Today’s Internet Service Model
While egalitarian notions like equality and neutrality underpin the Internet, the reality today in the online services realm is that users need to make choices about which platform they want to join. Whether the decision favours Apple (iOS), Google (Android) or Microsoft (Windows), the more the user adopts their chosen platform, the harder it is to leave.
This is a super-sized version of what already exists in other ‘platform’ markets – like console games where, having made a significant investment in an Xbox ONE, for example, the user is not only limited to games published by Microsoft – but is unable to benefit from innovations that occur outside that platform.
The mega-service delivery platforms that dominate the web today seek to achieve ‘platform lock-in’ which is a condition where users become so tightly bonded to a given platform that they cannot seriously contemplate leaving.
Companies like Amazon, Facebook, Apple, Google, Microsoft and others are all slogging it out using the same strategy which involves a race to scale: scale in users, breadth of devices and an ever-expanding range of services and apps.
But this is not stopping a steady stream of start-ups coming into the market. However, the investors who have backed those start-ups mostly see the big platforms as an exit route. Facebook is the best example of this – a core component of the company’s strategy is to invest heavily to acquire rival services. In contrast, Google tends to acquire start-ups for their technology and/or people while Apple is spending such a lot on product development that there are few start-ups that have anything valuable to offer.
This model is shown below:
Do Apple and Google have sustainable service delivery platforms?
There are two reasons why the service and application development infrastructures operated by Google (Android) and Apple (iOS) cannot rule on a sustainable basis:
Firstly, as the complexity of their platform models increase so they become harder to manage: bugs take longer to fix, changes have increasingly wider ramifications and new products and services take longer to introduce, time to market will slow.
In other words, the bigger these platforms become the slower they will get and, eventually, Google and Apple will move from growth to value stocks. That is not to say that Google and Apple will not still be fantastically profitable – just as Microsoft remains staggeringly profitable today. But these giants will no longer be able to control service innovation.
The second reason relates to something that was missing in the above diagram: which is the data that is being harvested:
Specifically, we are talking about the ‘big data’ part of the IoT: clearly pure connectivity by itself is not valuable. Instead it is the data that is transferred as a result of connectivity that is valuable. Ultimately, the term ‘IoT’ is a misnomer: in hindsight perhaps we should have used the term ‘IoD’, or Infinity of Data.
The data volumes we are talking about are truly mind-boggling. Firstly, the total number of connected objects breaks down as follows:
- Users: 1.2 billion adults were living in homes that had a fixed broadband connection by the end of 2014. This number will rise to 1.5 billion by 2018;
- Devices: If we just include personal devices and home connected devices then the total number of connected objects reached 5.3 billion at the end of 2014 and will reach 8.6 billion by the end of 2018;
- Services: There were 139 million online services worldwide at the end of 2014 and this will rise to 201 million by the end of 2018.
The metadata alone required to describe these objects and their interrelationships is mind-boggling: our estimation is 20 trillion discrete data files and over 900 PetaBytes (PB) of data. Remember, this is just metadata.
But if we look at the underlying data then we are already in the Exabytes league. The following chart shows how rapidly the total volume of data could increase in the coming years:
If we assume that the total volume of data that is available (although not necessarily accessible) triples each year (which is approximately exponential) then this will mean that the volume of data available in 10 years will be approaching 20,000 times the volume today.
Ultimately, the rate of human advancement is limited by the speed with which we can identify problems and then develop solutions to those problems, or otherwise to see ways to make things better. It seems pretty clear, therefore, that there is a direct correlation between the amount of data we efficiently access and our rate of advancement.
We are not talking about access to rather trivial data such as a user’s web browsing history or their location. We are instead talking about data like real-time access to the internal biochemistry of 100s of millions of opt-in users, and much, much more.
The commercial value that is locked up in all this data will be immense. In so far as innovations that will be enabled by this data then we should expect:
- Early warning systems for critical illnesses that will save trillions of dollars in healthcare costs and prolong life for millions of people;
- New ways to commercialise and distribute digital media that will make services like Netflix and Pandora irrelevant;
- A range of new online service categories, each of which could grow to the size of today’s mobile communications industry;
- The financial services and insurance industries could be transformed by eliminating zero value-added trading activities and increasing the efficiency with which investment capital is allocated.
I do not think it is realistic to expect that Google, Apple or Facebook will be able to control how these and other similar-scale opportunities are exploited.
In this view of the market’s development, the increasingly slow platforms operated by Google, Apple etc. will not just become frustrating for developers who want to bring new services to market but they will be suffocated by the weight of exponentially-increasing volumes of data – which, even with machine intelligence, will be impossible for a few senior managers to properly analyse.
Instead, an army of thousands of developers will use cloud-based AIs to do their own analysis in order to spot and exploit opportunities themselves.
Case Study: Mobile internet
One way of looking at this would be to think of the islands of connectivity that have formed around modern developer platforms – like Android, iOS, Facebook and Amazon – as being analogous to the content portals that were launched by mobile network operators around 2002- 2005.
At this time mobile operators thought that they would be able to control access to their users and so they set up mobile content portals through which users would access a wide range of content and content-based services.
The only way for a brand like ‘The New York Times’ to distribute their content to mobile devices was to make a deal with a mobile network operator who insisted on a share of any resulting service revenues.
To us, this failed model looks rather similar to the platform models that are being operated by Apple, Google and others, albeit at a higher level.
The mobile content portal model began to fall apart when users were able to access third party services using their mobile data plans. This allowed them to completely bypass the mobile network operator’s content portal. At first this was not a problem because mobile data charges were high and the number of users having suitable phones was low.
But then, slowly but surely, more and more users began to access services independently. And then flat-rate mobile data plans arrived, thereby increasing the number of users who wanted to access third-party services. By the time Apple and Google had introduced their respective mobile devices, just 5 years later in 2007, an initial trickle of leakage had turned into a raging torrent.
Another way of looking at this is that there was no way that a small team of product managers working for a mobile operator could match the collective creativity of 1000s of external developers and so the mobile content portal model was never sustainable.
Returning to IoT, when we begin to see the first signs that it becomes possible for third parties to enter the market and develop apps that span multiple devices and services – as opposed to being restricted to one service domain (e.g. Apple), then a similar trend will take hold.
New Service Vision
Rather as happened with mobile content portals and online service portals, developers will feel themselves being magnetically drawn towards an ever-expanding volume of data on users, devices and services – and on the relationships between these.
While the big platforms will try to force developers to work via their own platforms, this will prove fruitless in the end.
A new generation of entrepreneurs and developers will want to directly harvest their share of trillions of dollars of commercial value that is locked up in a vast trove of data. When the pressure reaches a critical point, a rupture will occur in the fabric of the market.
At this point, I expect to see the emergence of a new layer of service-delivery and application development infrastructure that will sit between users and developers – and will be independent of Apple, Google, Facebook and the rest – which will then be regarded as legacy platforms.
This infrastructure needed for this new platform could be put in place by a start-up which gains critical mass on the basis of a radically different proposition to developers, rather than just offering another version of what already exists. Alternatively, this could arise as a result of large players in the market who are currently locked out of the Apple/Google service platform market (or who are too dependent on it). Examples would include Amazon, Alibaba, Samsung and Huawei.
Either way, while the big platform providers are likely to live on as successful, highly-profitable companies – just as has proven the case with Microsoft – they will not be able to control the service innovation process indefinitely.
