292 of 327 people found the following review helpful
Is it a new brand of cereal? Or maybe it's a granola bar, or a chunky peanut butter spread? Then again, could it be the latest infomercial exercise device designed to give you the six pack abs you've always dreamed of but know in your heart of hearts you'll never achieve? Actually, it's a book - the title a product of the very methods the book describes. Here's what SUPER CRUNCHERS says.
(1) Mathematical regression models generated from large datasets often generate better predictions than human experts, and they provide supporting information on the predictive weight and reliability of each explanatory variable.
(2) Well-crafted experiments using randomized trials and control groups provide good market research and behavioral analysis results.
(3) Technological advances - the Internet, massive data storage devices, rapid computation, broadband telecommunication - are making it possible to share more sources of information and create ever-larger databases for analysis.
(4) Today's companies engage in multiple forms of market research by creating and using large databases and large-scale randomized trials.
(5) Many phenomena conform to normal distributions in which 95% of the population will be found within two standard deviations of the mean, the5% balance generally divided evenly in the two tails.
That's it. I just saved you $25.00 U.S. and a half-dozen or more hours learning how a guy from Yale named Ian Ayres collected a bit of information about applied mathematical techniques that have been in practical use for decades, packaged them up, palmed them off as something new, and cooked up the ridiculous name Super Crunching to describe an ostensibly new technological development. Yet "Super Crunching" is nothing more than the author's marketing hype for a couple of standard mathematical methodologies, a creation of nothing from something. There's no new breakthrough here, no new paradigm.
Yes, the anecdotal information about the future prices of wine vintages, Capital One's teaser offerings, and evidence-based medical diagnosis are interesting (hence the two stars rating). The rest, however, is neither prescriptive nor sufficiently critically analytical. Should we go out shopping for a Super Cruncher tomorrow? Should we delight in the increased accuracy of data-driven modeling and prediction, or should we fear the implied manipulation of our desires and the incessant, single-minded drive toward maximum profit at the expense of creativity? Do we really want movies and books to be developed from mathematical models like Epagogix? Do we really want our every keystroke on the Internet to be fodder for market research that manipulates us in response? John Kenneth Galbraith, among others, warned of exogenous, manufactured demand decades ago.
SUPER CRUNCHERS is part business tome, part econometric paean, and part sociology book, but not fully any of the three. No matter how many time the author uses words like "cool" and "humongous" and "amazing," it's still regrettably a "No Sale" even for someone like me who enjoys reading about applied mathematics.
247 of 301 people found the following review helpful
on October 6, 2007
I read a blurb on this book in the Economist and bought it for that reason. When I read it however, it failed to deliver. It is similar to the Tom Peter's "Search of Excellence" type book with anecdotal stories with little substance. It is overgeneralized and overhypes the models it discusses. The models Ayres discusses are also NOT NEW. I personnally have been creating these types of system for nearly 30 years. What has changed over the years, of course, is greater accessibility of data and a greater capacity to process that data economically. But we still struggle with quality of data issues and appropriateness of model issues -- especially when the models begin to be used by people other than the model creators. The book glosses over this, only providing an example of how Choicepoint used a poor matching algorithm when eliminating felons from Florida's voting roles and even then the author minimizes the problem.
There is no discussion of how these models become abused when implemented as tools where the user of the tool has no knowledge of its limitations, when the model provides suboptimal solutions or what "outliers" are and how to deal with them (although you know immediately when you ARE the outlier and are trapped dealing with a company using a model designed for a population you don't belong to).
This leads us to becoming a nation of people who read off a screen and do what the computer says to do, while turning off our brain. Any wonder you can get outsourced in that scenario? But it must be right -- we Super Crunched it!
55 of 67 people found the following review helpful
on January 3, 2008
Like "Freakonomics," this book over-relies on a catchy phrase as a substitute for a thorough exploration of the concepts and issues. The list of concerns includes:
1. Vague definition of the term "supercrunching." Is it "super" because the author wants us to think all statistics are super, or (what I had hoped) is there something about the type of statistics to which he refers that are in fact different from statistics in decision making for the last 40 years? All the talk of large datasets implies that supercrunching is a matter of size, but then why does the very first example of regression involve a model that has only 2 predictors? No need for large data sets for this kind of a model, right? Unless the effect size is tiny, but then, what good is the model? Tell us how things really are new and different now.
2. The book reads like a list of (mostly internet) companies and how fabulous and smart they are for using statistics. Actuarial science has been around for many, many years and again we see little discussion of how the actuarial tradition has become more available outside of the insurance industry. The whole book seems more like a stream of consciousness than an organized conceptual framework about the emergence of statistics as a guide to commercial, medical, and policy making over time.
3. While perhaps an excellent lawyer and professor, the author makes so many misleading or inaccurate remarks about statistics that it could be difficult for someone with a statistics background to enjoy the book. For example, regression is discussed as a technique that is different from the "randomized test," when in fact the randomized test is a design, and the regression (more commonly the "general linear model," including regression, analysis of variance, linear and structural modeling) is the inferential statistical technique used to evaluate the results of the test design. Early in the book, the author talks about how amazing regression is, and then gives and example of how a bank evaluates probability of future actions on the phone based on past behaviors on the phone. This very first example after introducing regression does not involve regression as a prediction technique, but rather actuarial base rates! In fact, I found it very disappointing that actuarial science, probability, and Bayes' theorem (all at least as relevant to data-driven decision-making as the randomized trial) were given so little attention in the book.
4. Finally, the great irony--and part of the "this book is so bad I have to finish it" quality--is that the author writes in an incredibly intuitive manner. The book is full of cognitively biased representation of the topic, owing mainly to "availability" heuristics, for example, the authors' excessive attention to the work of his friends, his roommates, his enemies, his daughter, or the companies he shops from. Better scholarship (or at least more rational) would have involved an initial sampling of all the relevant examples and final selection of the ones that would best illustrate the concepts (which I never really understood--see points 1 and 2). As other reviewers have pointed out, there is also "confirmatory bias" all over the place (presenting only the facts that fit with one's idea)--why aren't the counter arguments and counter-evidence better presented? The author must know that people buying a book on statistics will actually be smart enough to weigh the different sides of an issue. Like I said, I read to the end just to see if there was a "punch line" where the author confesses about his unapologetically intuitive approach to writing--that the book was a joke on the reader.
I would recommend this only to people who know very little about statistics and are unaware how companies like amazon.com use statistics to improve business. Such readers will be impressed. For everyone else...there are so many better books out there. Paul Meehl would be super-disappointed in this work.
83 of 104 people found the following review helpful
"Super Crunchers" provides a very readable summary of what can be done to improve performance using the incredible volumes of data accumulated in business, government, health care, and education. Why now? One reason is that the massive amounts of data now available make randomization (essential to valid conclusions) much more achievable than in the past; the other is the low and continually falling costs of computers and storage media.
The bulk of Ayres' work consists of examples (names both companies and the software involved) within each of the sectors previously mentioned. Capital One has been running randomized tests since at least 1995 - tests include page layout, and type and size of offers. Google uses data analysis to fuel its web accelerator (uses your past browsing history to predict pages to be called up next), Wal-Mart's analysis of responses to various employment questions is used to rank potential employees, and Continental Airlines follows up on its own data to design follow-up programs for complaining fliers. Capital One's approach has also been used to evaluate various charity donation-matching programs, and could also be used to evaluate potential billboard and magazine ads. (Similarly, TiVo is now being used to evaluate various TV ads, using the same approach and measuring the relative frequency with which various ads are fast-forwarded through.)
"Offermatica" software not only automates randomization (format, type of offer) for a number of firms, it also analyzes the responses in real time, dramatically cutting the cost of experiments. Thus, no more waiting for hyper-controlled experiments in universities and laboratories that conclude, ALL OTHER THINGS BEING EQUAL (that never happens), eg. red is preferred to blue.
Randomized tests are also increasingly being used to evaluate various government programs, finding eg. that additional job location assistance more than paid for itself for those receiving unemployment benefits, guiding HeadStart programs to target those most likely to benefit.
"Super Crunchers'" health care examples were the most impressive. Don Berwick's "100,000 lives" campaign saved 122,342 lives in an 18 month period through persuading about 3,000 hospitals representing 75% of all beds to focus on six areas of improvement identified through data analyses. These included antiseptic placement of central line catheters in ICUs, elevating heads and washing the mouths of those on respirators, adoption of the latest heart attack treatments, and rapid response teams to patent beds.
Bottom Line: "Super Crunchers" is an exciting vision of what is already possible!
44 of 54 people found the following review helpful
Ayres argues that decisions in business and government should be made through the creative utilization of data analysis rather than as the result of anecdotal observation. While this may seem to be almost a truism, Ayres begins by demonstrating how older enterprises like the wine industry and professional baseball both rely more on feeling and experience than on the quantitative method. Both also rejected initial efforts to move towards a more data centric model.
The difference in the two approaches is not just a matter of managerial preference according to the author: "We are in a historic moment of horse vs. locomotive competition where intuitive and experiential expertise is losing out time and time again to number crunching." Examples include hedge fund experts who create value by finding empirical correlations between unrelated factors and the consumer lending business where front line loan officer judgement has been replaced by more reliable centralized formulas.
I have long worked in the telecommunications business in which a surprising number of important decisions such as constructing channel line ups or marketing products is based to a large degree on experience or feeling. As we have moved to a more data-based model, we continue to struggle to achieve the balance Ayres describes as comfort with both numbers and ideas.
Ayres discusses some of the institutional and ideological barriers to such a transition. The shift to Direct Instruction in primary schools, for example, pits "the brute force of numbers" against the professional experience of teachers and the philosophical inclinations of education professionals. In the commercial lending business, super crunching (defined simply enough as "statistical analysis that impacts real-world decisions") has effectively shifted discretion from front line employees to centralized experts, has deflated salaries and has created the potential to export jobs overseas.
Overall, this is a useful discussion of the challenge of blending science and art in management. It brings a wide range of examples into play and achieves balance in its conclusions. Aside from the pure reading experience, I left with some definite plans to explore the use of randomized trials in my business in place of focus groups and simple historical analysis.
12 of 13 people found the following review helpful
Much of this book is unsurprising, describing the awesome amount of data and processing power available to us. I'm an academic; I'm familiar with statistical analysis. Yet there were anecdotes in here that floored me: Websites playing with different layouts to manipulate their users' behavior; companies using randomized trials to figure out how best to serve (and extract money from) their customers; even the book's title was tested by taking out ads on Google under several different names. "Super Crunchers" wasn't the author's top choice, but it was the apparent of the masses, so that was that.
And we're only at the beginning of this phenomenon. The stories told in Super Crunchers seem fresh now; in a few decades, these powerful methods will be ubiquitous. If you're in business, this book will put you ahead of the curve. And if you're just a customer, this book will help you to be prepared.
18 of 22 people found the following review helpful
on June 12, 2009
When a customer deleted the cookies on his computer which identified him as a regular Amazon customer, he discovered that Amazon's quoted price for DVDs fell significantly. This prompted Amazon CEO Jeff Bezos to declare: "we've never tested and we will never test prices based on customer demographics."
This excerpt from Super Crunchers introduces two techniques that form the focus of the book: regression analysis and randomized trials. Regressions are a widely used statistical technique that can be used for prediction, inference, hypothesis testing and modeling of causal relationships. The term "regression" was coined in 1877 by Francis Galton, a cousin of Charles Darwin, when he estimated a formula to predict the size of sweet pea seeds based on the size of their parent seeds (there was "regression toward the mean": peas didn't grow into balloons). Applied to DVDs and book sales, regression analysis helps predict a consumer's willingness to pay and leads to a pricing policy that maximize the value of sales based on a consumer's characteristics and buying pattern. It also allows websites like Amazon to make buying suggestions based on the observation that "consumers who bought this also bought that".
Randomized trials, used in the testing of pharmaceuticals, takes the analysis one step further. Instead of analyzing historical patterns, they produce their own data in an experimental setting that involves the random allocation of different treatments to subjects. The key word is random: if two groups of subjects with identical characteristics are exposed to a different intervention or condition, with all other things being held equal, then we can be confident that any change in the two group's outcome was caused by their different treatment. Randomized testing can also be used to test how much money can be extracted from a consumer by exposing similar buyers to different bundles of products and prices.
Ian Ayres, a law professor with a taste for numbers, has applied statistical testing to a variety of subjects: taxicab tipping, affirmative action programs, car theft, baseball card selling on eBay, weight reduction programs, etc. He doesn't shy away from sensitive issues. He was the first to expose the higher price markups that women and minorities had to pay at car dealerships. His research shows that the impact of concealed handgun laws on crime is inconclusive and doesn't validate the "More Guns, Less Crime" hypothesis.
But as he demonstrates in his book, applying statistical techniques to social issues is not the prerogative of academics. These techniques have now moved out of the ivory tower, as business and government professionals are relying more and more on databases to guide their decisions. Number crunchers "are not just invading and displacing traditional experts; they are changing our lives. They are not just changing the way that decisions are made; they are changing the decisions themselves." From teaching methods to health care, management is now backed by rigorous data analysis. Evidence-Based Medicine or Direct Instruction force doctors and teachers to follow a script, like a flight attendant reading FAA safety warning word for word at the beginning of each flight.
Data-driven decision making sometimes faces strong resistance. In the medical sciences for instance, the idea that doctors should give special emphasis to statistical techniques remains controversial to this day. The author notes that doctors are less likely than pilots to accept the drills of decision support software: "unlike pilots, doctors don't go down with their planes". In 1840, Ignaz Semmelweis was the Viennese physician who recommended doctors and nurses at clinics to wash their hands before surgery, after having observed that mortality rates dropped from 12 percent to 2 percent if they did. He was ridiculed by other physicians who considered hand-washing several times a day a waste of time, and after a nervous breakdown he ended up in a mental hospital, where he died at the age of forty-seven. How ironic: the reference for the diagnosis of mental disorders is now the Diagnostic and Statistical Manual of Mental Disorders or DSM, which evolved out of systems for collecting census psychiatric hospital statistics.
Super Crunchers can be read as a sequel to the hugely popular essay Freakonomics (the authors of the two books share the same blog on the New York Times website). Freakonomics focussed on how statistical analysis can reveal unexpected relations of causation, like the link between the abortion rate in 1970 and the crime rate in 1990. It also played the forensic economist trying to expose criminal frauds, like match rigging by Japanese sumo wrestlers or the rewriting of multiple choice questions by teachers in the Chicago school system. The book was "freakish" in a way as it presented unconventional academic results that sometimes had only a faint relation to economics as a science. By contrast, Super Crunchers focusses on real-world decisions and how they are being impacted by data-driven management. It is even further away from the academic discipline of economics. But the techniques and results that it covers are highly relevant for policy makers and business executives.
9 of 10 people found the following review helpful
on March 27, 2008
By Ayres' definition, I guess I am a professional Super Cruncher. I really didn't enjoy this book. While some of the anecdotes were interesting, he doesn't go into enough depth with each story and after a while the book just feels like a grab bag of anecdotes. Sure, I would have appreciated a more accurate and less sensational use of technical terms, but I would have walked away from this book with a much better feeling if he fleshed out more context with each anecdote. As a counterpoint, take "Money Ball" by Michael Lewis - this book is a single Super Crunching story and is a fantastic read.
It is a shame that such a recent book fails to discuss any of the serious debate of the issues that arise from data mining techniques in hypothesis formation and testing. For a fairly lay description of these issue in the medical field, track down the paper "Why Most Published Research Findings Are False" by John Ioannidis.
All in all, this text succeeds as a quick run through of various applications of data mining. But as a book, it isn't satisfying.
29 of 37 people found the following review helpful
on September 23, 2007
Format: HardcoverVerified Purchase
it's a shame that nowadays famous authors and professors dose out praises so easily that innocent readers like me were lured to purchase otherwise unimpressive work,
as the author himself admits, the theme of the book is based on a short 1954 article by psychologist paul meehl, which is much better explained and explored in the book "rational choices in an uncertain world", this book is at best just a loose collection of how number crunching wins over intuition, at times, the anecdotes cited are not relating to number crunching at all, e.g., his own story of finding his lost cell phone by making a number of calls ....
i really hope serious authors and professors could in future take their compliments are seriously
5 of 5 people found the following review helpful
on May 2, 2009
A book about statistics would not seem to interest most readers, but Ian Ayer's Super Crunchers, will interest many. From gambling habits to medical research, from job-seeking to titling a book, Ayers shows us the power, importance, and objectivity of statistical analysis. He also briefly touches upon the ethics associated with the use of statistical information.
In general, Ayers's uses plain, but complete, language to explain mathematical concepts such as random assignment. However, his use of breast cancer detection rates as a vehicle to explain the concept of "initial assumptions" was presented poorly. At the end of his explanation he reported that many doctors are confused by the concept of "initial assumptions" in regards to breast cancer detection. By his presentation, I can understand why doctors, so often misunderstand the concept of "initial assumptions."
However, in general, as Ayers so compelling shows us the power, importance, and objectivity of statistical analysis, college math professors should consider requiring it as a supplemental text for courses in statistics (particularly courses for non-math majors). For the non-mathematically inclined, Super Crunchers, would help take away some of the confusion and mystery of "number crunching."
However, Super Crunchers is no "textbook." It is an insightful and informative book for the consumer, the patient, the job-seeker, gambler etc. In short, it is a book almost every adult would find relevant in one way or another.