April 02, 2004

Of Music and Divx

CNet is reporting that Microsoft is looking to get into the music rental business.

I must say, I simply don't understand their business plan.

The iPod and Its Importance

Apple's iPod is big news in the computer industry. I see it as important on multiple fronts:

1. It's a tremendous hit, and the first really successfully computer-industry-makes-a-consumer-appliance story. With so many tech companies losing money right now, there's a lot of hope that companies can start moving into the consumer appliance category of the market to reclaim some of the sales. You can already see this with Gateway selling plasma TVs, Dell selling LCD TVs, and so forth.

2. The iPod's integration with the iTunes Music Store is the first real success story to come from digital music. When Time names you the Coolest Invention of the Year, you know you're on to something big. But it goes beyond music. The iTMS is the first successful online media distribution system. Music was a logical first step because of the relative small size of the files. But companies are eagerly watching, with dreams of selling video and other media in a similar fashion.

3. It reinforced the notion that design matters. This is all the more true with appliances, laptops, and other items you physically interact with, compared with a desktop computer that sits on or beside your desk. When you add in syncing your music or buying from the iTMS, this becomes simplicity matters. Downloading a song from the music store and getting it to your iPod is about as painless a process as anything the computer industry has ever done. This one will factor largely into my thoughts below.

Enter Microsoft

Microsoft, unsurprisingly, wants a piece of the pie. But there's a new angle: they apparently want to enable music rentals on portable devices. This recalls memories of Divx, the company that wanted to revolutionize video rentals back in the mid-nineties, and failed spectacularly.

Call me biased, but I just don't see the appeal here.

As I see it, the point of a music rental service is the ability to have access to any song; you have access to a much, much more broad library of music than you could ever afford to purchase. Seems like a great thing, right? But without a persistent network connection on the music player device, you can't offer this. It's simply not possible. Perhaps you could offer a reasonable approximation, say, by having it download some large subset of songs on every sync maybe having it auto-download the top 100 songs in a genre, for example. But the bandwidth costs for such an ability would be tremendous! For $10 of profit, the iTunes Music Store has to send across 10 files, plus however many free previews. A $10 rental service, however, might have to send across thousands of full-length song files! From the server side, that's a lot less appetizing, as you'd be paying for the costs of sending hundreds of files the user might never even listen to!

Now, you might say, you can analyze their listening habits and learn what types of files to send, minimizing the "wasted" bandwidth. Or hope that my habits don't change much month-to-month. But that's just it! I might be in the minority here, but my habits change all the time. If I'm going out on the weekends with my friends a lot, I'll listen to a lot more party music. If I'm lovesick, I'll listen to that sort of music. When I'm driving home late at night and want to stay awake, I usually pop on a musical or something with a plot to keep my attention. When the new Dream Theater album comes out, I find myself going back and listening to all their old albums again.

This means that they'll be sending me an inordinate amount of music, much of which I'll never listen to, in the hopes that my experience is close to the "listen to any music I want" experience that makes subscription services appealing. All because the xPod is essentially a cache for their larger library, with the inability to get new music when it's the most important when I want to listen to it.

And then, of course, I have to deal with the actual downloading and syncing. Even if I changed just 100 songs in a given month, that's still going to take me over half an hour to sync, on my DSL line! This is getting less and less attractive.

I just don't see why people would want this, especially given the widespread of radio as a free "get a bunch of various music" service and the small expense of buying music relative to the number of times you listen to it. One of the main reasons Divx failed is because people perceived it as an assault on media ownership. I see this as similar, except that now there are even more technical hurdles.

Maybe once there are persistent, fast network connections everything this sort of thing could work better, provided it was a hybrid model where you could also buy certain songs/albums. But I certainly don't see the appeal right now.

This computer industy opinion was posted at 11:37 AM