Chris Anderson, author of the Long Tail, writes about the Wharton team that analyzed the Netflix Prize data and discovered the following change to customer rental patterns over five years:
The Wharton researchers also disagree with Anderson's theory and its implicit challenge to the Pareto principle, or so-called 80-20 rule, which in this case would state that 20% of the movie titles generate 80% of sales. Anderson argues that as demand shifts down the tail, the effect would diminish. Using Netflix data, Netessine and Tan show the opposite -- an even stronger effect, with demand for the top 20% of movies increasing from 86% in 2000 to 90% in 2005.
Anderson did his own analysis using the Wharton data and found lower demand for the top 500 products and more interest in the middle part of the curve. He also points out that a Long Tail adding up to 15% of total demand came from titles beyond the top 3,000 -- the amount typically stocked in a video store.
I'd agree with the long tail, when members first join they want all those new releases but over time their interest in 'middle' and 'long' tail increases as they discover more titles...
Posted by: shetaan819 | September 29, 2009 at 01:13 PM
"Anderson did his own analysis"
This chart has a mathematical defect. You can't compare trends based on a fixed unit of 500 top movies. As titles are added to the Netflix library, you will begin to see distortion that has nothing to do with demand. You have to use a percentage of top movies.
For example, if Netflix had 10 movies in their library and viewers were most interested in the top 5... and then Netflix added 10 more titles totaling 20, demand for the top 5 would appear to go down.
Posted by: Rube | September 29, 2009 at 08:28 PM
What appears to be missing is a control group that could get all the New Releases that they wanted. I wonder what would happen to the 'long tail' for the New Releases control group?
Posted by: Edward R Murrow | September 29, 2009 at 09:13 PM