Note: My thanks to Orin Zebest for all the photographs, provided via Flickr on a Creative Commons Attribution Licence. Orin, you’re Ze Best. And I’ve left all your original titles in!.
Note: I had some trouble with the photographs when viewed via the permalink. I’ve reloaded each one from a different “source” and with standardised parameters and it seems to work. Let me know if you have any trouble.
2010: The Year of Platforms
I think 2010 is going to be the Year of Platforms. Not Snake-Oil-as-A-Service. Real honest-to-goodness heavy-lifting platforms. The stuff that makes it possible for everyone to have Everything-As-A-Service.
Some of you think that platforms are passe, so 2007. Some of you think that platforms are cloud-cuckoo-land, to be filed alongside the Paperless Office and the Paperless Loo. To my mind there’s something very William Gibson-ish about platforms: the future’s already here, it’s just unevenly distributed.
In 2010, we will see this distribution become more even. We use platforms every day, it’s just not that obvious to us. A credit card is a platform, as Richard Schmalensee and David Evans pointed out so vividly in Paying With Plastic. An airport is a platform. Facebook is a platform, as is Twitter. As is LinkedIn. If you’re using a smartphone to read this, then you’re probably using a platform: both iPhone as well as Android are platforms. If you’re at a desktop and using Firefox and WordPress, as I am, they’re platforms as well. Amazon.com is a platform, as is Force.com. Ribbit, the reason I spend a good deal of time in San Francisco, is a platform. Each with its own ecosystem. Each working with other platforms in a co-opetitive, almost fractal way.
So just what is a platform? A place. A device. A company. An everyday item. Bits of software. All of the above.
When I say “platform” I mean:
- something that is a foundation, an enabling environment, upon which others can build things, make things
- something that exists for a specific purpose (or set of purposes), and which invests in capabilities related to those purposes
- something that then makes it easy for people to use those capabilities
- something that does all this in a commercial model that facilitates the creation and development of new products, new services, new markets, new marketplaces
- something that can coexist with other platforms and ecosystems
Humour me for a minute or two. Imagine what would happen if enterprise IT departments started behaving like the platforms that I defined above. A foundation. An enabling environment. Something that exists for a specific set of capabilities, that executes on those capabilities, that makes it easy for people to use those capabilities, that supports the creation and development of new products and services. You know something? I think many boards would be happy to have an IT department that did just that, that behaved like the platforms defined above.
It’s not just about IT departments. It’s about all shared services. Actually it’s about all services. You see, whenever something gets produced and consumed at the point of production, whenever something cannot be inventoried or “bottled”, there is something quintessentially human about the phenomenon. So we have more to learn from biology than we have from physics, something we’re slowly getting better at. Slowly.
The importance of trust
Service is a human concept. Human beings concern themselves about all sorts of things above and beyond the “fit-for-purpose-ness” of the service. They care about their personal safety and security, about fairness and equality in the environment around them, about simplicity and convenience of use, about many such things. And they care about the exchange of value taking place, what they have to give up, what they gain for it. Humans want to trust the people who provide them with the services, they want to trust the people who provide the platforms underpinning the services.
In a hierarchical world, with deep vertical integration and end-to-end control, this may have seemed easy. In a networked world all this becomes a lot harder, as vertical integration becomes less feasible, as services become more and more “horizontal”, as end-to-end control becomes a nonsense. In such a context, trust becomes more and more important, a point that Chris Brogan makes eloquently in his book Trust Agents. Let me give an example. People can give me a million reasons why Facebook should be considered “closed” and “evil” and whatever else. But to me Facebook is the place where Dave Morin works, where David Recordon works, where Chris Kelly works. They become the face of Facebook to me, and if I trust them I trust Facebook. I cannot do otherwise. It’s the same with Amazon. Every time I meet Werner Vogels I meet Amazon. Trust agents. If I don’t like something I am free to express it; if enough people express themselves similarly then things change. Customer-driven change, built around trust relationships. That’s the way it is nowadays.
Anything that aspires to be a platform needs to engender this trust. So when you look at “platform APIs” don’t be surprised at what they do at their core. They’re usually about a very small number of things:
- user directories, adding and removing people, grouping and classification
- identity, authentication and permissioning
- service and data inventorying, cataloguing and access
- publishing of things digital
- distribution of things digital
The need for openness and transparency
Much of this is done to satisfy the security, safety, privacy and confidentiality aspects of human needs. It’s not about control. It’s about what people want. Of course the platforms can do this more openly, more effectively. But we have to remember these are pioneering times for open platforms. Marty Cooper made the first mobile phone call in 1973. Tim Berners-Lee wrote his Web paper in 1989. Software-based open multisided platforms are relatively new in comparison, and they will adapt to achieving the trust levels necessary.
Back to the IT department. One of the reasons people distrusted the IT department was the smoke-and-mirrors black-box nature of the beast. What was not expressed clearly was not understood. What was not understood was not trusted. Back to the trust issue. Do what you say you’re going to do. Do it. Prove you did it.
This requires something somewhat alien to the command-and-control nature of the traditional firm. Openness and transparency.
Verifiability
That’s why you can find open, accessible and extensive documentation on APIs in places like the Facebook Developer Wiki. But it goes further than that, because trust works in daisy chains. So Facebook have to say “policy” things like “You must not use a user’s session key to make an API call on behalf of another user.”. Why? So that their identification, authentication and permissioning is seen to work. And seen to work verifiably.
Another example. This is what Apple has to say as part of the documentation for the iPhone Dev Center, under Fast Launch, Short Use:
The strength of iPhone OS–based devices is their immediacy. A typical user pulls a device out of a pocket or bag and uses it for a few seconds, or maybe a few minutes, before putting it away again. The user might be taking a phone call, looking up a contact, changing the current song, or getting some piece of information during that time.
In iPhone OS, only one foreground application runs at a time. This means that every time the user taps your application’s icon on the Home screen, your application must launch and initialize itself quickly to minimize the delay. If your application takes a long time to launch, the user may be less inclined to use it.
In addition to launching quickly, your application must be prepared to exit quickly too. Whenever the user leaves the context of your application, whether by pressing the Home button or by using a feature that opens content in another application, iPhone OS tells your application to quit. At that time, you need to save any unsaved changes to disk and exit as quickly as possible. If your application takes more than 5 seconds to quit, the system may terminate it outright.
So Apple take care of the user experience through the policies and guidelines of their platform.
As I said before, it’s not just about the IT department, I used them as an example. Every firm is a platform. Why stop at firms? This thing is fractal. Aggregations of firms, entire markets, are platforms.
Even governments are platforms. Platforms that identify, authenticate and permission people to use products and services, that allow them to publish services and data, to subscribe to services and data, in a controlled manner. Platforms that allow people to build new services simply and efficiently, that allow markets to form and be formed.
Yes, governments too are platforms. Something that Tim O’Reilly has been driving for quite some time, and something that the current administration appears to be taking seriously. But open government is no simple matter, even with all the heart and will in place. We use terms like collaboration and teamwork and innovation freely, but making them work in a government context is easier said than done.
Small pieces loosely joined
Yesterday my daughter wanted me to buy something from AllPosters,, and when I did I was faced with a variety of payment options. Not just the traditional Visa or Mastercard. But stuff like PayPal and Amazon Payments. Sometime before that I was using InstaPaper to bookmark stuff I wanted to read later, and I watched some stuff on Boxee. In both cases I think I used Facebook Connect. Some of you have heard me speak about using last.fm and audioscrobbler and FoxyTunes and TwittyTunes and Firefox and Twitter in a simple chain before.
It’s where things are going. Sets of small horizontal services doing simple but important things, with the customer having a level of choice at each stage rather than being faced with lock-in. Platforms have to be about choice, no one wants to learn the AOL way again. Walled gardens do a prison make.
It’s not enough to be open, the platforms have to focus on open innovation. As the saying goes, none of us is as smart as all of us. Whatever set we belong to, the aggregate of smart people outside the set will usually overwhelm the aggregate of smart people within the set. Innovation takes place most effectively at the edge where two well-bounded domains meet, and collaboration becomes even more important as a result.
I guess all of this strikes some of you as utopian and rose-tinted and overly optimistic, but I urge you to look for yourself. Think about mail. Think about publishing. About “sharing”. About bookmarking. About paying. About watching. About reading. In a digital context you have choice for every one of these activities. Sure there are dominant market leaders. And sure there is immense resistance to their dominance. Monocultures will not be tolerated, not by the public, not by the regulators, not by the competitors. It’s only a matter of time.
Which means that interoperability and standards become very important.
Interoperability, standardisation and convergence
It has taken a long time for people to figure out that the data centre and the exchange are like Kipling’s Judy O’Grady and the Colonel’s Lady, sisters under the skin. The very concept of cloud computing, and of cloud services, has been a long time in the making. And we’re going to need a lot of work done to get interoperability right, to get the standards right. And the standards aren’t just about formats and protocols, they’re about the data. Which is why microformats are going to grow in importance, why Linked Data will become critical, why the Web Science Trust set up by Tim Berners-Lee is such an exciting proposition.
As all this takes place, we have to keep reminding ourselves of the biggest change that has taken place as a result of the Web. The power of Us.
Customer power and rights
Anybody can build a digital bookstore, but they can’t get millions of reviews overnight. Anyone can build a photo site, but not get a bazillion tags overnight. Anyone can build an auction house, but not get millions of buyer and seller ratings. Anyone can build a social network, but not get yottabytes of user-generated content stored with them. In today’s world, 20 million is a public beta and 500 million the table stakes for entry into global marketplaces. People will come where they can deposit their data easily and take it out as easily. They know that they are instrumental in creating value. So initiatives like Doc Searls’ VRM will become very important. [Sometimes people get hung up about the name. Don’t. The concept is important, not the name.]
Scale and its implications
As I said before, all this is taking place at scales where we’ve never operated before, or even conceived of operating before. Skype, Apple, Facebook, Amazon, PayPal, they’re all showing us scale in a way we’ve never seen before. But up to now much of the scale has been achieved in some sort of walled garden. That’s going to change. Google. What we’re seeing with OpenSocial and Android is not to be taken lightly, what we’re seeing with Facebook Connect and Amazon Web Services is going to get bigger and bigger.
Platform-based scale has its effects on cost points and price points, on coverage and availability. And the changes have already taken place. It takes nearly nine minutes for light from the sun to reach us. Something similar is happening with platforms. Platforms beget scale. And ecosystems.
IT departments will have a choice. Firms will have a choice. Governments will have a choice. To paraphrase Gandhi, they can be the change they want to see. Or fossilise watching.
Health. Education. Welfare. Communications. Transportation. Welcome to the world of platforms. Or…..
Back to the old IT department. Creating and operating an enabling environment. Handling the directories and catalogues and relationships. In some cases operating Apple-like and “certifying” the applications, in other cases taking a laissez faire approach like Facebook does. Leaving the choice of device to the individual. Letting that individual select the services she wants. Relaxed about the hosting of those services, making that the responsibility of the application provider. Focusing on doing the core things well, in an open multisided marketplace.
So what’s wrong with the picture?
Yup, the sky’s got a fold in it.
If we don’t get the cloud computing environment right, we will hold all this up for a few more years. Which would be a terrible waste. A waste of energy, of scarce resource, of money, of time. Of everything.
Platforms are the way forward. Platforms can and will happen. Platforms are happening. We just have to make sure that the infrastructure for the platforms is done right. Infrastructure in terms of compute cycles, storage and bandwidth. Infrastructure in terms of interoperability protocols, standards and guidelines. Infrastructure in terms of duties and rights and regulations. Infrastructure in terms of sustainability and affordability.
If I had to choose one word for 2010 it would be Stewardship.
Stewardship made possible by open platforms on reliable infrastructure.
Have a great 2010, everyone. Especially if you got this far reading this post.
My thanks again to Orin Zebest for all the photographs, provided via Flickr on a Creative Commons Attribution Licence. Orin, you’re Ze Best.
Which brings me all the way back to Bob Frankston and the reason for this post. Bob’s always drilled into my head the concept that the addresses and numbers we use should never be considered routing; instead, I should consider such things to be nothing more than hints, clues as to the best way to get something to someone. Reading about SendSocial reminded me about his dicta, with their focus on getting things from person to person without an address.