Thursday, July 21, 2005

Network effects...

Windows is the best operating system in the world.
Windows is the most useful operating system in the world.

Now the geeks among you are saying WTF!!!

I HATE windows, it's a piece of junk.

Guess what, so is Linux.

So is OS X.

They ALL SUCK!

But Windows sucks the least for the most amount of people.

The issue at hand here is network effects.

In a market like computing, unless all vendors implement standards interchangeably (which will allow forcommoditization), then a single vendor will always develop an advantage. Because a vendor does not wish to be commoditized, unless they are competing on price alone, they will NOT willingly implement full standards; thus preserving their advantages. These advantage will result in further adoption of their product over other products, which increases their advantage, because no more people have access to identical experience/advantages. This will generally happen with the product that has the most advantages, for the most amount of people (more on this later)

At this point a network effect kicks in. The more individual users using a product, the more value that product has to the individual user. This gives individual users the incentive to switch to that dominant product.

Generally speaking, a viable, but not dominant minority product will rise in popularity, in direct opposition to the dominant product, with niche products developed to meet special needs.

This happens because of that “most advantages for the most people” issue. Yes, one dominant product can server MOST people needs, but the ones who were left out want OTHER advantages. This results in a strong minority product, and multiple niche products, but combined they will almost never achieve more than 25% or so market share (except in certain special circumstances or markets), because of the large network effect of the existing dominant product, which in and of itself has become the biggest advantage for the most people.

Markets without a dominant product tend to stagnate, or fluctuate wildly, unless all products submit to commoditization. This is due to consumer uncertainty, and competing standards. It is only once a defacto standard is estabished that markets tend to exhibit consistent growth, or measurable market patterns.

The "problem" in the OS market, is that this strong minority competitor has not happened with Windows, because the alternatives have been fragmentary and unfocused, or simply not viable. Now, the network effect for Windows is so large at this point, that only a needs based solution will use anything different.

Of course this excludes minority user populations like geeks and hackers (the good kind), who will use a technology for it’s technical superiority. They are by definition a niche market.

The only way this will change, is if Microsoft is unable to satisfy their large user bases needs to a sufficient degree. It doesn't depend on technical superiority, because the network effect is its own inbuilt technical advantage. The degree of technical inferiority required for a network effect to be negated is VERY significant.

This for example is why until recently Macs were still dominant in the graphics design world, and why they have lost that dominance.

For a long time Microsoft was not able to provide a viable solution to designers, who built a strong network with their Macs. This network effect grew stronger as more designers switched to the Mac platform, or made further investments in their already established Macs.

The problem is, as the MS solution got better and better, Apples solution did not improve at all. In fact relative to PCs (which wer eadvancing rapidly), apples solution worsened significantly.

Eventually, rapidly improving PC hardware, combined with a 3 year stagnation in Apples hardware capabilities, the near parity of software with Apple, and the compliance to standards of both major platforms; overcame this strong network effect, and now Macs are being replaced more and more by PCs; some of which run windows, and some runninglinux (for rendering and image processing).

This could easily happen with Windows (and even easier with Internet Explorere for example. Non I.E. browsers already have between 10 and 20% of the market, and there share is growing rapidly).

As more and more services and systems become networked, embedded, and appliance-ized (ugly pseudo-word that), the desktop operating system becomes less an independent operating environment, and more a network hub for managing the input and output of these various devices and services.

This has MS scared stupid, because the network effect works both ways, and is NOT transitive across needs.

These systems and services are standardizing on non-proprietary formats, or on proprietary formats which MS does not control; which means to enter these markets MS has to implement these standards, or to attemtp to establish a network effect with a competing standard, and other competitors have already established strong network effects and standards control in these markets.

MS is trying as hard as possible, as fast as possible, to build strong positions in all these secondary markets, to allow network effects to build. For the most part they are failing. Yes they make huge announcements and push big programs, but basically they are going nowhere. The ones that ARE viable, are moving to more standards compliance with the rest of the market so they can increase their market share.

This has Bill Gates personally frightened. No, I’m serious. Bill is PARANOID about competition and failure.

The next few years should be interesting

Oh and in case anyone missed it, we just had a discussion about market capitalism and microeconomics; not technology.

Now heres a fun thought experiment for you, especially to libertarians out there. Replace "linux" with "libertarian"... e.g. Apply network effect economics as I described above to a multiparty political market with a "first past the post" electoral system.

Yup, I'm an evil bastard.