Dear Mr. Barton,

As you may or may not know, I emailed you guys a couple of weeks before you unveiled your product to suggest that you consult at least one real estate expert, before going public. I further suggested that since I have sold real estate in five states from coast to coast, that I might be able to help you tweak your product before its unveiling. I feel very badly that some are poking fun at your great real estate adventure, by coining the phrase “You’ve been ZILLOWED!”

Here are a couple of tips for you, (or for Dustin and Galen and Robbie) If you modify your application of data according to these guidelines, you will likely increase the reliability of your online Zestimate by as much as 50%.

Seattle area: Yes, you can value property fairly accurately using the tax data in the Seattle area. But the first step is to determine the appropriate factor. Many will value out at between 1.2 and 1.4 times the assessed value. Hot areas, like downtown Kirkland or parts of Queen Anne, etc will value at 1.5 to 1.6 times assessed value. Don’t take the comps out too far, keep your radius small. Stay as close to the subject property as possible and STOP when you have 5-8 comps after throwing out the High and the Low. DON’T average the sale price of the comps one to another to determine the value of the subject property. DON’T use price per square foot as a guide. Take each sale price and divide by THAT SAME PROPERTY’S assessed value to come up with the factor. If all of the properties in that neighborhood sold at 1.44 times assessed value, then your ZESTIMATE should be 1.44 times the assessed value of the subject property. You can average the factor, but not the price. Then use a range. Chuck the high and the low, the way I learned in grade school from the good Catholic sisters who taught me well.

Example: Data equals 1.8, 1.4, 1.42, 1.43, 1.44, 1.44, 1.45, 1.47, 1.1

Throw out 1.8 and 1.1. as the high is a massive remodel and the low is a fixer. Factor becomes 1.4357142. Assessed value of subject property is $313,000. Zestimate is $449,378.54 or between $438,200 and $460,110.

When inputting tax data, do not overlook the “effective year built”. Currently your program is not noticing that very important date, and reverting to the original year built, throwing the numbers way off on 80% remodels. You can use the 1.8 and 1.1 in the sample above by saying “Your home is valued at between $438,200 and $460,110. If you have just remodeled the interior, the price might be as high as $563,400 (1.8 X $313,000). If it is a fixer it may be as low as $344,300 (1.1 X $313,000).

For Seattle area, always use the assessed value of the subject property against the neighborhood factor.

Briefly, for Los Angeles beach areas: DO use price per square foot, as by and large that area does not have underground basements and the tax assessment increases to sale price every time a property sells (unlike Seattle and many other areas)

Florida: Do use price per square foot and keep the comps apples to apples. Watch the lakes. Price properties on lakes against other property on the lake and interior against interior. You are already OK in FL for the most part, so you can leave that alone.

PA, NJ and most of the Northeast of the country, keep the radius short and use price per square foot. Then find and apply the neighborhood factor and average the two answers.

Hope that helps you, Mr. Barton. Or maybe it will help Robbie and Galen come up with their own “Better than Zillow!”

Have a great sunny day in Seattle!

24 thoughts on “Dear Mr. Barton,

  1. I have a question I hoping some experts would comment for me:

    In the same condo complex,
    two condos of same size:
    Option A; one with view of alley.
    Option B; and one with excellent view of Space Needle/city.

    Which would rise faster in value?
    Generally, does condos with view rise faster in value, or the cheapest condo in the complex rises faster in value?

  2. View, except for the obstructable feature. For instance if you are in the xxxx building looking at the space needle, I am fairly certain that the new 35 story to be built on that side will obstruct the space needle view. So if you pay for the space needle view going in, you will lose that value going out. In that case the one without the view cleaned up and remodeled modestly will gain more value than the “view” condo, whose view was taken away by new construction.

    “Unobstructable” view will always rise faster than no view, but they are harder to come by and pricey from the get-go.

    But that’s a whole new post for another day…how to accurately value a view ๐Ÿ™‚

  3. thanks for the comment.

    just a bit scary considering Option A is : $382 per square feet, vs. option B (with view) is $550 per square feet?

    Is that a typical price difference for view vs. no view?

  4. New construction doesn’t value the same as resale. Kind of like buying a new car vs.buying a used car. The New Car has to hit the street for a month or more before its true value is known. Same regarding new construction, as it is purely a supply and demand market. The builder controls the price internally.

    You don’t have a Plan B with new construction the way you do with resale. If prices don’t go up in the area on a resale product, you can always paint it or remodel something and improve value without relying on area prices going up generally.

    You buy new housing for the same reason that you buy new cars…because you want new. In a flat market new will always go down. So the appreciation factor has to compensate. If resale goes up 8%, new will go up less than that, once the carpet is stained and the appliances and countertops are no longer the style of the day. You have the downward pressure of things getting old working against the upward pressure of the market in general.

    Anyway, to answer your question, hedge your bets and go in the middle. Find something in the building you like in the range of $466 a square foot, if that is an option.

  5. Hey Ardell….

    go onnnn with yo bad self!!!


    PS. I don’t know why these guys didn’t hire local advisory boards to help with the algorithm per city or per jurisdiction even.

    RE is and will always be local to some degree.

  6. Well, mostly, think about what will make you happy ๐Ÿ™‚ It’s your home first and an “investment” second. Make sure your payment leaves you enough money to include other things in your life. That’s the most important thing.

  7. That was a great post Ardell. Plus I picked up a nice nugget of information in the comment to boot!

    It will be interesting to watch for tweaks in the Zestimate formula in the coming weeks and months.

Leave a Reply