Book Review: Dr. Deming
This was the first non-database related book that I've read in a long time, and its full title is actually "Dr. Deming: The American Who Taught the Japanese About Quality". I was expecting to read about the life of W. Edwards Deming, but instead the book was more of a summary of his teachings and management methods. I was still glad I read it, as it takes an insightful look at the differences between the American and Japanese auto industries as well as management methods in general and how Deming's methods improved both in Japan.
While I can imagine his name might be covered in business school classes, techies like myself may be in need of an introduction. William Edwards Deming was a professor, statistician, and consultant with a focus on improving quality through both statistical and managerial methods. His earlier years were spent at Western Electric's Hawthorne Works in Cicero, Illinois, the remnants of which I pass by daily on the train ride to work. While at Western Electric, he worked with Walter Shewhart, known by many as the father of statistical quality control. Deming developed the statistical sampling methods used for the 1940 U.S. Census, and his experiences led to him being selected by the U.S. Army to help plan for the 1951 Japanese Census. In Japan, Deming gave lectures on process control through statistics and the overall idea of quality to hundreds of scientists and engineers. He also lectured top executives on how improving quality will reduce expenses while increasing both productivity and market share.
Deming later developed his practices into a series of lectures which he was hired to give to companies all over the World. The details of his lectures that were mentioned in the book are too numerous to cover here, but many of them centered around his list of 14 principles of management. Despite being geared towards industrial and manufacturing jobs, they can be applied to any position. I drew a lot of parallels to IT jobs, and a few of them piqued my interest:
You can't inspect quality into a product. Deming argued that no level of inspection or testing could ensure a quality product. Quality must instead be built into a product from the very beginning. Drawing parallels to software design, I'd argue he's spot on. Unit testing and other methods can't ensure software will be error-free, as there are too many possible situations to cover in a test environment. I've seen lots of buggy software that has passed unit testing because either the tests weren't sufficient or unforeseen changes happened elsewhere that caused the product to fail.
Eliminate management by numbers and numerical goals. Substitute leadership. I once worked in a shop where upper management was extremely numbers-oriented. If a decision could in any way be made based on values calculated in a spreadsheet, they'd find a way to do that. Deming argued that true leaders will know what needs to be done in order to further the business, and wouldn't have to depend on numeric values for justification.
Cooperation over competition. Cooperation is a fundamental ingredient that leads to improvement, even when occurring between rival companies. Cooperation between all players helps to ensure that the industry as a whole moves forward, and everyone will benefit as a result. A lot of standards (e.g. Compact Discs, the USB interface, network protocols) have come from cooperation between competitors. Similarly a lot of promising technologies have gotten off to a very rocky start when competitors didn't cooperate – one such example is the great VCR war between Betamax and VHS. In the beginning, consumers weren't sure which one to buy, movie studios had to produce two versions of each movie on tape, and in the end a lot of people were left with Betamax VCRs that nobody made tapes for. I think the SQL Server community is an excellent example of cooperation over competition. The community has so many professionals with a lot of great tricks up their sleeve, but rather than keep them a secret they share their knowledge freely with anyone who's interested. The more DBAs who can effectively solve problems as a result of this knowledge, the better off the SQL Server platform will be.
Dr. Deming was born on October 14, 1900; yesterday would have been his 110th birthday.
As this book was written in 1991, it does an excellent job of showing how the quality of Japanese products (particularly automobiles) is leading to Japan straining the US auto industry, and correctly predicts that Japan's auto sales will overtake the US in the future. The emergence of China wasn't really on the radar 19 years ago, and I'd be interested to see what the author's take on more current events is. I'd imagine one could successfully argue that China's manufacturing dominance is more linked to dirt-cheap prices than to quality, which pretty much throws the book's main argument out the window. That being said, I won't hold a book liable for not being able to predict all aspects of the future.