How Brokers Lose Home Search (Again & Permanently)

knockoutBrokers made two tremendous strategic errors in the prior two decades: 1) not taking control of the internet business channel in 1996 when realtor.com was launched, and 2) losing home search in 2006 when Zillow was launched. Their current strategic error will cause them to completely and permanently lose home search by the end of 2017.

Brokers will lose control of internet home search as the four major portals (Zillow/Trulia, Realtor, and Redfin) attack home search with Artificial Intelligence (AI) based tools, pulling their combined market power to 90% by the end of 2017.

The top four real estate sites currently own about 44% of the total market share for real estate search, but that is the wrong way to perceive their prominence. Once you drop below the top ten ranked real estate sites the market share of the remaining thousands of sites rapidly drops from 1% to .0001%. Essentially these other sites don’t exist; they are irrelevant; they possess no market power. When you consider who shapes the industry the top four competitors control 90% of the power affecting the industry.

All of the leaders, Trulia, Realtor, and Redfin are launching or have already launched AI-based home search. Zillow has not formally announced their tool but that’s part of the sneak attack, as they already have access to Trulia’s AI work and they are hard-core technologists themselves. Their silence on this issue should be a deafening alarm to the industry. These will be the only players of consequence for internet real estate search by the end of 2017. (Reading the linked data above for each company will be an eye-opening experience and well worth your time.)

Unless a large franchise has secretly spent the last five years and tens of millions of dollars in R&D, and has already created an AI based home search tool–and it is on the launch pad ready to go today, their market status for internet home search is irrelevant and irretrievable. Not having done this is the brokers’ third strategic error and a fatal one.

In his book “The Innovator’s Dilemma” Clayton Christensen discusses in proven detail how several industry’s dominant firms focused on sustaining technologies and did not invest in disruptive technologies. Think of a sustaining technology as any improvements introduced which perpetuates the current market dynamics: the market leaders remain the leaders.

In the real estate industry’s case the franchises and brokerages invested in the sustaining technologies represented by bigger-prettier-faster websites, CRM tools, electronic contracts, faster response systems, communications software, team collaboration software, and various hardware improvements. These keep the broker-centric model in place and crawling along but never advancing.

Disrupters on the other hand attack the market from below—striking with new technologies never seen by the marketplace. Our primary example is Zillow’s Automated Valuation Model. Although every real estate website on the planet has at least as good of quality of home data as Zillow does, they were still destroyed as the market ran to the disruptive AVM technology they could get could not get elsewhere.

Disruption does not attack by making incremental improvements to existing technologies and systems. It attacks by changing the rules of how the marketplace is served. Disruptive innovation is much like a shark: coming up from the darkness below and hitting the incumbent with such force that even if they survive they are maimed for life—never as strong again.

This is now happening with AI based home search.

Home searches on the internet began around 1996 when realtor.com first appeared, and were additionally promulgated through a multitude of broker-based sites with the advent of IDX (broker reciprocity) in 2000. These market offerings were fundamentally watered-down versions of the MLS searches brokers used based on location, a price range, bedroom and bathroom count, and home size. These search features set a technological standard twenty years ago and are still used by 99% of real estate sites today.

Other than a few incremental sustaining improvements no franchise, brokerage, or the National Association of Realtors has advanced home search beyond this antiquated, programmer-based search criteria in twenty years. Every brokerage and franchise and broker IDX site is based on this ancient search technology which will be destroyed within two years.

Within Christensen’s book he demonstrates how incumbents vying with each other using sustaining technologies never substantially gain or keep a strategic advantage over one another. The combined efforts of franchises perfectly fit his description. Yet when an outsider introduces a disruptive technology the strategic advantage is so great that no existing market leader ever catches up with them in market demand. Again consider how Zillow’s AVM introduction has allowed them to hold captive internet use for an entire decade with no franchise even coming close to catching up with them.

This is now happening with home search. With the years of research required to develop such sophisticated software, along with the tens of millions of dollars investment to pay for this—all topped off by the patent protection created, no one not already in this game can win.

So what is AI based search? For the real estate industry it is essentially recommendation engines based on machine learning (AI).

Recommendation engines for real estate buyers is very tricky stuff—as you really can’t ask a home buyer what they want. Due to a problem commonly called survey bias you will never get an accurate self-assessment of what a home buyer truly desires. You’ve seen this yourself where a buyer describes what they want in your first meeting only to buy something substantially different six months later.

So you really are left with tracking a person’s actions and then inferring intent based on your algorithm’s analysis. Essentially you are trying to interpret what the buyer likes based on where they go and what they do on your website. You also, over time, have the advantage of “seeing” what similar buyers did and liked—creating a “standard” for this “type” of buyer. It’s labor intensive—at least for your computer. The technology world likes to call this kind of research “big data”. Again, consider following the links above, particularly for Trulia and Zillow and you will barely begin to understand the complexity of AI search.

Imagine the AI software noting your affinity for pictures of Victorian homes—or at the opposite end of the spectrum industrial lofts. Further the system tracks locations you like to pull these homes from and captures the price range you spend the most time at. The software perpetually correlates what you appear to like to locations and price ranges you search, proximity to schools, parks, restaurants, cycling trails, transportation, and every amenity each home has as captured by the software from the description brokers provided on each house.

The AI system does what a buyer agent does over a period of months as they show buyers dozens of homes. Through much time spent you note what really excites the buyers regardless of what they originally told you they wanted in a home months before. Except that the AI system tracks perhaps a thousand potential criteria relative to you focusing on a half-dozen, and finds correlation as the buyers search hundreds of homes on the internet within just a couple of days. The AI system discovers what the buyers really like based on their actions in a couple of days and starts making recommendations quickly thereafter. It accelerates finding the home the buyers really like—all while they are playing on the internet site.

The various types of homes have commonalities—such as our Victorian being turn-of-the-century; renovated—but not remodeled; along with key attributes like 10-inch based moldings, period fixtures, hand-carved wood finishes—and the like. The AI system notes, remembers, and searches for other homes meeting these criteria within acceptable tolerances and makes recommendation to the buyer.

The system might note the industrial loft buyer certainly spending a lot of time in hard-core urban areas, and key terms like concrete floors, steel beams, 20-foot ceilings, underground parking, and walking distance to museums and restaurants—and making recommendations of homes based on these criteria, and thousands of others it tracks.

Once a website offers even a non-perfect version of home search being able to learn from a user’s actions and offer suggestions of homes they may like, you have touched on a series of benefits no search based on price range and bedroom count can ever match.

This strategic advantage, even with its shortcomings while it is being refined, will destroy home search at all sites using the current price-range and bedroom count search criteria, and you simply cannot compete with it. The top four sites will use AI search to sustain and grow their 90% market power, with the buyers and your advertising dollars all converging there.

Leave a Reply

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