Sunday, January 18, 2004

Telirati Newsletter #35

Here I make a singularly safe prediction: Unless AOL/Netscape plays to win, they should just give up. Subsequently, the collapse of that entire "threat" to Microsoft was more complete than Bill Gates's fondest hopes. Even if the illustration turns out to lack suspense, the lesson remains: Don't take on large problems with small plans.

Telirati Newsletter #35: New horizons in competition

There is a theory I developed as I plied my craft early in my career: Most management failures are a result of insufficient radicalism. I was a hothead, a trouble-maker. I was always being told that, in the “real world” you had to be patient and that people could not digest too much new technology at once. You can see this mind set writ large if you pick up any trade publication oriented to corporate data processing: “Why do they keep revising software?” is a persistent theme of columnists in those pages. Those bad, bad, software people just do it to make problems for the decent folk that help keep computers running. What is really wrong with the world is that software changes too slowly.

Consider hardware: having crashed through the $1000 barrier, computers plummet toward the frightening figure of $500 per. Hardware doesn’t have to change, it just has to scale. (The fact that all the pieces, from CPU, to ram, to rotating memory, to CRTs, to printers have all kept pace with this same plunging price curve is another marvel which will have to get it’s own dedicated newsletter topic.) So hardware can change the economics of operating computers, change the kinds of applications everyone can use, change everything without themselves changing in any fundamental, architectural way.

No so with software. Make a word processor four times as fast and lo, I still type like leper that dropped a few digits. In order to take advantage of the new hardware, and the new economics of hardware, software has to change, deep down, architecturally, fundamentally. This happens, but it happens slowly compared with the changes in hardware price and performance. There are effects such as enabling nonlinear video editing for the masses that cause software to be pulled into the mainstream as hardware becomes bigger cheaper faster, but, in general, the software that runs adequately on today’s machine will not change your life on a new machine. Software, as an industry does not scale, it deforms.

This means that the nature of competition in software does not scale, it shifts. Which brings us to our example: AOL/Netscape is a potential competitor to Microsoft not just in competition with MSN, but also in competition with the way software is delivered. Even more fundamentally, AOL/Netscape can change what people expect to pay for software and how they expect to pay it. Mark this then: Microsoft’s most important software product is not Windows 2000, or Office 2000. It is Windows Update.

Given this potential for tectonic shifts, how can the DoJ build a case against Microsoft based on an economic theory that says Microsoft can build on its advantages? Every company can build on its advantages. If AOL/Netscape fail to change the software landscape it will be due to failure to build on the advantages of realizing early on that any level of investment is justified in pursuing dominance in Internet access. Every shift in the industry places new players on the field with interesting advantages. Execution, and not the advantages themselves is what makes the difference.

And if AOL/Netscape succeed what might be different? First ask, what is a client environment worth? Why is Windows worth its present price? What if the valuation of a company is based on something other than the price of a desktop OS. Why not give away a client environment? Radical? Someone was sufficiently radical to ask: “Why not give away a Web browsers? Why not integrate it tightly with the OS?” Any response would have to be sufficiently radical as well, and open source isn’t enough.

What are the tests for plausibility here? Size: A browser, a full suite of plug-ins, and a Java environment (especially a full Java 2.0 environment) is about the size of the rest of Windows98. Applications: Is the Internet not application enough? How many people who buy computers are retail stores need Office? Installed base: 15 million AOL users is larger than the Macintosh installed base.

All the ingredients are in place for making life interesting. The only question is whether AOL/Netscape has sufficient audacity. Will they see that, like the early stages of AOL, any level of investment is worth the prize? Can they get a hardware maker to market computers that don’t run Windows or find some other path to a native platform? Can they get their arms around all the potential changes to their business model such as challenge would entail? A winning formulation here could take a variety of forms. But it is certain that the cautious forms, like those pursued by Sun, cannot win.

One last question: Are we not glossing over the matter of an operating system here? A Web browser is fine, Java provides an API, AOL is a place to go and a means of connection, but what takes care of the hardware? The answer is that AOL/Netscape are spoiled for choice: Be, Linux, perhaps the carcass of OS/2 could be disinterred, QNX, etc., could all serve as an underlying layer.

If they don’t rise to the occasion, you heard it here first: It was not Microsoft, not the consumers, not the channel, not the Asian crisis, Y2K, or bad karma that prevent AOL/Netscape from mounting a credible challenge to Microsoft. If you are not radical enough, don’t take on big problems.

Copyright 1999 Zigurd Mednieks.