Help -
Help for Webmasters
« back to results for ""
Below is a cache of http://www.fma.org/Orlando/Papers/GreatDataMining.pdf. It's a snapshot of the page taken as our search engine crawled the Web.
The web site itself may have changed. You can check the current page or check for previous versions at the Internet Archive. Yahoo! is not affiliated with the authors of this page or responsible for its content. Good to Great , or Great Data Mining? Good to Great, or Great Data Mining?
Kristine Beck Associate Professor of Finance University of Wisconsin Oshkosh 800 Algoma Blvd. Oshkosh, Wisconsin, 54901 beck@uwosh.edu
Bruce Niendorf Associate Professor of Finance University of Wisconsin Oshkosh niendorf@uwosh.edu
September, 2007 Good to Great, or Great Data Mining?
Abstract The primary objective of this paper is to discuss the dangers of data mining using the book Good
to Great as an example. Good to Great has been on Business Week</i>s business best-seller list
since its October, 2001 release. This study considers the validity and generalizability of author
Jim Collins conclusions concerning the greatness of the eleven companies identified in the
book. The second objective is to examine the impact of financial risk on the returns of the
eleven Good to Great firms. Although Collins and his research staff undertook a rigorous
screening process to identify the best firms, their selection criteria omitted a critical variable:
risk. The third objective is to offer a homework or classroom exercise in which students can test
whether risk-adjusted stock returns outperform a market benchmark. This exercise also can be
used to test the performance of the Good to Great firms in other time periods.
Good to Great, or Great Data Mining?
INTRODUCTION Jim Collins has written two business management books and numerous articles that have received much acclaim in recent years. Built to Last has been on various national bestseller lists for more than five years, and Good to Great has been on the business best-seller lists of USA Today, Business Week, and the Wall Street Journal since its October, 2001 publication. Good to Great (GTG hereafter) is now in its 75 th printing, with 2.5 million copies in print. GTG has become popular in the boardroom, and is being used extensively in classrooms across the country. This popular management tome has inspired millions of managers and has inspired author Jim Collins to rework his ideas for the non-profit and governmental sectors. In Good to Great, Collins identified a set of firms that made the leap to business greatness and sustained those results for at least 15 years. One of the criteria he used was that the firms had generated cumulative stock returns at least three times as large as those generated by the general stock market over 15 years. Other screens included firms that outperformed their industry as well as the general stock market, and firms that were well established (not start-ups). Finally, the firms were required to be on an upward trend at the end of the 15 years. Collins then contrasted these firms with a set of comparison companies that failed to make the leap from good to great. He looked for commonalities among the firms that distinguished them from the comparison firms, identified five management practices common to these firms, and concluded that these practices lead to greatness. The eleven companies identified in GTG are presented below:
1
Good to Great Companies Abbott
Altria (formerly Philip Morris)
Circuit City
Federal National Mortgage Association (Fannie Mae)
Gillette
Kimberly-Clark
Kroger
Nucor
Pitney Bowes
Walgreens
Wells Fargo We question the validity of the process Collins used to select the firms and identify the practices that lead to greatness. First, Collins technique identifies firms on the basis of sustained high stock returns (in addition to other criteria), and observes five characteristics common to these firms, but this process does not prove that these are the characteristics that made the firms great. He may have established a correlation between the five characteristics and greatness, but he has not established the causation he claims between the five management tenets and greatness. Second, although Collins used somewhat extensive screens in selecting the companies, he did not screen for one critical factor: financial risk. Without considering risk, Collins cannot conclude whether or not the shareholders of the firms were adequately compensated for the risk undertaken. If not, then it may well be that the eleven firms he identifies as great, out of a population of thousands, were just those that got lucky, not necessarily those that are run by skillful managers. What is Data Mining? Data mining also is called data dredging, data grubbing, or fishing in statistics. It is the act of trying to find patterns in data by trawling through information. This process is regarded as 2
productive in marketing and is often referenced with the euphemism knowledge discovery. For statistical inference, however, data mining is used as a derogatory term. It has a derogatory connotation because a sufficiently exhaustive search will
certainly throw up patterns of some kind by definition data that are not simply
uniform have differences which can be interpreted as patterns. The trouble is that
many of these patterns will simply be a product of random fluctuations, and will
not represent any underlying structure. The object of data analysis is not to model
the fleeting random patterns of the moment, but to model the underlying
structures which give rise to consistent and replicable patterns. To statisticians,
then, the term data mining conveys the sense of na
The web site itself may have changed. You can check the current page or check for previous versions at the Internet Archive. Yahoo! is not affiliated with the authors of this page or responsible for its content. Good to Great , or Great Data Mining? Good to Great, or Great Data Mining?
Kristine Beck Associate Professor of Finance University of Wisconsin Oshkosh 800 Algoma Blvd. Oshkosh, Wisconsin, 54901 beck@uwosh.edu
Bruce Niendorf Associate Professor of Finance University of Wisconsin Oshkosh niendorf@uwosh.edu
September, 2007 Good to Great, or Great Data Mining?
Abstract The primary objective of this paper is to discuss the dangers of data mining using the book Good
to Great as an example. Good to Great has been on Business Week</i>s business best-seller list
since its October, 2001 release. This study considers the validity and generalizability of author
Jim Collins conclusions concerning the greatness of the eleven companies identified in the
book. The second objective is to examine the impact of financial risk on the returns of the
eleven Good to Great firms. Although Collins and his research staff undertook a rigorous
screening process to identify the best firms, their selection criteria omitted a critical variable:
risk. The third objective is to offer a homework or classroom exercise in which students can test
whether risk-adjusted stock returns outperform a market benchmark. This exercise also can be
used to test the performance of the Good to Great firms in other time periods.
Good to Great, or Great Data Mining?
INTRODUCTION Jim Collins has written two business management books and numerous articles that have received much acclaim in recent years. Built to Last has been on various national bestseller lists for more than five years, and Good to Great has been on the business best-seller lists of USA Today, Business Week, and the Wall Street Journal since its October, 2001 publication. Good to Great (GTG hereafter) is now in its 75 th printing, with 2.5 million copies in print. GTG has become popular in the boardroom, and is being used extensively in classrooms across the country. This popular management tome has inspired millions of managers and has inspired author Jim Collins to rework his ideas for the non-profit and governmental sectors. In Good to Great, Collins identified a set of firms that made the leap to business greatness and sustained those results for at least 15 years. One of the criteria he used was that the firms had generated cumulative stock returns at least three times as large as those generated by the general stock market over 15 years. Other screens included firms that outperformed their industry as well as the general stock market, and firms that were well established (not start-ups). Finally, the firms were required to be on an upward trend at the end of the 15 years. Collins then contrasted these firms with a set of comparison companies that failed to make the leap from good to great. He looked for commonalities among the firms that distinguished them from the comparison firms, identified five management practices common to these firms, and concluded that these practices lead to greatness. The eleven companies identified in GTG are presented below:
1
Good to Great Companies Abbott
Altria (formerly Philip Morris)
Circuit City
Federal National Mortgage Association (Fannie Mae)
Gillette
Kimberly-Clark
Kroger
Nucor
Pitney Bowes
Walgreens
Wells Fargo We question the validity of the process Collins used to select the firms and identify the practices that lead to greatness. First, Collins technique identifies firms on the basis of sustained high stock returns (in addition to other criteria), and observes five characteristics common to these firms, but this process does not prove that these are the characteristics that made the firms great. He may have established a correlation between the five characteristics and greatness, but he has not established the causation he claims between the five management tenets and greatness. Second, although Collins used somewhat extensive screens in selecting the companies, he did not screen for one critical factor: financial risk. Without considering risk, Collins cannot conclude whether or not the shareholders of the firms were adequately compensated for the risk undertaken. If not, then it may well be that the eleven firms he identifies as great, out of a population of thousands, were just those that got lucky, not necessarily those that are run by skillful managers. What is Data Mining? Data mining also is called data dredging, data grubbing, or fishing in statistics. It is the act of trying to find patterns in data by trawling through information. This process is regarded as 2
productive in marketing and is often referenced with the euphemism knowledge discovery. For statistical inference, however, data mining is used as a derogatory term. It has a derogatory connotation because a sufficiently exhaustive search will
certainly throw up patterns of some kind by definition data that are not simply
uniform have differences which can be interpreted as patterns. The trouble is that
many of these patterns will simply be a product of random fluctuations, and will
not represent any underlying structure. The object of data analysis is not to model
the fleeting random patterns of the moment, but to model the underlying
structures which give rise to consistent and replicable patterns. To statisticians,
then, the term data mining conveys the sense of na
