Re: The long tail

Alexander Johannesen wrote:
>
> to look to the commercial world for examples of how libraries can
> improve), the lack of a long-tail would put you all out of work.

Well, not necessarily. Let's deconstruct what we mean about the 'long 
tail'.

First is the hypothesis that usage conforms to a power-law 
distribution.  Certain items are used (purchased) many times, and other 
items are used/purchased much less, and the fall-off roughly corresponds 
to a power-law distribution.

If this isn't true for library materials, does it matter?  Maybe library 
material usage falls off less abruptly, and all library materials are 
used more or less equally. (Of course, this seems unlikely to be true). 
Would that hurt libraries? Nah.  Or let's say usage falls off even more 
abruptly than power-law, and the little used materials are hardly EVER 
used. Would that hurt libraries?  I don't think so, although public 
libraries are more likely to de-accession these materials -- as they 
always have.  We can look at the usage distribution of our materials by 
looking at circ statistics, and we in fact always have done this (with 
varying degrees of rigor, accuracy, or insightfulness), and public 
libraries especially always have used this to make accessioning and 
deaccessioning decisions.

The next hypothesis, and why the 'long tail' has become of trendy 
interest, is specifically about online electronic distribution. It's the 
hypothesis that as the _marginal cost_ of increasing your inventory 
decreases, you can increasingly make money off the long tail. While 
items at the end of the tail individually have very low usage/purchase, 
if you add them ALL up, you can make as much money as carrying the 
popular ones.  If the cost to carry (or publish) a million items isn't 
that much different than the cost to carry a thousand, then you can make 
money carrying that million items in the long tail.

Now that argument applies specifically to electronic items, and is 
related to the much cheaper marginal cost of electronic items. It's 
irrelevant to libraries traditional print business, and may be 
irrelevant to libraries in general since libraries generally don't use a 
business model where they charge per-use.  It's relevant only to 
businesses that make money (or recover cost) per-use.  And it's also 
specifically about electronic items.

So the article cited suggests that this isn't true, that the usage curve 
falls off even more drastically than power-law distribution, that the 
marginal costs are non-zero, and thus it may not be as easy as some 
enthusiasts think to make money from the 'long tail'.

It is not obvious to me how this would effect libraries.  And Alex's 
supposition that it would "put us all out of work" is _definitely_ not 
obvious to me. In fact, it might be the other way around -- libraries 
traditionally exist in part to serve a mission of distributing that 
"long tail" _despite_ the lack of commercial viability in distributing 
it. The "long tail" argument is that it will soon be commercially viable 
to distribute that long tail.  If that's wrong -- well, business as 
usual for libraries. If that's _right_ , THAT might be more of a game 
changer for libraries!  But either way, the impact or significance on 
libraries is not entirely obvious to me, it requires thought, and more 
careful analysis of our own usage statistics.  Which is definitely a 
useful thing to do, I'm sure we can learn things about how to serve our 
users better by analyzing usage statistics (the little used 'long tail' 
stuff is an obvious choice for moving to off-site storage or 
de-accessioning, depending on the mission and resources of the library 
involved, for instance). 

Jonathan