CSC2231 - The long tail review

The Long Tail
--------------
Chris Anderson

I find this paper to be a good discussion on physical vs online sales.
The paper emphasizes the power of the long tail in online distribution
and retail and the fact that there is no scarcity here like in the
hit-driven economy.

Since there is no shelf-fee, limitations of geography or scale, online
retailers get their success from the power of the long tail and the way
they manage to use this power. The paper states three rules that help a
company take advantage of the long tail: making everything available,
cutting the price in half and lowering it, pointing down in the long tail
at searches (e.g Rhapsody has for each artist a "similar artists" feature
that are artists lower on the long tail than the targeted one).

One thing I don't get is why the titles ranked by popularity graph has
a zipf-like behavior. Like in P2P systems (such as Kazaa, analyzed in the
other paper), the fetch-at-most-once rule applies here also.

We can put the discussion into a P2P environment, although the content
popularity may not be exactly the same (may not follow the zipf curve
from this paper). The "lot of crap" that the paper mentions can be the
unpopular content that noone downloads. I guess 2 of the three rules
presented in the paper can also aply to P2P. Making everything available
is the first one. The issue here is the long tail since the popular
content gets replicated by default. Like most successful bussiness on
the Internet a strong P2P system should be designed with the long tail
in mind. The second rule is the "help me find it" one. What Rhapsody and
Amazon do could be used in a P2P architecture.
Received on Mon Nov 28 2005 - 10:38:31 EST