A couple quick quetions I have about Zillow's claim is 1) the sample size for their survey and 2) when those samples are selected and compared - meaning Zillow said that they used their historical estimate to compare, what date did they pick as historical and how many houses did they count for each area? We all know the more recent sample the more accurate the estimates are.
My clients were just making comments about how the propertieson their block of four homes (all very similar, not unique homes) are differnt in prices from $630Ks to $1.4M+. (said Zestimate was $1.8M and sold for $1.4M).
Of course, the $1.4M property is a pure error in this case, human or computer, probably both.
I do agree though that as time goes on and more sales are on board, the data may be more accurate - only if Zillow bothers to program their data to include and consider all the manually updated details, such as granite counters, limestone tiles, SS appliances, brand of appliances, ..etc.
Because unless there is a huge amount of turn over every year in a certain area so Zillow can account for , without human intervention - Realtors coming in and actually evaluate and compare the houses - the data will soon be outdated.
Real estate appraisals are as good as the people who make them. It's not unlike weather forecasts. Sometimes you are on and sometimes you are off.
Many people poo-poo Zillow because they have no clue of what Zillow does. Zillow uses algorithms to predict resale value based on historical data. The data they use isn't complete and nor is it 100% accurate. But they give you a range for the neighborhood based on what has sold in the past. They obviously don't go into evey single comp and check out what type of renovation were made.
I am willing to bet that many appraisers and many real estate agents do not do that also. Zillow is just another piece of information that prospective buyers and sellers should consider to make up their minds about the value of a house.
I wish that we could get all the appraisers of this world tell us about their record in terms of their ability to predict accurately the sale price of the value that they appraised. I would be willing to bet that they too would find sometimes quite a bit of variance between what they predicted and what happened. All I can tell you is that as a former scientist, I am only capable of accurately predicting the past.
As far as making an offer, nothing ties you solely to what a licensed appraiser or Zillow or your real estate agent tell you. Your offer should take into account your desire to get that home, your ability to purchase it and the level of motivation of the sellers. Those cannot be measured accurately and this is why most predictions are off target.
You'll see that in most regions, their median error is about 10 percent. In some regions (San Jose, Seattle) they do quite a bit better, 6 or 7 percent.
I would be quite curious to see what the median error would be for a typical real estate professional touring a house and trying to estimate what it would sell for. It is hard to imagine that many real estate professionals could do better than 5 percent. There is just too much randomness in the game (which buyers happen to show up, how the seller behaves emotionally, what the agents recommend listing for, etc.)
Are there any real estate agents out there who think they can beat Zillow and get within 5 percent more than half the time? You can test yourself on this, you know. Tour and write down your best guess for twenty homes listed by your agency by agents other than yourself (do this before you know the listing price) and see how well you do. You might be surprised to discover how many agents underperform Zillow (or not... I've never done this experiment...)
Although its usefulness for individual homes is debatable, I think Zillow is EXTREMELY useful for answering broad questions (Is one neighborhood pricier than other? Have prices in a town gone up or down this year? Which areas have the biggest houses? Which have the highest price per square foot?) that are sometimes tough to answer by eyeballing a smattering of dissimilar comparables. As Michel says below, it really is a remarkable tool.
I believe Zestimates can be used a data points, but they are most valualbe when supported by MLS comps sales, etc.
Hope this helped and good luck!
I've looked at a number of properties, then looked at the Zestimates and found the numbers far off. The same is true for the other online automated services. In fact, I just wrote a newspaper column about inaccurate ways to attempt to determine a home's worth. It hasn't been published yet, or I'd link to it. But here's what I wrote:
The first strategy which can produce misleading or inaccurate results is to rely on online, automated web sites. There are a number of them, including Zillow, RealEstateABC, and Eppraisal.com. A consumer logs onto the site, enters a property address, and in a few seconds is presented with an estimate of the propertyâ€™s current value. Often thereâ€™s other data as well, including price trends for surrounding areas, a list of â€œcomparableâ€ properties, and more. The sites can be fun to use, and they often have good information about the home buying or selling process.
The sites calculate values by attempting to identify comparable propertiesâ€”nearby similar-sized homes, often built around the same time, that have recently sold. All this information is run through proprietary formulas to produce a suggested value. However, those suggested values can vary greatlyâ€”both above and below the actual value. For example, I selected a recent home sale in Zip Code 22030 close to the NVAR average sales price of $621,570. It had taken 76 days to sell, right in line with the average days on market of 91.
In this case, the property was a 4 bedroom, 2.5 bath home built in 1986 that had sold for $605,000 in mid-June. RealEstateABC estimated its value at $539,000. Eppraisal estimated it at $558,947. Bank of Americaâ€™s service estimated it at $560,045. And Zillow estimated it at $783,000.
A home seller relying on those online estimates might have priced the home as much as $66,000 under its fair market value. A home buyer relying on those figures might have overpaid by $178,000. Thatâ€™s just one property, and certainly not a scientific sample. Sometimes the online calculations are more accurate. Sometimes theyâ€™re less accurate. The point is: A seller canâ€™t rely on those services for an accurate valuation.
Zillow does publish data indicating how accurate/inaccurate its numbers are. Even their published numbers generally indicate an unacceptably wide range, and I wonder (but can't outright dispute) their calculations regarding their accuracy.
Hope that helps.
Zillow pretty much takes the information on the tax record and does estimate accordingly. Since a lot of information are old and, as you said, they (the computer) only has Numbers (# bedrooms, bathrooms, sf, lot size, year build, etc), it's really for the computer to come up with a number to represent the real value of the home.
I just did one for my own neighborhood (I always do this to see how far off or close Zillow is), and so far, it's still very far off from real value. It does allow you to go in and manually change the assumptions (such as major kitchen remodel, addition, etc), but does not account for the view, specific location (right on the golf course, etc) and everything that need human interpretation.
Yes, do adjust the price according to how nice the property is compare to others. Part of how we, as Realtors, learn and estimate a property.
Sylvia . .
If any of us could consistently and accurately pinpoint predicted resale home values we would be in Las Vegas and making a heck of a lot more than most your average appraiser or REALTOR. My only point is that when making a critical financial decision like how much to pay for your home, it behooves you to have the opinion of a qualified professional in your corner.
In your example above, Zillow most likely has no idea the home in question is "nicer inside and has been taken care of well". It may very well makes sense to pay more than what the Zestimate says. Only a real estate professional, familiar with the area, can tell you for sure.
Who knows, maybe Zestimates are good in Sarasota, but I know in my market, they are more like a weather forecast than an appraisal.
Zillow is a useful tool.....................however, it is critical to know the personality of your target area. Zestimates have been known to be way off. It helps to know the area so that bad comps can be cast off and compensations made.
Yes....be on guard.
Never base any pricing decisions on Zillow's Zestimates. I am a Certified Residential Real Estate Appraiser as well as a real estate agent. I have personally appraised over 1,500 properties in Massachusetts. DO NOT rely on Zestimates for accurate market values. They are notoriously off target. It is impossible for a computer algorithm to accurately take into account all the intricacies of a particular market. Zestimates should come with a disclaimer "For Entertainment Purposes Only" because I agree, they are fun to check out. Consult with a local real estate agent, or better yet a local real estate appraiser, if you are interested in determining a TRUE market value. ESPECIALLY if you are considering buying this property!
Hope this help.