A practical model for analyzing long tails

First Monday

Volumn 12, Issue 5, 2007, Pages

  Kilkki, Kalevi ^a  

^a NOKIA SIEMENS NETWORKS   (Finland)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 34248174776     PISSN: 13960466     EISSN: None     Source Type: Journal    
DOI: 10.5210/fm.v12i5.1832     Document Type: Article

References (24)

1. C. Anderson, 2004. "The Long Tail," Wired, volume 12, number 10 (October), at http://www.wired.com/wired/archive/12.10/tail.html, accessed 9 October 2006.
2. C. Anderson, 2006. The Long Tail. London: Random House.
3. The phrase "long tail" is not a new term. It has been used in statistics to describe distributions that have essentially thicker tail than exponential distribution, see e.g., http://en.wikipedia.org/wiki/ Long_tail. Still, without a doubt, the long tail concept has become much more popular after the article by C. Anderson.
4. As Stephen Hawking expressed it in A Brief History of Time: "Someone told me that each equation I included in the book would halve the sales." Anderson's book has only one equation and it was 47th in the list of Amazon.com's most popular books in August 2006 (http://www.amazon.com/gp/ bestsellers/).
5. The data is taken from the long tail book p. 121, and is originally from Book Industry Data Group and obviously refers to book sales in the United States.
6. See e.g. the analysis by Morris Rosenthal at http://www. fonerbooks.com/surfing.htm, accessed 9 October 2006.
7. The data was obtained from http://www.accessabc.com/reader/top150. htm, accessed September 2006.
8. Statistics are available at http://www.ses.fi/fi/tilastot.asp, under topic: "Vuonna 2003 ensi-illan saaneet elokuvat. (pdf)," accessed 9 October 2006.
9. The statistics are available at http://www.movieweb.com/movies/ boxoffice/alltime.php?q=&page=1. The presented results were accessed in August 2006.
10. Last.fm statistics are available at http://www.last.fm/charts/.
11. The data was obtained from http://www.last.fm/charts/music/artist/, accessed September 2006.
12. The data was from http://www.last.fm/charts/music/tag, accessed 7 September 2006.
13. The data was obtained from http://en.wiktionary.org/wiki/Wiktionary: Frequency_list, most common words (TV and movie scripts), accessed in September 2006.
14. The data was obtained from "Wigblog - Things Internet and Otherwise" by Richard Wiggins, at http://wigblog.blogspot.com/2005/08/ long-tail-and-short-head-of-zipf-curve.html, accessed in 9 October 2006.
15. The data about Google hits (or results) was gathered in August and September 2006 using the Finnish language version of Google at http://www.google.fi/, The number of hits somewhat depends on the selected language. There also are variations between different dates that seem to be quite random.
16. The data was obtained using Google at http://www.google.fi/, accessed in September 2006.
17. The data was obtained using Yahoo at http://www.yahoo.com/, accessed in September 2006.
18. The data was obtained using MSN Search at http://www.live.com/? searchony=true, accessed in September 2006.
19. The data was obtained from http://www.census.gov/genealogy/names/ names_files.html, accessed in September 2006.
20. The data was obtained from http://www.census.gov/genealogy/names/ names_files.html. File: dist.female.first, 29-Sep-94, accessed in September 2006.
21. The data was obtained from http://money.cnn.com/magazines/fortune/ globa1500/2006/full_list/, accessed 9 October 2006.
22. The data was published by Talouselämä magazine. Data is available at http://www.talouselama.fi/te500list.te, accessed in September 2006.
23. The data has been provided by Statistics Finland, http://www.stat.fi/ index_ex.html.
24. This choice of including the volumes with k < 1 in the volume of the volume of most popular object raises an interesting philosophical question of what are those objects with a rank of below 1 in the theoretical model. We may, for instance, say they are those objects that the audience is still waiting for. Because that need does not yet have any concrete target, the most popular object gets extra interest. Regardless of the speculativeness of the idea, the model often works amazingly well.