I’m reading The Long Tail (the book, not the article), and something struck me. On p. 119, Anderson writes:

As the Tail gets longer, the signal-to-noise ratio gets worse. Thus, the only way a consumer can maintain a consistently good enough signal to find what he or she wants is if the filters get increasingly powerful.

I wonder if this is actually true. It seems to me that a significant flaw in Anderson’s argument is that he conflates the mathematical power curve with the economic Long Tail. Not everywhere, but pretty consistently. A power curve is infinite. The Long Tail of products and consumption is finite, limited by the amount of stuff that’s available and the number of people available to consume it and the amount of time and money those people have to spend. All of those pools are large, but finite. Anderson writes about abundance, and how that’s changing economics. And I agree with him there: decreasing costs and risk in creation, dissemination, and selection of stuff will, I believe, deeply change how markets operate. Are in fact already, and will continue to do so. But the Long Tail of economics does not imply that there is suddenly an infinite supply of everything, nor an infinite number of consumers for it. Under ideal conditions, with free tools to facilitate creation and a completely frictionless market, even then, the Long Tail is merely unlimited, it is not infinite.

So, my question. Given that the Long Tail is finite (if potentially very very long), is it really true that filters must forever increase in power? Way way way down the tail, where there are very very narrow niches, will there ever be enough stuff to make a filter necessary? Of course I suppose this will change over time: as more stuff is developed, the need for filters increases. So I suppose my question is, where is the tipping point, so to speak, at which a filter becomes necessary in a narrow market sector? Or, since this is probably not a question solely of the amount of stuff available, what factors contribute to the need for filters in a narrow market sector?