How a trash can in Iceland will force Amazon to adapt its business model
The Internet of Things, Distributed Computing and their impact on the Platform Economy
Everywhere you go in the tech world (and now even way beyond) people are either thinking about building a platform, have started on it, or already operate one. Consulting companies like Accenture are advising everyone should have a platform strategy by 2018 at the latest and offer 100 day roadmaps. CDOs of all sorts of companies shout it from the rooftop, and popular blogs like TechCrunch are full of articles about the “platform economy”.
Right now there is an entire industry singing the gospel of the Holy Platform: It will boost your revenues, it will help cut costs and it works with any business model in any industry. It’s really too easy: Just offer the infrastructure for vendors and buyers to conduct their business, charge a fee/ premium, use data as leverage to reduce (or use) inefficiencies and cash out twice in the process. And it would surely lead into the golden age of the holy platform for everybody if there was not one tiny little problem:
The time of owning users’ data as a company and capitalising on it is coming to an end, and the reason is the rise of Distributed Computing through the Internet of Things.
You might very well think now “Wait… what? Are you crazy? That is not connected at all!” Well, I assure you it is, and I will lay out why. But now please consider the following thought for a moment:
A networked trash can in Iceland will eventually force Amazon to adapt its business model.
Let it sink in, in all its strangeness. And now, let’s back up for a second, have a quick look at platforms, and come back to that later.
The infrastructure on which companies like Amazon and Google have founded their business are dubbed “platform economy” today. The idea is that the layout of the infrastructure is so specific and crucial to these companies’ success that it constitutes a business model in itself. In other words: Amazon and Google are their own business models, they do not operate any. That at least is behind the idea that there is a “platform economy”. And attached to that is the belief that platforms shape the way we do business. But is that really true? Well, I for one think it is the other way around.
Our needs and behaviour shape platforms, they do not shape us.
Let’s think about why platforms exist in the first place, and why they look they way they do: they are here because they allow for us to complete different kinds of transactions very efficiently. And efficiency (or comfort) is really the key here, because it explains the motivation of a user to go to a platform-based service and not use an alternative.
The main reason is that utilising Google’s and Amazon’s products is an excellent use of one’s time and therefore highly advantageous. It produces a fantastic Return on Investment. I use email (especially Gmail) because it has countless advantages over a letter. I buy things on Amazon because of its superior offer range and availability of products. And for this surplus in efficiency and comfort I pay a premium which is charged in something far more valuable than money: personal data.
Services who can offer a significant advantage in efficiency over alternatives will attract users.
Since giving away personal data is easy and painless, customers flock to services who offer the trade-off mentioned, being happy about the efficiency gained. These services however need to be catered to, and they need an infrastructure that produces precisely the efficiency that attracts users in the first place. That is why the infrastructure of platforms takes a shape that ultimately supports the users’ needs, mirrored in the different services that are offered.
Moreover, the layout and functionality of platforms is determined by the form of the services and the way they are conducted. That is why Amazon’s technical and organisational layout is fundamentally different from Google’s. Emailing and internet search require a different structure than selling physical and virtual goods online. In other words: platform does not equal platform.
So the very reason platforms came to be is the same reason why their shape will change fundamentally: Because transactions shape services. Services in turn shape platforms. Transactions change according to users’ needs, and users’ needs are about to change dramatically. Here is why:
Every single one of us will produce too much data to be sent to any cloud for processing because we will own more and more devices that generate and compute data.
I am not talking about a second or third smartphone, I am talking about the Internet of Things (IoT). And although it is getting more attention these days, most people do not realize the true impact it will have on our way of living. Looking at some numbers might help to grasp what is to come though:
In 2015, we had 4.9 billion (bn) connected things in the world, of which an approximate 1.4 bn were smartphones. Cisco and some others predict that by 2020 the number of internet connected things will have exceeded 50 bn, of which 6.1 bn are smartphones. Gartner believes that a quarter of a billion of these devices will be connected vehicles with higher and higher degrees of autonomous driving.
In 2020 there will be 6.4 times more networked devices than people on earth, and counting.
Even with a continuous evolution in network technology and computing power it is quite likely that the sheer amount of data produced by all these devices can neither be sent to centralised cloud infrastructures anymore nor be processed there in time and sent back. Just think of ONE self driving car and the amount it generates in one day. Estimations today circle around 40 Terabytes (TB) per eight hours of driving. It will have to send around 4,000 Gigabyte (GB) for one hour of driving. And that is just ONE car. Multiply this with 250 million (mn) cars in 2020 and you get a staggering 1 bn Terabyte or 1 mn Petabyte (PB) of data output for one hour. As a comparison: at the moment we use 650 Megabyte (MB) on average per day and person. That means ONE hour of driving with ONE car equals the combined daily data use of around 6,150 people right now.
The “Send — Calculate — Respond”- loop to the cloud will simply take too long — or will even be interrupted due to bad network coverage and overload.
With a self driving car it becomes painfully obvious: it cannot rely on a network connection to make decisions. If a kid is running onto a street and the car has to react, there is simply no time to wait until a faraway instance will have calculated the best possible scenario. And that leads to one very likely outcome: the data storage and processing has to become localized again.
Sending data to a central processing and storage unit away will be disadvantageous.
The shift has a historic parallel: the first computers were mainframes that stored and processed all data, accessed by “dumb” terminals. After that, PCs were the expression of decentralised data processing, at home or in the office. With the rise of the internet and the accumulation of enormous computing power through big companies like Google and Amazon, cloud computing was (and is) an efficient way of using infrastructure that is already in place as a commodity, both for storage and computing. But through the sheer masses of data and the need to process it on the spot (or in our own network at least), we will experience a shift to local computing at the edge of the network — hence the synonym “edge computing”. And that will lead to one significant change:
We will own our data again — and the services will come to us, not the other way around.
Services come where the users are. And if the users (especially their data) stay in their own network, then services will ultimately show up there. We will still want and need services that are hosted on or attached to big platforms today, and big platforms will not die all of a sudden, but the way we use them will be very different.
Muneeb Ali, a Distributed Systems expert at Blockstack predicts that among other things cloud storage providers will be “dumbed down” in order to serve as encrypted storage boxes while we will own our “personal cloud servers” to process our data, much like the Virtual Machines of today but owned by the user. Programs like Boxcryptor allow for this already, turning a Google Drive into an encrypted box only the user can look into. Moreover, crypto tokens will enable us to use third party software on these machines, so we move the service to our domain as long as we need it and then switch it off when we’re done. Structures like that also exist already, with Microsoft Office for example. You install the software locally and get keys for certain usages. A natural step for platforms to evolve to is therefore token exchange marketplaces where “personal clouds” can go shopping according to their needs.
In way we are going back to Personal Computers — we own all components, but the barrier between hard- and software is mostly virtual.
A popular developer joke goes “There is no cloud — it’s just somebody else’s computer.” And that is true. But with the rise of distributed computing we will own our own computers again, and also our data through advanced encryption and the use of tokens.
That will make it much harder for any platform and/ or Software as a Service (SaaS) companies to acquire data that can be monetised apart from the transactional value and the records that occur in any marketplace. So, platform companies like Google and Amazon will have to adapt to reach their users and come where the computing and data storage is done. They will also have to deal with the fact that they will not be able to identify users anymore as they can do it today.
The Internet of Things will do more for privacy than any data protection law could.
Remember the trash can from Iceland we started out with? Well, I have to admit I am not certain that it will be a particular trash can right there who will make the mighty Google and Amazon change their ways, but it definitely will play a part in tipping the balance from the era of centralised data storage and cloud computing to decentralised data and computing with a clear ownership on the user’s side. And that is good news for all of us, both users and companies.
