚

  21. My experience has been that no one takes a Zillow Zestimate as “gospelish” and no one takes the agent’s CMA value as “gospelish” either.

    Whether or not Zillow is accurate has to do with the “subject property” and the variance of value at Zillow can be the same if three professionals come in with three different values, as is often the case.

    To say that “professionals” are always “spot on”, or are always accurate and more accurate than Zillow is not realistic either. Never say never or always πŸ™‚

  22. In my opinion, the Zillow Zestimates are seriously flawed. Since Zillow doesn’t reveal the algorithm for open scrutiny, we can’t evaluate where the flaws are kicking in. That, in itself, is cause for concern.

    We’ve been following this over at the UCR Sloan Center for Internet Retailing, and so far, Zillow doesn’t have a very good answer to the discrepancies.

    See for yourself: http://sloan.ucr.edu/2008/03/09/why-are-zillow-zestimates-so-wrong/

    Professor Donna L. Hoffman
    UCR Sloan Center for Internet Retailing

  23. Q-Diddy,

    That answer is often more property specific. It depends on how many like kind sales there are in a given area.

    In a housing complex where five sales of the same model have sold within a six month timeframe, there would be very little difference. The agent version would have more of a “leaning”. If the agent represents the seller, they would try to add value based on supply and demand and increase condition factors to support a higher than last one price. If the agent represents the buyer, they would try to hold closer to the comps and work more off of a “price per square foot” rationale.

    Zillow gives a range rather than tilt it toward “Who’s Asking?”. Appraisers will lean toward who hired them and why. Estate Valuation? Divorce Valuation? Refi? Purchase Loan? On a Purchase Loan often the downpayment will be a factor and an Appraiser, whose primary job is to protect the lender, will tilt differently at zero down than at 50% down, as the lender’s exposure is the primary factor.

Leave a Reply