The Innovation of Open vs Closed Systems on the Web

I had an interesting chat yesterday with a guy that thinks about the Web probably as much as I do (or more).  The subject of the conversation was Open vs. Closed systems, a very prevalent discussion right now.

On the Web, there seems to be different universes forming with different fundamental laws governing them.  The most prominent of these “universes” being Apple’s media ecosystem, the Twitter-verse, and Facebook.  People enter these universes and interact differently with the Web and other users.

Defining Open and Closed is a bit of a discussion in itself, but I’d describe Twitter as being completely Open.  The Twitter API is like electricity; Twitter powers any piece of software that is designed to use it. But, it provides such a fundamental functionality, it can power many things (just as electricity delivers energy).

Apple maintains a high level of control over their system through their vertical integration from hardware, software/media, to distribution. They have certain rules for what can be done on their devices.  You manage and obtain shows, music, and apps through iTunes.  This is a closed system.

The major question posed during this conversation was, “How do Open and Closed systems affect innovation”?

Here’s my answer after this conversation (thanks to Brad for framing it within the context of evolutionary processes):

Early in their formation, closed systems produce the most innovation and value. With a closed system, if the designer gets it right, clearly defined rules and uses allow people to immediately build within the system and harness it.  If the system really gets it right (and I think Apple did), it will be the optimal system at that point in time. Consequently, Apple’s universe exploded very quickly.

Early in the development of open systems, people aren’t really sure exactly what it really is for and what it can do. Twitter started as a micro-blogging system with an open API (other programs could interact with it), so you could post short messages like, “I’m eating soup.”  However eventually people started doing very interesting things with this system, people started to figure out what it really was, Twitter developed more capabilities, and people started better understanding the laws of the Twitter-verse and how to operate within this universe.  Then Twitter suddenly had a trove of openly available data.  This further enhanced the platform.  The platform (Twitter) and the users co-evolved.  It took a long time for this to come about.

Intelligent design of closed systems is much faster at producing growth, and optimal in the short term.  Thus, closed systems can really hit the nail on the head and deliver in a short burst.  However, they suffer long-term risk. Apple has to get things right since they impose constraints. They don’t leave it to others to find their way.

Open Systems don’t have constraints, and the agents within can pursue whatever they desire within a loose set of fundamental laws. Open systems evolve and take longer.  But in the long term, evolutionary processes are very powerful.  In the beginning things are a mess, but once people start to get it, and since people can pursue whatever they desire, the activity grows explosively. Consequently, innovation comes about by sheer brute force of having so much activity.  And when selective pressures come along, something will stand.  There is less long-term risk.

You could think of a closed system as an arrow. If it’s on the mark, it’s beautiful and elegant. But if it’s off target, it produces nothing.  You better have someone skilled shooting the arrow.

Open systems however are more like a shotgun, once the powder is ignited (think of an old gun that takes a long time to load), it fires, and hurls a mess of small projectiles at the target.  It’ll hit something.  Although it’s not nearly as pretty or elegant, almost anyone could hit the target to some degree.

Not a perfect analogy, but it’s the best I could do in a limited amount of time.  It’s difficult without having two similar systems (Apple and Twitter do different things).  Here’s one that could provide a model: Encylopedia Britannica and Wikipedia.  Once Wikipedia got going, it blew encyclopedias out of the water.

Thoughts? Clarifications for me?

-Kevin
5.7.2010