Some time back IT departments stopped thinking about services as one half of goods-and-services, and started thinking about them as a way to provided everything. When that happened many intractable problems became, well, tractable. Today the service-oriented operational model for IT departments seems like a no-brainer. But it took a while for the new approach to feel normal.
Anyway, I’m not here to talk about services. And I am sorry about the bait and switch but I want to prepare you to think about another very familiar topic in an unfamiliar way. And before I do, I want to remind you of how well that can sometimes work out.
My topic is platforms.
If I asked you to name a few for me most of you would clearly start with Windows. Android and iOS will certainly get a look in. In IT we associate platforms with software environments such as operating systems, application development frameworks and middleware. Iately we started using the term platform to refer to new kinds of services. Facebook, Amazon AWS, and Google are emblematic of these new platforms. But things we once thought of as applications now have a kind of dual identity as both applications and platforms. SalesForce CRM, BlackBoard LMS and now even SAP ERP are all applications that have been repositioned as platforms.
In IT we understand the importance of platforms better than we ever have.
Still, I we limit our understanding of platforms if we only associate them with software systems.
Platform is a very metaphorical word. A platform is something on which other things rest. Platform is a good solid weighty word. If it had a smell it would be of freshly sawn timber. The word itself encourages us to think of platforms structurally. It makes intuitive sense to think of platforms as infrastructure – something akin to roads or a power grid. But while platforms can provide infrastructure, they are much more than that.
Platforms are Economies
Take a look at this picture of the Prahran Market in inner Melbourne, where I shop (and enjoy excellent coffee) most weekends.
Do you see a business or a platform?
Underneath all that activity there is a shared superstructure of pipes, electrical wires, garbage disposal services, and storage facilities. All those service combine to make it possible to set up many different business and to bring those business together to create a vibrant community of traders and customers. The individually owned stalls in this market could not operate without those services.
They would not be viable on their own.
It’s easier to see how the market is a platform if we look at it as a map.
See those extra bits, the drinking water, the toilets, the free wifi, and even a place to recharge your electric wheelchair. It’s not just a collection of stalls. And don’t forget the business services as well. They add things like security, advertising, and insurance, to the mix.
The market actually works much the way an IT platform works. The market provides the operating environment in which stall-holders run their business. The market is the operating system and the stalls are the applications.
Car manufacturers have platforms too. And so do drug companies, airlines and armies. In these industries platforms are collections of shared technologies, components and knowledge that can be combined in different ways to meet different needs without having to build everything from scratch.
Here is another kind of platform with which I am very familiar.
This is Monash University where I currently work.
Universities, like markets and operating systems, are also platforms. A University provides resources to create applications. In this case courses and research projects. Universities are very versatile platforms. On what other platform can you build a symphony orchestra a hospital, and have resources left over to let scientists walk around inside terabytes of data?
(those people in the picture are wearing 3d glasses.)
A Platform is an encapsulated economy.
Platforms are arbitrary, but deliberate, collections of resources that create economies in which other enterprises can be viable.
I am not sure we can neatly define what a platform is and isn’t in all cases. Some things fit the idea very well, others seem just a bit ‘platformy’. But we can make some useful observations:
Platforms remove scarcity
As an economy, a platform, like all economies, is a solution to the problem of scarcity.
We should not, however, confuse them with markets. Platforms are often built and run more on the command economy model than the free market one.
Platforms don’t prescribe their uses
An application or service provides clearly defined products and behaviours. A platform succeeds best when it provides tools that others can use to solve their own problems. It is not possible to exhaustively predict what happens on a platform. They allow some level of functional autonomy. In short the let users hack.
Platforms are superstructures
Platforms allow users to solve problems with some level of autonomy and to create new things, but they remain unchanged by use.
This is a debatable point and some may argue that platforms are even better when they are self-hackable. Some software-systems or services are actually designed this way.
I suspect that while some method for structural change is needed for real world platforms, that shouldn’t override the functional raison d’etre of a platform. Which is the viability of the enterprises it hosts. Protecting that viability would inevitably mean that platforms should be built so that the functionality and viability of the platform is independent of the functionality and viability of the enterprises it hosts.
At the very least structural separation of the viability of the platform and the hosted enterprises should be mark the difference between well built and poorly built platforms.
Platforms create network effects
The more users (or applications) a platform has the more value each tends to derive from the platform.
However, the network effect can also create lock-in. On the one hand it can be difficult to change or replace a platform that is supporting a large number of enterprises. On the other hand it can be equally challenging to transfer a successful enterprise to an alternative platform.
When the platform itself is a proprietary for-profit enterprise in its own right the owners of a platform can use network effect to create a localised monopoly. In such cases cost of migration is the true price signal and not cost of production. The owner will motivated to maintain low production costs and high migration costs.
Network effects can created social equity and access challenges. Within a platform scarcity of opportunity can be removed to the advantage of the platform’s hosted enterprises while barriers on entry to the platform by new enterprises can remain high, restricting opportunity in the wider community.
Platforms are hierarchical
Platforms can be built on other platforms. In the terms I have outlined here, it is perfectly valid to think of nations as platforms. Likewise organisations, such as Monash University described above. Monash University is a platform that makes the creation of research projects and courses viable. Monash University’s own viability is secured by the resources and services the national government provides, and which in this country, it shares with 38 other universities. In turn the University owns and operates many information services, that are, in a more limited way platforms.
Platforms can collapse
Platforms can collapse, that is become unsuited to securing the viability of their hosted enterprises. This can happen in different ways:
- An underlying platform collapses or is significantly degraded.
- The scarcities a platform removes are no longer relevant.
- A bigger more open or efficient platform becomes available.
And the winner is…..
The internet of course. The internet is the most interesting platform to emerge so far.
It removed two scarcities that were so fundamental to our experiences we barely considered them until the internet removed it – proximity and information.
It turned out that the cost of finding things out and establishing relationships was actually quite high.
The internet has driven down the cost of proximity and information to almost nothing. And look at what that’s done.