I, for one, bought a lot of music in my late teens and early twenties, and now I am content to listen to my collection, the radio, and the occasional purchase. Some people do buy music in a fairly regular way throughout their lives, but judging by the actions of the studios, they at least believe that many people spend like me. [Hans Fugal, writing in the comments to a previous post on this blog.]
This brings up an important flaw in all current music subscription efforts I know of. And that is that they are all one-size-fits all. Hans, and many others who feel the same way he does, may really have no need to spent $180/year on a music subscription going forward, because they just don't care to spend that much on music. They are happy with a relatively small collection of owned music plus the radio.
So, it's true that today, one-at-a-time purchase models have the advantage that you have complete control over what you spend in a given year. And for many people, that is indeed a huge advantage over current subscription offerings. But there is no reason whatsoever to believe that subscription services won't do something equivalent. That seeming limitation of subscription services is in fact only an artifact of the reality that they are still in their infancy and have yet to display the features that they will inevitably acquire in time.
For instance, one can envision a subscription service with bronze, silver, and gold levels. The gold level might work as Napster To Go does now: you can listen to as many songs as you want in their collection per year.
The silver level might be: You can listen to all songs from 200 different artists in a given month.
The bronze model might be: You can listen to all songs from 50 different artists in a given month.
Now suppose you're someone like Hans, and you're happy with an occasional purchase plus listening to the radio. Then you really don't need to listen to 200 artists every month in addition to what you year on the radio. You probably don't spend enormous amounts of time listening to music; it's usually more of background thing.
Now consider what your experience would be with the silver or bronze model. You don't have to commit to the specific artists in advance as you would if you were buying CD's. Rather, you listen to whatever you want when you want to. But, your listening habits are such that you don't listen that much, so you don't hit the limit of artists you can listen to in a given month. In other words, experientially, your bronze or silver subscription would be identical to what your experience would have been if you had chosen the gold level.
The only real difference would be that it would cost less. The subscription vendor could charge much less for the silver and bronze levels than for the gold level because you would not be using the same degree of "music resources" (i.e. engendered copyright fees) as a gold-level subscriber would.
If your listening habits changed, and you wanted to bump your level up, that would be fine, you could do so at any time.
Eventually some rough equivalent to the levels described above will be part of the offerings of any competitive subscription service.
An advantage over buying albums is that it realistically addresses the facts that a) tastes change, and b) people get tired of albums they've heard too many times before, and so their frequency of listening to even albums that are still well-loved therefore decreases over time.
A purchased album is a binary choice that says that you want infinite access to that particular album forever. The money spent on it can no longer be spent on other music that you might prefer hearing in the future. But that's not a good match to actual listening habits. In reality, that money should ideally be divided up between the music you want to hear now, and the different music you'll want to hear in the future.
That's the real difference between a subscription service and buying albums.
Let's "Do the math" again, considering Napster To Go to be the gold-level service. It would take 55 years of paying for Napster To Go at $180/yr to spend the $10,000 it would take to buy 10,000 iTMS songs to fill up an iPod. Let's consider that to be your "lifetime music cost".
The cost is the same $10,000 either way. But with Napster To Go, that cost is distributed between the songs you want to year at various stages of your life. With iTMS, it is 100% spent on those 10,000 songs, period.
(Note that with something like iTMS, you are not only committing to the particular songs you buy, but also to their not-great 128kbs sound quality, which will be considered very low as bandwidth resources increase.)
Now consider the silver and bronze levels. With those levels, you might still end up listening to many more than 10,000 songs whenever you want in the course of your life -- but you don't have to spend nearly so much because your listening habits don't require you to have access to thousands of different artists in each month of that life.
So, while subscription services now appear to be too expensive relative to the listening habits of people such as Hans, we can expect that to change in the future. The current advantage is simply because subscriptions services are very new, and getting them working at all with a broad catalog has been a major achievement in the negotiations with the major labels. One of the next steps will be offering different levels of service.