Everybody is Worried About the Listing Data; but it’s the Sold Data That’s Going to Kill You

jingaWhile franchises, brokerages, and agents fume and plot how to regain control of the listing data, Zillow (ZG), Realtor, CoreLogic, and others are sucking down the MLS style sold data and that’s what is going to kill you.

Think back to the industry’s golden age of full monopoly status on the MLS systems and all active and sold home information.  Aside from a few for-sale-by-owner sellers everyone had to hire a broker to get their home valued, “listed” in the MLS, or homes searched on behalf of a buyer.  The MLS along with some classified ads and an open house was the entire advertising platform for the real estate industry.

Then one day the industry launched “reciprocity” and IDX feeds.  A few brokers ran with this opportunity to create new business models; other brokerages attacked and boycotted them, MLS’s cut data to them, and the USDOJ eventually stepped in, sued NAR for antitrust, and we ended up with the VOW agreement and 2 million IDX sites which all look and function the same.  Whew, innovation averted by the industry.

But the public display of previously secret MLS information pulled a couple of supports right out from the very bottom of the broker model—but nobody really noticed.  It seemed that as long as all 2 million brokers had the same boring IDX websites that no one could gain an advantage in the marketplace, that everything would continue as it always had, and most importantly brokers could move their heads from their business cards and signs over to their websites.

However outside of the industry along came Zillow and Trulia—and Realtor, but NAR made sure Realtor didn’t do anything to please the marketplace.

Zillow and Trulia snagged a bunch of county record data, and a bunch of demographic data, and crime data, and value trend data, and anything else they felt would attract the eyes of the marketplace, and jammed this all up on public access websites.  Oh no, a twinge of innovation in the marketplace.

Then the biggie: Zillow launched home value estimates and the ability to stare at your neighbors’ values and backyards and rooftops.  What fun, and the best part was that the marketplace could do it all without a broker.  The marketplace loved the Zestimate but brokers hated it, arguing it was too inaccurate.

The brokerage industry is now running scared because the four main portals capture about fifty-percent of all real estate related internet traffic, with Zillow Group owning most of that.

ZG is better at marketing the industry’s own information than brokers are.  Zillow loves snagging and publishing active listing data and loves the profits they make through advertising fees paid by brokers for a shot at the business they weren’t skilled enough to attract on their own.  But published listing data is not your real problem.

The fees earned through advertising paid on actively for-sale homes is simply the short-term cash flow to keep the ZG machine running in order to capture the full transaction processes of selling and buying real estate.

You see ZG has survived off dusty old county record data for a decade.  But with each property displayed at their sites and eventually sold, they have slowly built up a national database of rich sales information, and this is accelerating exponentially today.

The grungy old county records data barely offered them a style, size, bed and bath count, lot size, and some government conjured tax valuation for creating their Zestimate.  But oh that rich MLS style data with text descriptions, and pictures, and lists of inclusions which comes as each active property becomes a sold property—that’s gold or presently will be.

The Zestimate will soon integrate all of the luxurious text descriptions and inclusions from all of those sold properties into its valuation model–another huge chunk removed from the broker model’s foundation.  The other companies with access to this kind of data will do the same.

From analyzing all of that sold data the ZG sites will shortly integrate key word matching algorithms into their buyer search tools—another chunk out of the broker model’s foundation.  Oh thank you for that rich sold data, mmm.  Buyers will flock to ZG, thus more sellers will demand to be there, thus more brokers must advertise there—on and on.

Currently the Zestimate claims a median error rate of eight-percent, and brokers scream that’s not good enough.  It will get much more accurate with a decade of rich sold data—and you have another problem.

Brokers’ own estimates of value have a median error rate of five-percent of actual sale price.  I evaluated 4,802 recently sold properties from all areas and all price ranges from the MLS and guess what?  The original list price was within a median five-percent of actual sale price.  What, Zillow is within three points of brokers’ own estimate of values and they’re just now getting ready to use MLS style data to improve that!

Here are the stats: 4,802 MLS homes analyzed (all areas, all price ranges), compared original list price to actual sale price, and the median difference was five-percent.  The breakdown is: 609 properties sold for list price; 710 properties were undervalued by an average of five-percent; 3483 were overvalued by an average of eight-percent.

What, Zillow is not tremendously out of line with brokers for valuation accuracy?  I hear the arguments now.

  1. “You did the analysis wrong.” Maybe, I’m not a statistician, here’s the spreadsheet. All data came directly from MLS records downloaded on 4/19/2015. I invite you to investigate it. I further invite you to pull a few thousand records yourself, or have your MLS do it. Pull all price ranges and locations of both attached and detached residential properties and compare original list price to actual sold price.
  2. “Hey those aren’t the real broker estimates; the original list prices are what the sellers demanded!” You don’t want to go here. If that’s the case then sellers don’t need you or your estimates of value as apparently their homes will sell for market value whether they initially price them a bit high or low. That makes the Zestimate and no commission paid look really shiny to them.

Another massive piece of the broker model foundation was just removed, and it’s about to get worse.

Now that ZG is making a fortune on listing data, and now that they can improve their buyer search tools and value estimates with a decade of accumulated MLS style sale data, they can turn their efforts toward controlling the entire sale process.

Most likely every time a potential buyer touches a ZG site it’s tracked.  I am sure ZG is creating learning algorithms and artificial intelligence to understand what actions buyers take just before becoming serious.  ZG has all of the active properties, a decade of sold properties, and gets to witness every action a potential buyer takes.  They know who’s about to buy; what’s that worth to you?  Oh, it’s going to be more than what it costs to paste your picture next to somebody else’s’ listing; place your bid.

Every time a potential seller touches a ZG site it’s tracked.  You get the idea: ZG knows their address, they now have high quality sold data offering superior valuations, and can tell when a potential seller is about to get really serious and put their home on the market—want in on this action: place your bid.

Reiterate the same processes above using the “by owner” and the “make me move” people.  ZG can tell when these guys are ready to hire out some help; place your bid.

Renters…every move tracked…when they are about to get serious and become buyers ZG knows; place your bid.

Eventually you drop 40% of your commission on getting one of these hot leads.  Guess what, you’ll be using the ZG transaction platform software.  Now guess who’s being tracked and analyzed as part of this entire process.

A few years later buyers become sellers, ZG’s CRM software has kept them well informed of everything going on in their marketplace, and guess who’s once again being tracked and analyzed as a probable seller; place your bid.

Everyone demanded ZG’s data become more accurate.  You now have what you desired.

A kinder-gentler-edited version of this article hit Inman News  5/4/15, and wow did the marketplace react!  Check out the comments from the Inman readers if you get a chance.


Disclosure: Creed Smith created and owns the QValue AVM, which can value properties and search homes as described in this article; patent pending.  Creed Smith has no affiliation with ZG or any other company mentioned in this article as of the publication date of this article.


  1. says

    Appreciating the time and energy you put into your blog and detailed information you provide.
    It’s good to come across a blog every once in a
    while that isn’t the same out of date rehashed material.
    Great read! I’ve saved your site and I’m adding your RSS feeds to
    my Google account.

  2. says

    92 3-6-12 spune: In mod normal, daca era sursa, se restarta PC-ul asa ca ramane varianta drvierelor de la placa video, sau chiar placa in sine. +25V-a ajutat acest raspuns?

Leave a Reply

Your email address will not be published. Required fields are marked *