Rain, Steam and Speed, JMW Turner, 1844
This is the text of a talk I gave at The Chairman's Dinner, prior to the Urban Land Institute's Annual UK Conference on the 4th June, 2018.
There is an old Danish proverb that states ‘It is difficult to make predictions, especially about the future.’ So… obviously I am going to now make a lot of predictions.
Worse than that I am going to make predictions with a five year timespan. As any sane ‘futurist’ would tell you the trick to making predictions is to make them for a long time in the future. That’s why you have ‘the end of the world is nigh’ not ‘the end of the world is tomorrow’, as in the Monty Python sketch.
Near term predictions can make you look very stupid.
But I have a trick up my sleeve… and that trick is technology.
You see, within the tech sector, there is a very well established and pretty much rock solid adoption map. You start out with a product that gets picked up by innovators, and there are always some of them for just about every product, and then with a bit of luck you build a user base amongst the ‘early adopters’. Very soon after that though you reach what Geoffrey Moore, in his great book ‘Crossing the Chasm’ described as… the Chasm. This is where you come up against economics; you need enough traction to fund, or raise the funding for, continued growth.
The other side of the chasm is the ‘Early majority’ which is a market significantly larger than the early adopters. But most products never make it across the chasm. For whatever reason they fail to get enough traction to escape the world of startups. Just like some birds jump the nest before they are able to fly.
BUT for those that do Cross the Chasm, you know what their market development is going to look like. Growth amongst early adopters takes them to about 50% market penetration, and that is normally where they peak in terms of gaining new users. After that it is a matter of hoovering up the late majority, and even later, the laggards.
So, looking at life through a tech lens, yes you can predict the future. You just need to determine which products have ‘crossed the chasm’. Yes there is still the issue of determining speed through the stages, but far less uncertainty of outcome than is often assumed.
What has happened over time is that adoption of new technologies has got faster and faster. So, for example the Internet rose to 50% market penetration far faster than did electricity, the telephone or indeed the washing machine.
In similar fashion it took mobile phones about 17 years to reach 100 million users. The internet took about 8 years, and Facebook just 5. All though are knocked into a cocked hat by the Indian phone network Jio that reached 100 million users in just 170 days, after its launch in September 2016. That’s what free voice and data can do!
We can then see where the near future is headed, and we can be assured that it is going to get there faster than we think.
Last week I gave a talk and used Turners 1844 painting Rain, Steam, Speed for a slide. This is the one where the steam engine is flying across the Maidenhead bridge on the recently completed Great Western Railway. It sums up brilliantly what it must have felt like to Victorians caught up in the First Industrial Revolution. The future rushing towards you and everything in a bit of a blur. But with a feeling of palpable excitement.
That was then and this is now.
Generally speaking that feeling of palpable excitement has diminished, in many cases to be replaced by a feeling of angst and consternation. What were the election of Trump and the vote for Brexit if not a reflection of angst and consternation.
What has also diminished though is an appreciation of just how fast technology is developing, behind the scenes. As Jeff Bezos has said, we are all only impressed for a short while. So, even though you probably all have in your pocket a smartphone that in 1985 would have outgunned a Supercomputer and cost $25 million dollars, you don’t give it a second thought. Whereas 25 years ago you could wait 28 days for a mail order delivery, we now complain that our Amazon order only arrived at 8pm on Sunday, as opposed to 8 am. After all we did order it from our sofa whilst watching a streaming movie 24 whole hours ago.
And that is the flashing red light for the real estate industry.
Regardless of whether or not the populace is viewing technological change through the excitable lens of Rain, Steam, Speed, it is happening faster than ever.
And that is why our safe, if cyclical, slow world of real estate, where you build something you designed ten years ago, largely unchanged, and with no penalty, is set for a disruption that very few believe possible, let alone likely.
The key reason this is both possible and likely is that, contrary to many peoples understanding, human behaviour follows on from technological developments and not vice versa. Technology is not invented in response to human demand. Technology is invented when its invention is possible.
For example, you probably remember the dot com crash of 2000 where trillions of dollars of investment capital evaporated over the following year. Much merriment was had at the tens of millions lost by online pet food companies and video on demand operators. What you might be less likely to know is that the online pet food company Chewy was sold for $3.35 billion last year and that Netflix is currently valued at $153 billion.
The difference? Infrastructure.
In 2000 broadband was in about 26% of homes, and computer chips contained about 100 million transistors. Today broadband is near ubiquitous, at about 96% penetration, and computer chips have 10 billion transistors.
Chewy and Netflix exist because they can exist, not because any human asked for them.
The technology that crosses the chasm does have to either address an existing need or create a new need, but most importantly it needs to work, at least in the context of something users value. For example, the original iPhone was a terrible phone, as Microsoft’s Steve Ballmer very loudly shouted about as he mocked the then not very important Apple. But Steve Jobs knew that people loved being able to take their music around with them and loved taking photos, and on those two vectors the original iPhone was terrific. That took it across the chasm.
It is really unwise to bet against progress in technology.
A few years ago there was much talk about Moore’s Law coming to an end. This was the theory, from Intel co-founder Gordon Moore, that technology at a given price would double in power every two years, and it had held for fifty years. The talk though was that because of the laws of physics transistors simply could not get any smaller and 10 billion on a chip was about the limit. The logic seemed hard to refute. But things have not panned out like that. Instead of progress getting slower, it has sped up.
And for that we have to thank the computer gaming industry.
With a fortuitous twist of fate, it turns out that the type of computer chips that Artificial Intelligence works best on are not the PC chips that Gordon Moore and Intel had been referring to but GPU’s, which are the primary processing units for dedicated gaming machines. Simplistically, in computer games it is vital to be able to do thousands of things at the same time, as opposed to one after another, and that encapsulates the difference between GPU’s and CPU’s.
The consequence of this, the use of custom designed chips to do very specific tasks is that the training of Neural Nets, which are foundational to Machine Learning, increased in speed 60 times in just three years from 2013 to 2016, with the bulk of that happening in one wonder year, 2015.
At the same time the scale of the smartphone industry has increased so dramatically that the cost of any component in one of them has reduced to being ‘cheap as chips’, as the saying goes.
So any sensor in a smartphone, for example GPS, accelerometer, proximity, barometer, biometric, compass, and of course high definition slow motion and high speed still and video cameras is now available in very high quantities and at a very low price.
Combine all of this together and we are entering a world where anything that is ‘structured, repeatable or predictable’ is going to be automated, and if you believe McKinsey, they said in January last year that that applies to 49% of all the tasks people across the globe are paid to do.
The RICS last year published a report saying that 88% of everything a surveyor does could be automated within 10 years.
Does that therefore mean that half of all jobs are going the way of the Dodo and that Surveying might not be the best choice of career?
In short No - because life never works out like that. What it does mean however is that the work we all do, and the way the economy works, will change fundamentally and in so doing will upend the value chain within the real estate industry.
Lets go through some of the sectors:
The Office:
When half the tasks that humans do now are no longer within the purview of humans, and anything ‘structured, repeatable or predictable’ is done, in one way or another, by a ‘machine’ the fundamental purpose of an office changes. ‘New’ work, as I like to call it, in a digital world will be all about human skills. Design, Imagination, Inspiration, Creation, Empathy, Intuition, Innovation, Collaboration, Social intelligence - those are the killer skills of the next 10 years. And they require a new form of office, one that harnesses data and design to catalyse just those skills.
42% of employees work for companies that employ over 250 people. Those companies should be able to muster up the investment, inclination and skill to create the sort of ‘Imaginarium’ that allows their employees to be as productive as they can be. As we know, many of them fail on this at the moment but my feeling is that the top two quartiles will create great spaces in the future, if only because without them they won’t be able to hire the talent without which they will collapse. Faster than ever before.
But 48% of people work for companies that employ less than 50 people, and in many areas, such as the City of London, some 70% of occupied units are less than 10,000 sq ft. 50% are less than 5000 sq ft.
In these companies they most likely will not have the space or skills, even if they do have the inclination, to create great modern, flexible, adaptive, high quality, human-centered spaces.
My prediction is that this sector of the market is a prime target for whoever can come up with the business models and #SpaceAsAService thinking that delivers a Product that fits their requirements. ‘Everyone deserves a fantastic workplace’ to quote Neil Usher, but most small companies, half or more of the market, cannot deliver this for themselves. Someone needs to do it. And whoever does it best will build a great Brand with great value.
The flip side of this is that there is going to be a huge amount of obsolete space around. We need fewer but better offices.
From an investment standpoint the big trend will be the shortening of leases, and the increasing change from a valuation model with bond like characteristics to one where the operator really matters, as they will be the driver of revenue. The curator of the user experience, the UX, will be perhaps the most important component of an investment. The owner of the asset, if not the curator, will move down the pecking order.
Retail:
In the same way that, if we admit to reality, we do not need offices to work, we do not need shops to shop. In both cases we need to be made to want them. It is interesting to see and hear the attention the Grimsey Report 2 is getting at the moment. Five years ago I was at the launch of the original report at the Houses of Parliament. Thereafter Parliament completely ignored it, whilst indulging in endless fantasies about ‘saving the High Street’. That was an obvious wrong move five years ago and thankfully today you hear less nonsense along those lines, even if not a great deal of coherent action.
The reality of retail is very simple. Mostly shopping is dull, if necessary. For those essential items that we forget about and run out of, and for anything ‘cheap’, too cheap to ship, that we want or need to buy, there is a solid future for physical outlets to supply. At the other end of the market, for shopping that is fun, we both need and desire great retail experiences that pander to our innate human need to socialise and acquire. Everyone, or almost everyone, loves some form of shopping. Give them a great experience and they will part with their cash.
For everything in between cheap and everyday, or fun and an experience, we can rely on Mr Bezos to handle it.
Amazon spent $13 billion on Whole Foods and Alibaba have spent $9.3 billion on offline retail. Do not be misled by this - they have not done so because they think offline retail is great and an essential asset. They have done so because, with their vast technological capabilities, mountains of data, and personal relationship with millions upon millions of online shoppers, they believe they can build a ‘New Retail’ model that is much superior to what the incumbent offline industry provides.
The thing to worry about is whether the software industry can learn retail quicker than the retail industry can learn software.
Industrial:
Three words covers all you really need to know about the industrial market: ‘robots’ is one and ‘last mile’ are the other two.
The industrial market is all about automation. Amazon may put out PR pieces about the number of people they employ in warehouses but that is hardly the end game. Complete automation of warehouses is simply a technological challenge. And when you are getting 60 times better at machine learning every three years it is only a matter of time before the best warehouses, run by the best companies, will be entirely dark. Many already are. Many will follow.
It is about more than automation though. A company called Clutter in the US is a new breed of ‘Big Yellow’ type self storage operation. With two differences to the standard operating manual. First, they use AI to help them optimise storage; they don’t store all your goods together but mattresses with mattresses, bicycles with bicycles etc. This means they can utilise their space much more efficiently. And they have their warehouses many miles outside of the cities they serve, where they are cheap. Using an app you can see and track everything you have in store and request delivery within hourly slots. What they lose in proximity they make up for with technology.
Cracking the last mile though, is the big problem the industrial real estate sector needs to help solve. How can we help enable Amazon, or whoever, to get an order to a customer in an hour? And no, Click and Collect is not the answer to that; it is simply a temporary fix that only those in the real estate industry think is an end point.
Alibaba own a supermarket chain in China called Hema. If you live within three kilometres of one of their stores they will deliver to you in…. 30 minutes.
Amazon, even prior to the acquisition of Whole Foods, had a distribution centre within 20 miles of 45% of the US population.
Industrial real estate strategies simply need to focus on those three words. Robots and last mile.
Residential & Build to Rent
Clearly we are not building enough homes, and have not done so for decades. Even if we did they would today be beyond the reach of most young people.
So it is inevitable that the Build to Rent sector will grow, probably dramatically. It has crossed the chasm. Regulatory issues need to be sorted out but I cannot see how they will not be.
This sector has the chance to replace prime offices as the default home of large chunks of institutional money. As we have seen the office market is, for a large percentage of it, dying in terms of providing Bond like returns whereas whilst one can quite easily consider buying office space on-demand that is harder to square with where one lives. As such, stable long term returns should be on offer and might well become the asset of choice to institutions that require stable long term returns.
It is also highly likely, desirable and sensible that large scale build to rent developments will be mixed use and include significant amounts of office or other commercial space. As one does not need to be in a human-centred, deeply collaborative working environment five days a week it makes much sense to add facilities near to where people live.
Hotels
Not my field but demand for Hotels rises inexorably, despite the rise of AirBnb, largely because the world is getting richer and more people are travelling. Barring cataclysmic events, hotels have a strong future. Prosperity, as well as demography, is destiny.
Planning, Design, Construction
If you want one sure bet within the real estate sector then the increasing use of technology in Planning, Design and Construction is it. Together with these silos becoming increasingly joined up. We will not be worrying about the supply of construction workers in ten years. Fortunately, as there won’t be enough. These sectors are being pushed by human constraints and pulled by technological advances. They are only moving in one direction.
In conclusion then we are in a world that Turner would have understood, where everything is a bit of a blur but palpably exciting. Even if we are begrudging in accepting that that is the case, it really is.
Steam engines transformed Victorian society and AI and the 4th Industrial revolution is changing ours. To quote though another great artist, Picasso; ‘Computers are stupid - they can only give you answers’ - it is up to us, their masters, to realise that where we end up, as individuals and as a society, is down to our use of that human skill where we, at least for now, greatly outshine the machines. Judgement. As AI reduces the price of Prediction, the value of Judgement goes up.
And finally, to paraphrase Gary Kasparov, the great chess player and thinker, a human plus a computer, trumps any computer on its own.
Cogito ergo sum
Antony