He cites some sales statistics from Amazon which indicate that by using things like customer suggestions, Amazon has been able to boost the sale of items that other online retailers don’t even have in their inventory to the point that these so called "long tail" items account for roughly 57% of Amazon’s sales.
Understanding what these "long tail" items are is a little tricky (at least for me at the moment). As I understand it, the reference to "long tail" is the reference to the large majority of markets on the Web that are not pursued because they are not focused on the more popular items. For example, the word ‘sex’ is the most popular word used on search engines and there is therefore quite a bit of attention paid to that term as a search time. The market for things sex-related is pretty big but there are millions of markets for millions of things that are not sex-related.
Put another way, there is a small number of markets that draw a significant number of people. One such market is for sex-related products and services. At the same time there is a large number of markets for other products and services that don’t draw large numbers of people per market but when you add up the interest in these less popular markets, you are talking about a much larger group of people who are largely untapped.
Using this graphic, you can see what he is talking about. The 97% part of the graph represents the 97% of searches conducted on Excite that Krauss and his colleagues couldn’t work out how to tap, back in the day. They focused on the 3% which also represents the most popular searches.
All in all, this is a very interesting post and certainly worthwhile paying attention to because it defies the overworked traditional models and is the key to a largely untapped and potentially very lucrative group of markets.
As an aside, Chris Anderson of Wired wrote about this in October 2004. Take a look at this article too.