Watson, And the Future of Real Estate Technology

 

Say hello to Watson

Over on Facebook, in one of the real estate discussion groups I’m involved with, Loren Sanders posted the following:

I saw a documentary called “The Pit” about floor traders on the New York Board of Trade. It shows before and after they were bought out by an electronic trading company ICE. It is not an exact parallel to real estate but it shows clearly that progress waits for no one and no one (or company) is bigger than than market. That is my long winded lead in to: What are you doing now so your service will be relevant 2 years from now?

That sparked a bunch of discussion. The trailer for The Pit, by the way is here:

The movie is about the impact of technology on the Wall Street traders in the actual trading pits. Loren raises a really interesting question.

As it happens, I just contributed a chapter to Stefan Swanepoel’s upcoming report, speculating on the state of real estate in 2022. I’d call it science fiction, since ten years might as well be fifty years for trying to see where technology is headed. But I did find something while researching and thinking about that which is worth thinking about more.

Because of space limitations in print format, I couldn’t really get into all of the detail I’ve been thinking about. So I figured I’d think out loud here with you all.

A Little Bit About Watson

What most people know about Watson is that it played Jeopardy against human opponents. “Hey, how cute! A computer that plays Jeopardy!” is likely the extent of the average person’s reaction to Watson. But why would IBM spend tens if not hundreds of millions of dollars, and five to six years of time, using some of its best computer scientists and engineers… just to play a game show?

The answer is that Watson represents a breakthrough in how human beings interact with computers:

Watson’s ability to understand the meaning and context of human language, and rapidly process information to find precise answers to complex questions, holds enormous potential to transform how computers help people accomplish tasks in business and their personal lives. Watson will enable people to rapidly find specific answers to complex questions. The technology could be applied in areas such as healthcare, for accurately diagnosing patients, to improve online self-service help desks, to provide tourists and citizens with specific information regarding cities, prompt customer support via phone, and much more. [Emphasis mine]

So what is this Watson thing anyhow?

Well, it’s a supercomputer comprised of 90 servers with 2,880 processor cores, running IBM’s proprietary DeepQA software. That computing power is married to a very large and very fast storage system that holds the equivalent of a million books. And IBM claims to have spent years feeding mountains of information into Watson’s database.

Right now, you’re thinking — as I did — um, so what? It’s a big computer tied to a big database. Well, check out what the IBM people think Watson will be able to do.

So, Watson will be able to sift through thousands, millions of facts to help a doctor diagnose the specific symptoms of a specific person. It will be able to answer questions and provide confidence scores, and the reasons for why Watson believes that the answer is the right one. Watson will be able to digest, remember, and process more facts, more information, and more data than any human being possibly could.

Then there’s this:

“Watson would change the business model of how call centers are set up.”

The reason is that Watson is the first question answering computer. The current technology is limited to pointing humans to documents, websites, and places where the answer might reside. Think of Google. You go on Google and ask it, “What’s the best flatscreen TV for under $1000?” and you get a whole list of links back. You don’t actually get the answer from Google. Or Bing. Or any search engine, even those on a retailer’s website. You get a list of webpages, documents, or whatever with the right keywords and you get to figure out if that answers your question.

More often than not, those webpages and documents do not answer the question. Hence, the frustration of using Google to find a specific answer to a specific question, and the related human skill known as “Google-Fu”.

Let’s Think About Real Estate For A Moment Here

With Watson in mind, let’s think about real estate. What is it that a broker or agent actually does day to day? As I see it, most of what a realtor does is answer questions. Managing the transaction itself — meaning, filling out paperwork, and managing the workflow — is such a minor part of the practice of real estate that most successful agents have transaction assistants who would handle all or most of that work.

If a realtor is working with a seller, she has to answer questions on pricing, on market trends, buyer preferences, on what marketing strategies would be most effective, whether to drop the price and if so, by how much, etc. etc.

If she is working with a buyer, she might be asked about neighborhoods, schools, local market trends, and specific properties — “How big is the backyard on that colonial we looked at on Friday?” Based on what I heard from the Charleston HearItDirect event, some buyers really want to know about their potential neighbors: what are they like, will I fit in, is it a strong community or one where people don’t talk to their neighbors, etc. etc. One consumer actually said that the job of a realtor is to build communities, not just to sell houses. The Fair Housing Administration might have an opinion on that, but he has a point. Maybe she has to handle questions about what the HOA rules are, what the taxes are, about local zoning regulations, any environmental hazards, soil composition… who the hell knows the infinite variety of things that a buyer might want to know?

A major shift in the post-Internet era has been the ability of consumers to get the answers to some, though by no means all, of their questions from the Web. The “data is everywhere” meme may or may not hold true in the future, but there is no denying that more and more consumers come to their real estate agent having done all of their own research and homework.

And you know what? It is unclear to me whether consumers actually enjoy doing that research or not. Again, referring back to the Charleston HearItDirect event, more than one of the buyers mentioned how the search for the house became a part-time job. It became a hobby of sorts… albeit, not necessarily an enjoyable one.

The reason, I believe, is because current search technology does not provide answers; it only provides pointers to places where answers might be found. It is then up to the consumer, with his limited knowledge and ability to analyze the actual data (Zestimates, market trends, etc.) to figure out what all this data means. The smart consumer of today, like yours truly, just skips over all that frustration and hires a buyer agent to go do it all for him. Yay for realtors… right?

Enter Watson

As far as I can tell, here we have a computer system that is able to store, digest, analyze, and weigh terabytes upon terabytes of data. It is able to understand natural language queries, even through speech recognition, enough to run hundreds of parallel algorithms against the database, weigh the possible answers, reach a confidence score, and then provide the answer and the reasons why it believe that answer is correct. Watson is able to do this for doctors who need to quickly diagnose what is wrong with a particular human being with a particular set of symptoms given his age, diet, general health, previous medical history, etc. etc. and so forth.

Here we have a machine that major corporations will use to replace thousands of human call center operators, because customers can just call in, ask a question (“My TV is broken”), go through a set of back-and-forth with Watson, and then get an answer.

If a machine is able to provide answers as personal, as specific, as unique as what a single patient’s symptoms mean… is it really unthinkable that it could provide answers to what house is right for a particular family? Connect up your personal datasets — from Facebook, from Twitter, from your travel sites, from Spotify or Pandora, from your magazine subscriptions, etc. — and Watson can tailor the answer to your specific lifestyle. Is that impossible to imagine? I can imagine it.

If Watson works via natural language speech recognition, well enough to play Jeopardy, and major computer scientists are hard at work right now tweaking and improving that speech recognition capability, we are no longer talking about having some consumer go visit a website, type in queries, and such. We’re talking about calling up a phone number, and conversing with Watson and getting an answer. (By the way, Siri + Watson makes a whole lot of sense now, doesn’t it?)

From where I sit, so much of the value of having a realtor is replaced by something like Watson. Of course, there’s the irreplaceable human element — the emotional support, the counseling, and having someone to make sure the transaction is going to close — but so much of what a realtor does for the consumer is answering questions. Replace that, and one wonders what the value would be.

Which brings us full circle back to Loren Sanders’s question: What are you doing now so your service will be relevant 2 years from now?

I’m sure one of the answers someone will bring up is, “Pfah, technology, schmechnology — this here is a relationship business, son, which you would understand if you had ever gone belly to belly with a real buyer or seller.” Well, I got news for ya. I have great relationships with my personal realtors — Sue Adler in NJ, and Blayne Vackar in Texas. They’re both good friends of mine, and wonderful people, who also happen to be experts in the real estate game.

But if I can get truly accurate pricing, backed up with evidence and confidence scores, from Watson; if I can get all my questions about neighborhoods, local restaurants, zoning regulations, HOA rules, and so on and so forth answered by Watson; if I can have Watson pull together all of the available economic data and market data from throughout the country and in my local marketplace, weigh the evidence, analyze for reliability, and provide me the answer to “How’s the market doing?”… let’s just say that my good friends with whom I have strong relationships would need to adjust their expectations on what I’m willing to pay in commissions.

And I’m just about the most industry and realtor-friendly consumer you’re going to find.

To Make Thing Even More Interesting…

The first and most natural response to the above should be, “You’re crazy, Rob. A computer will never replace a human being, no matter how connected it is.” The more knowledgeable people — many who tend to read this blog — will point out all the various flaws of Watson. It doesn’t get everything right. It doesn’t really understand human language. It lacks the ability to anticipate the question. And it doesn’t know to ask the right questions — something that is uniquely human, so far.

All true. But consider: What made the breakthrough that is Watson possible? Why did we see Watson emerge in 2011?

Here’s the New York Times in 2010, writing about Watson:

Ferrucci’s main breakthrough was not the design of any single, brilliant new technique for analyzing language. Indeed, many of the statistical techniques Watson employs were already well known by computer scientists. One important thing that makes Watson so different is its enormous speed and memory. Taking advantage of I.B.M.’s supercomputing heft, Ferrucci’s team input millions of documents into Watson to build up its knowledge base — including, he says, “books, reference material, any sort of dictionary, thesauri, folksonomies, taxonomies, encyclopedias, any kind of reference material you can imagine getting your hands on or licensing. Novels, bibles, plays.” [Emphasis mine]

All that was necessary for Watson to become reality is “enormous speed and memory”. Ferrucci jokes in this TED panel about Watson that the Power750 chips that power Watson is “off the shelf” and he can get you a price for that.

Now… consider the phenomenon known as Moore’s Law:

Moore’s law describes a long-term trend in the history of computing hardware: the number of transistors that can be placed inexpensively on an integrated circuit doubles approximately every two years.

In practice, Moore’s Law has translated into computers doubling in speed and power roughly every two years. And… the real world implications of the Moore’s Law are as follows.

Ten years ago, in 2001, the top of the line Apple computer was the PowerMac G4 800 Dual Processor Quicksilver, a mid-tower desktop, with the following specs:

The Apple Power Macintosh G4/800 DP (Quicksilver) features dual 800 MHz PowerPC 7450 (G4) processors each with the AltiVec “Velocity Engine” vector processing unit, 256k “on chip” level 2 cache, and 2 MB of level 3 backside cache. It shipped configured with 256 MB of RAM, an 80 GB Ultra ATA/66 hard drive, a 2X DVD-R/CD-RW “SuperDrive”, and a 4X AGP NVIDIA GeForce 2 MX TwinView graphics card — capable of powering dual displays — with 64 MB of SDRAM. AirPort (802.11b) was available by custom configuration.

That was state of the art just ten years ago, and it cost $3,500. It did not come with a monitor or keyboard.

The least expensive MacBook Air in 2011 has these specs:

The Apple MacBook Air “Core i5″ 1.6 11″ (Mid-2011/Thunderbolt) features a 32-nm “Sandy Bridge” 1.6 GHz Intel “Core i5″ processor (2467M) with two independent processor “cores” on a single chip, a 3 MB shared level 3 cache, 2 GB or 4GB of onboard 1333 MHz DDR3 SDRAM (which cannot be upgraded after purchase), 64 GB or 128GB of flash storage, and an Intel HD Graphics 3000 graphics processor with either 256 MB or 384 MB of DDR3 SDRAM shared with system memory. This all is packed in a razor thin (0.11-0.68 inch), 2.3 pound, aluminum case with an integrated “FaceTime” video camera, a backlit full-size keyboard (the function keys are smaller, however) and an 11.6″ widescreen TFT LED backlit active-matrix “glossy” display (1366×768 native resolution).

The entry level 13” MacBook Air has 4GB of DDR3 RAM, and 128GB of flash storage, and retails for $1300. It is more than twice as powerful as the cutting edge desktop of 2001, has 16 times the RAM, which is faster and higher performing than the 256MB of RAM in the PowerMac, and comes with video camera, keyboard, and a monitor… in a 2.3lb package. For about a third of the price.

And that’s comparing computers. What about the iPad2? A device whose whole category didn’t even exist five years ago:

It is powered by a custom-designed 1 GHz dual-core Apple A5 “system on a chip”, has 16 GB, 32 GB, or 64 GB of flash memory, front and rear mounted cameras, 802.11a/b/g/n Wi-Fi support, an accelerometer, a three-axis gyroscope, an ambient light sensor, digital compass, a speaker and a built-in mic packed in a 0.34 inch thick, 1.33 pound glass and aluminum case with a black or white front and an aluminum back. The battery life of this iPad 2 is reportedly 10 hours “surfing the web on Wi-Fi, watching video, or listening to music.”

The tiny little iPad, which didn’t even exist as a category ten years ago, has faster CPU (dual-core = dual processor) than the top-of-the-line computer of 2001, double the RAM at 512MB , and nearly as much disk space – except that its “disk drive” is 64GB of flash memory, which is faster than spinning disk drives. And it costs $700.

Just two years from now, Watson will be twice as fast, twice as powerful, have twice the memory capacity, and have larger and faster storage. A mere six years from now, Watson will be eight times as powerful. And maybe everyone will have a Watson-powered Siri-type assistant on their mobile devices over 5G LTE networks….

Conclusions?

It is so easy to spiral into sci-fi land. Technology has a way of surprising us with how quickly it advances, and it also surprises us with how little it changes. The internal combustion engine is still with us, after all.

I don’t know that I draw any particular conclusion about what real estate will look like as technology advances. I do think, however, that the changes are likely to be orthogonal rather than linear — meaning, it isn’t enough to project what we have today forward. We have to think that technology will go sideways… into areas and capabilities we’re not expecting.

Strategy, then, dictates that those in charge of thinking longterm about their companies should pay some attention to these unexpected developments, and start thinking about how to answer Loren’s question: What are you doing now so your service will be relevant 2/4/6/8/10 years from now?

Oh yeah, go buy a copy of Stefan’s report in which my full piece appears. #shamelessplug #butitsforafriend