Stardate 20020110.2243

I've long since concluded that one of the differences between Mac aficionados and the rest of the world is that Mac people simply think about their computers in different ways than, well, geeks. A fine example of this is the contrast between James Lileks' answers to readers' comments about his thoughts on the new iMac (scroll down on both pages) and comments by Steven den Beste (one and two). Den Beste is all over the technological problems; Lileks is all over the user experience. While den Beste may well be right about some of the technical questions, he doesn't get where Apple seems to be going.

I am a geek, but I'm not being geekish about this. If I were reacting to the iMac as a geek, I'd be making fun of how silly it looked and making up disparaging nicknames for it like "iLamp" and "iSore". (Well, I admit that I did call it "a camel" and "a gargoyle"; chalk it up to geekish excess.)

I'm looking at Apple's recent products and general design strategies not from the point of view of an individual user (let alone an individual geek), but rather from the point of view of someone who's trying to predict where the company will end up in the long run. As such, it may well be the case that iPhoto is really cool (as James says, and I don't doubt it) but that may not actually be important in the long run.

There are larger issues involved. The question is whether Apple is in a hole (there's no question of it) and whether it can dig its way out. That latter does not have an obvious answer, and it's what I'm concentrating on. I am highly skeptical. While the company is in vastly better shape than it was four years ago, it faces enormous obstacles and I think some of them (e.g. Motorola) are insurmountable.

Some of my readers who have known me for years will recognize this as the kind of analysis I did about OS/2 starting in about 1995. Its fans were no happier about what I had to say than are Mac fans now, but I was right then, and I think I'm right this time, too: Apple is caught in a downward spiral and it's too late to escape. In ten years it will either be completely out of business, or will have switched to some other market. The Macintosh is doomed; it will follow the Amiga into the pages of history.

There is a marketing term: "network effect". It refers to the fact that for some kinds of products, the product becomes more valuable to each individual customer as more and more people buy it. Companies love this if they can get it, because there's no manufacturing cost associated with network effect, so the value (and potentially the sales price) of the product can rise and thus profit margins can increase.

The best example of this is the telephone: your phone is more valuable to you if nearly everyone has a phone and can be called than it would be if only 5% of the population had one.

In the case of telephones there is a literal network, but the marketing term "network effect" doesn't require that. For example, there's a minor network effect associated with car brands. If a car brand is popular, then there will be ample supplies of spare parts for it and many places capable of doing repairs. If a car is rare, then spare parts will be hard to come by and skilled repair personnel will be harder to find. So popular car brands are cheaper to maintain.

It turns out that computer architectures are particularly susceptible to network effect. Users of niche architectures often refer to those who use the majority platform as "sheep" and denigrate them as being cowards who only find safety in numbers or who are too stupid to think for themselves and just go along with the crowd. As such they completely miss the point that there is value in numbers for the users themselves.

It has to do with vendor psychology. The cost to develop a given software package for a given platform is not related to the number of potential customers. When a software developer makes its business plans, it balances the potential expense of developing a product against the expected sales, and then does some arcane calculations to determine the probability that the product will be profitable. There are always more opportunities than there are resources to apply, and so unless the company is mismanaged it will do this analysis for all its alternatives and then choose the ones among them which have the best chance of making money. The rest won't get done.

So suppose that the company has enough money and programmers to create software for one platform. It could be the Amiga, the Mac, or Windows. The cost of doing each version is about the same (within 25%), but the potential sales of the three is radically different because the user bases are radically different. At this point there may not even be half a million Amiga users left. Most apps are lucky to make half a percent market penetration. (Sometimes the penetration is vastly lower than that.) If you figure half a percent of half a million users then you've got 2500 copies sold. That isn't very many. If the product costs you $2 million to develop, you're going to lose your shirt.