A Demon article republished with permission from Inman News
After News Corp. took control of realtor.com and ListHub, players in the real estate industry started to wonder what would happen next. When Zillow and ListHub failed to renew their listing agreement, things really started to heat up. And now that Zillow and Trulia are merging, the battle is really going to get interesting.
News Corp. has structural advantages over Zillow through receiving nearly 100 percent of multiple listing service (MLS) data directly, and with internationally known media channels that they can use to disseminate real estate information and bias people’s attitudes toward a preferred source.
What it now must offer are two website tools built around the greatest desires of home sellers and buyers:
- For sellers, an automated valuation model (AVM) that actually displays what creates differences in home value.
- For buyers, a search tool based on what they truly desire in a home.
No one has offered an innovative valuation model since Zillow hit the market a decade ago, and no one has ever offered a search innovation. Whoever gets these two tools to market first takes a commanding lead in the portal war.
Offering the marketplace three avidly desired tools launched Zillow into existence, and everyone else’s lack of innovation allowed them to dominate the real estate portal business for a decade.
Two of those three advantages have now been matched within the marketplace. Complete and lengthy details about homes are commonplace (perhaps even in the overkill stage at this point). So are map-based searches and geographic displays of information. What used to be cutting-edge technology that initially dazzled the marketplace is now fully expected and no longer gives anyone a competitive advantage.
This erosion of a competitive advantage happens in every market and is visually represented in the Kano Model of Quality below. Essentially, Kano’s model demonstrates that what was once a leading-edge technology or service or other advantage loses its impact as competitors match or exceed each innovation over time. There’s a slightly more in-depth explanation of Kano here at the Demon site.
News Corp. sees the need for new tools on realtor.com, along with the Australian and Indian portals, as demonstrated in the company’s interest in artificial intelligence (AI) software to assist in seller and buyer decision-making, such as the Big-Decisions software package. But the AI that News Corp. is investing in is “rationally based” — comparing financial options and outcomes (for example, rent-vs.-buy financial comparisons). And while every one of us perceives ourselves as smart, rational decision-makers, we actually make decisions based on envisioned life events and the immediate perception of emotional rewards.
Why humans make emotionally based subconscious decisions is outlined by Daniel Kahneman’s research as explained in his book, “Thinking Fast and Slow.” A less in-depth and easier read on the subject can be found in Malcolm Gladwell’s book, “Blink.” I think the part about making decisions based on envisioned future life events while immediately receiving the emotional reward is the key. So let me touch on the concept of an AVM based on emotionally driven decisions.
No one has ever bought a three-bedroom, two-bathroom house. Those may have been physical requisites, but they do not represent how the buyer chose or valued one house over the dozens viewed. More than likely, all of the homes they looked at met the three-bed, two-bath requirements. They chose and appreciated one home over the others based on key attributes of the home’s flow, form and finish. The home they loved had aspects that triggered visions of future life events. These envisioned events created emotional responses. This flood of emotional responses caused one home to stand out from the others; these emotions caused one home to be chosen and valued more over all the others.
Where News Corp. can maim Zillow is by offering a superior AVM. On the other hand, it’s where Zillow can capitalize on its market momentum and cut News Corp.’s Achilles tendon by offering an AVM superior to the current Zestimate. Whoever wins this battle captures first-mover advantage and public adoration: the hearts and minds of the marketplace.
A superior AVM will capture and utilize the key attributes of the homes that equate to these emotional responses — how buyers choose and value a home. I call the aspects of homes that cause these reactions “emotional triggers” — attributes weighted in importance based on the emotional impact and how they correlate to a sale price: value.
Nobody is currently offering an AVM like this because it is very difficult to create. And using demographic data on age, income, crime rate or distance to parks is not the same thing. Buyers first choose and value a home based on its emotional triggers; then they will rationalize their decision using demographic data or using the purely financial data from a company like Big-Decisions.
Once you have this AVM, just reverse it to become an incredible buyer search tool. This is the second tool News Corp. could use to decimate Zillow, or where Zillow can slam realtor.com dead in its tracks.
If buyers do not choose or value a home based on square feet and bedroom or bathroom count, why does every real estate portal and website on the planet force them to search this way? Because it is easy to do from a programming standpoint: It’s precise, the data is readily available and it’s rational. But it’s not how buyers choose a home.
The ideal search tool allows buyers to spot any home they think is fantastic, then click one button and find every other home available with similar emotional triggers — and this is how they really choose and value a home.
It’s the same AVM simply working in reverse.
Is it radical innovation? Well, remember: To meet our definition of radical innovation, we need a body.
If News Corp. launches the innovation bomb, Zulia survives but within 24 months is relegated to the No. 2 portal position. Additionally, homes.com is dead along with all franchise portals.
If “Zulia” launches the bomb, realtor.com is a very distant No. 2 and waning; homes.com and all the franchise portals are dead. Additionally, Zillow gains structural strength: the new AVM-based tools attract many more buyers and sellers; the mass influx of users means all brokers and MLSs want their properties on Zillow. Zillow gets all of the direct feeds it desires.
If, miraculously, an insider, such as Realogy, launches the bomb, all other franchise sites are dead, as is homes.com. And now we have a three-way battle for top portal rankings. More importantly, this would give Realogy franchises a powerful market advantage, causing all other franchises to lose leads and sales to them.
How does this affect you? Why should you spend any valuable time fantasizing about the perfect AVM or home search tool?
Because this will probably be the first innovation bomb to hit the real estate industry, with either Zillow or News Corp. — or, at a long shot, Realogy — launching it.
Whoever does launches the bomb will set an entirely new standard for how properties are valued and searched for, and this will define where many of your Internet leads will come from in the future. Whoever creates these tools will control the real estate Internet. This is an industry-changing event that will direct how the entire real estate business functions moving forward. And it will be a life-changing event for you.
I think that sometimes the questions and my responses are the most fun part. Here they are if you are interested.
Q: My buyer had an “emotional trigger”. Unfortunately, the appraiser said it didn’t really matter and the home won’t appraise. What now Demon?
D: I hear you. I mentioned somewhere up above a potential new article that I already submitted to Inman that talks about why current AVMs fail (based on all physical criteria–from the appraisal model legacy), how to make a better AVM, then I kind of attack why the appraisal model is not really accurate. I am not sure that article will run as Inman is thinking I probably shouldn’t be writing under my pen name. So we’ll see.
Here’s a link showing why appraisers should not automatically adjust value based on their eight major physical criteria (square feet, beds, baths, basement, basement % finish, lot size, garage spaces, YOC. Here you can see that none of these criteria consistently correlate to value. It varies by area and price range–all based on the desires of the marketplace. http://qvalue.net/appraisals-fail-in-avm-use/
hat this means is that appraiser should not automatically adjust values based on differences between homes if the area has a low or no correlation factor.
I know, it’s too late on your deal. And I know that if you gave the appraiser a multiple regression report for a Demon, well it may have not gone over well. But everything in this market is changing. And technology, intelligently designed around how things really work, not what is easiest to count, will soon win out.
Q: The bigger threat to third party real estate portal survivability is the upper limit of broker/agent marketing spend on these sites, which depends heavily on assumptions derived from eyeball counts and projected revenue enhancements to appease certain institutional investors’ positions, decisions and expectations regarding their respective buy-hold-sell metrics.
For instance, dataquick, eppraisal, zestimate AVMs vary wildly for just about any subject property, rarely hinting at a consensus of average fair market value. Realist and RPR AVMs differ significantly in many instances as well.
The local real estate professional understands nuance best: juggling equity participation, appraisal curves, and current market dynamics in play at any given moment.
The dance between the rational and the intuitive means real estate professional had better love the amusement park for all its roller coaster rides, carousels, ferris wheels, funhouse mirrors, etc. while balancing dreams and market realities.
I am just not feeling a magic algorithm is even conceivable in this ether.
D: Hi Richard you are so correct on the variances between AVM products. If you look at a map view of the comps chosen by most you can quickly discern they simply used some kind of square foot range and bed and bath count, and then tried to keep some semblance of acceptability for distance. You and I know that 200 feet can make a difference between a $200K neighborhood and a gated $2M neighborhood. This is just the beginning of current AVM failures. I may have another article at Inman within the next 10 days with further details on why AVMs fail. You might find it interesting. I say “may” as right now Inman staff is deciding whether I can continue to write for them under my pen name.
And you are just so correct that the “magic” algorithm is perhaps inconceivable right now. I hope to show you something on that in perhaps another month or two.
I love discussing the hard stuff, so if I am still around here always throw me a heater. Thanks
Q: Great Post. gree 100% that the best AVM model going forward will make that portal the clear-cut, long-term favorite for consumer eyeballs. So then the question becomes (as you pointed out), how to create the best AVM that takes into account that emotional connection that Buyers have with properties.
I think one way to do this is to start collecting OFFERS data on comps. Offers data, while still not as early of a data point as what someone might be willing to pay based on just seeing a home, is still earlier in the process than sold data, and gives a very clear, black and white valuation as to what someone is willing to pay based on their emotional connection (and other factors) to a home. It’s also significantly more real-time than sold data (sold data is simply not timely and comes way too late in the process). If you could take other properties that had similar characteristics of the target property, and use the current Offers data on those other properties to determine the AVM, that would be much more accurate and a better representation of what the general population is thinking of regarding valuation (I understand at the end of the day that the sold price is the only one that matters because there is just one winner for each home sold, but if you could provide a min, a max, and a median dollar amount for Offers comps, that would be huge).
Shameless plug time — this is one of the reasons why we have the Offers feature within Reesio — we feel there is a HUGE opportunity to be able to capture Offers data and use that data (anonymously) for better comps and AVM’s, which would help all parties involved in a real estate transaction.
Btw, love the Demon Of Marketing avatar. Worth reading this post for that alone
D: Thanks Mark, you and I could fill up many pages on this discussion. I am unaware of any database where seeing and utilizing active offer prices is available; you have that? It would seem quite secretive and controlled–at least where I do business. But in that vein, here are a couple of scary Demon notes:
At my MLS asking price is within 5% of selling price about 97% of the time. That’s a strong predictor.
Current AVM’s using only physical data give you the base value; the level of finish relative to market desires creates the potential range (yea, in the article). But by taking the key attributes and running them through random-forest-regression tells you what the market desired and the impact of each on value. Thus you can solve for the price range of each home, albeit not in “real time”.
AVM’s and search related tools are the new arsenal for this battle. Watch for the mushroom clouds.
Q: Great article, sir. One point I don’t necessarily agree with..being first is not the end all be all. Lots of companies have been first but were beat out quickly when the serious players jumped in. Google wasn’t first. Neither was Amazon. First can give you a jump but it won’t win the race. Realtor.com may win the widget race but it still hasn’t figured out the “what the public wants” part. Remember when Blockbuster followed the Netflix model? It was too late. Damage had been done and too much $ had been lost in other ineffective areas…so they went belly up.
Love your insight, keep up the great work.
D: I fully–very fully agree.—on all of your points. Here I was referring to “first-mover” of two, perhaps three (if you include the entire Realogy group) mega companies. This whole portal thing may be in its “maturity” stage meaning new entrants will be few and at a tremendous disadvantage. There are perhaps 20-30 “real estate portals”, but most are quite tiny relative to our big two. I think ZG gets perhaps 28% of traffic, realtor.com 8-9%. I’d have to check in on Homes.com. But after that it gets pitiful. The largest five franchise portals combined are under three percent. So the first-mover strategy between these two monsters could kill everything under them. Then you’ll have the patent war to beat back any serious competition for some time. It would be a big advantage.
The blockbuster/netfilx comparison, I think, be less relevant to portals and far more relevant to the current broker model of listing/selling real estate.
And yea, realtor.com is still suffering from the NAR disease. It was never built to serve the needs of the market, just the needs of NAR. Unless News Corp gets serious and competes at the level I suggest, that image may keep them stagnant.
Thanks for reading and thinking enough of my stuff to comment. I love what is going on in the industry and think these will be the most exciting five years in its history.
Q: Great insight and great thinking! The dynamics of your proposition are obvious. What we do not know are the potential developments in virtual reality that could be implemented and become even more impacting than the references you indicate. Is there greater market power in “selling the sizzle and not the steak?”
D: Hi Larry, thanks for the feedback and kind remarks. You comment has me thinking about some various books where deep-thinking requires trying to determine the “unknown unknowns”–we not only don’t know everything that may happen but can hardly conceive beyond that. Here the “sizzle” in envision for an AVM is two-fold. Right now an AVM coughs up a dollar figure, then to meet a statistical “confidence” level has to offer a value spread with a low and high number. The machine and its inventors do not know why, just that statistically it is required. That spread are the “qualities” of the home the market most desired–the form, flow, and level of finish. These trigger the viewer to envision future life events. The secret is to capture these triggers and use them within the valuation model. As Richard above noted: nearly an inconceivable mechanism. I think that I can disprove that thought.
Finally the sizzle with such an AVM can then demonstrate to sellers exactly why one home was more valued than others. For a home seller the AVM gives a number but allows them to literally see the differences between values in a concise list (remodeled bath, etc). This leads to true marketing–letting the seller determine how they wish to prepare their home to meet the desires of the marketplace better than their competition.
This could go on for a really long time with some very complex discussions of psychology, buyer behavior, and programming. So I will stop here, but will ask you to keep an open mind about what will happen sooner than most folks can envision. Thanks
Q: I focus my business on unique properties often in the high end part of my market, and having this understanding of the consumer is what has set me apart from my competitors. I am not afraid to use my instinct to price and market properties that don’t fit the mold of an AVM. Sadly, appraisers are using more and more AVM’s and are producing highly inaccurate appraisals that are in stark contrast to actual market conditions. I completely agree that if a new AVM was created as you suggest, it would be the next game changer.
D: Hi Steve, thanks for the feedback. I actually have what I think is a very cool Inman article forthcoming that not only discusses why current AVMs fail, and what new AVMs must do, but even how the entire appraisal method fails. An appraisal is based 99% of physical criteria, with specific dollar adjustments made for square feet differences and different bedroom counts. However, if you run a multiple regression test on physically similar homes you find out that square feet, and bedroom and bathroom counts, and basement size and finish, and lot size and garage spaces and year of construction do not consistently correlate to value. It completely varies from location to location and property type based on what the market in that area most desired. That’s all in the article. So using “unusual” valuation methods is something you should keep doing.
I’m not sure that article will run as it sounds that Inman may no longer allow me write for them under my pen name of Demon. If it doesn’t I will be happy to get you a copy if you are interested.
Q: Wow. So simple. Glad to know your opinion isn’t biased against RDC or NAR… Well, actually, your words on your site state, “Your brand, your image, your stupid NAR and MLS rules are of no importance.
D: Hi Frank, I’m not particularly biased against any one group or company. What I am biased for is doing what is absolutely the best relative to the desires of the marketplace–the buyers and sellers. Whoever gives them the best I like; whoever tries to design something for themselves and not the marketplace…I just call it out. Thanks for reading.
Q: Another left field player is Facebook given their eyeballs, data and behavioral insights. Zillow is still the innovator in my book given consumer centric focus.
D: I agree. Yahoo! Homes could also make a move given their desire to catch back up to Google. But since their feeds come from Zillow that would create what they call “channel conflict”–taking business from those you are currently working with. Given the complexity of such a move they probably would not risk it.
Q: Great post. Don’t forget that the innovation may come out of left field. What company know what you want before you know what you want? GOOGLE.
D: Hi Vince: absolutely. In an article not on Inman I use Google as an exact potential candidate. I think that I say something like “if you sent the 100 best marketers and programmers at Google away for a three-day weekend, do you not think they could create something superior to the existing MLS systems?” My own retort was “that’s a silly question; they would create the superior system in one day and the other two would be a beer and pizza fest.”
Q: As an active broker here in Texas I don’t hear any of my clients talking about AVM and they all use different sites for one reason or another. They are simple and are using the sites for 2 reasons. To find something in their price range, basing that on what the house is listed for, and the second reason is to find something that looks good based on the pictures.
D: Sure, that is how it’s been done since public IDX fed sites hit the internet. Visualize those sites you just mentioned. Now visualize a site where a potential seller not only sees a value, but sees in plain text what created that value (remodeled kitchen or remodeled bath, etc). The seller can then get an idea of what really made the difference in value from one home to the next–and how to better prepare their home for market–and what to do if they would like a higher price by meeting the market’s desires. Which site will they like more?
The same for buyers. Instead of here’s a list of houses between $150K and $200K, good luck. They see a home they really like, click a button and get a list of other homes with the same kind of feature–emotionally based triggers. Which site do they like better?
Competition is always pushing innovation to better meet the needs of the marketplace. That’s why your smartphone does not weight three pounds and your car has an automatic transmission.ake other properties that had similar characteristics of the target property, and use the current Offers data on those other properties to determine the AVM, that would be much more accurate and a better representation of what the general population is thinking of regarding valuation (I understand at the end of the day that the sold price is the only one that matters because there is just one winner for each home sold, but if you could provide a min, a max, and a median dollar amount for Offers comps, that would be huge).