Datasets:
botchagalupe
/
ProfoundDeming

Dataset card Data Studio Files
xet
Community
2
text
stringlengths
0
1.44k
Shewhart classified defects as being caused by one of two things. The first was chance, what Deming would later call common cause. These were variations that could be predicted and should be planned for. The second was assignable, or what Deming would call special cause, causes that couldn’t be predicted and shouldn’t be planned for.
Variation always occurs in all processes. For Shewhart, as long as the variance fell within standard-deviation limits, the variance was inherent to the manufacturing process (i.e., assignable or common-cause variation). Sometimes the variation’s cause was an outlier, like an employee not being trained well or a machine that broke. That’s not something Hawthorne’s managers would have predicted. These special cases or anomalies are not part of normal or standard operations; hence, chance or special-cause variation.
The real value of Statistical Process Control is that it allows you to observe variation and look at random versus non-random patterns. A random pattern represents a stable process, a.k.a. a process “under control.” A non-random pattern is a useful predictor of potential defects, signaling an amount of uncertainty in the process. And here is the root of all evil: misidentifying variation.
Let me illustrate. Say I have an iPhone app for my thermostat. I like to keep my home at 70°F for the dogs when I’m away for the day. I expect the temperature to vary from 68°F to 75°F throughout the day. This is normal or common-cause variation; it’s to be expected. I don’t need to worry or intervene.
However, 80°F would be problematic for my pets. If I noticed the temperature on my iPhone app trending upward from 68°F to 72°F to 74°F—that is, a non-random pattern—I might suspect that something was off. Maybe my A/C is on the fritz. This might signal a problem leading to a special cause, meaning I might need to intervene before the temperature becomes harmful to my pets.
Shewhart understood that all processes have variation, but patterns of variation in a process can reveal insights into future defects. That’s why Statistical Process Control is so phenomenal: It allows you to statistically predict defects before they occur.
In The New Economics, Deming explained variation with the example of an insurance actuary who was constantly late. After reading his true-to-life example, I realized I’d had the exact same experience. At one of my startups, I had a software developer—let’s call him Bob—who was one of the best I’d ever worked with. Gifted, he took to code like a fish to water. He did have one annoying habit: he was chronically ten to fifteen minutes late every morning, getting to our thirty-minute morning team meeting halfway through. Then he’d always say, “You won’t believe what happened to me this morning!” and spend another five minutes regaling us with a story. One time he was late because a goth guy got his foot stuck in the subway door. Another time, there was a protest outside of a furrier. A neighborhood parade. A fight outside a bagel shop.
You see, Bob didn’t grow up in New York City. He didn’t know that these kinds of weird things happen all the time. While each occurrence was unique, overall, these were common occurrences. They always made him ten to fifteen minutes late. Even if he couldn’t predict what would happen, he could always count on something happening. As such, he should have planned the process of his morning commute to take fifteen minutes longer than expected. Every. Single. Day.
And every single day, the team would lose fifteen to twenty minutes of our thirty-minute meeting between waiting on Bob to show up and hearing the latest edition of “You won’t believe this . . . .” Firing Bob was out of the question; he was simply too valuable. And he was a hard worker; he spent hours of his own time at home working on problems and coding issues. There was no way I was going to chew him out for being a few minutes late every morning.
If I’d been smarter, I would have realized I was dealing with a common-cause problem: I could always count on Bob being ten to fifteen minutes late . . . so, why didn’t I just start our morning meeting at 8:15? Neither Bob nor I understood that his being late was common-cause variation—something normal in day-to-day commuting in the Big Apple, regardless of Bob’s exotic story du jour.
Here’s another example of misidentifying variation. The manager of a datacenter misidentified a special case as something common to the system. The datacenter was built out in the backwoods of an area that rarely experienced snow. Well, one day, there was a freak snowstorm. It was so bad that no cars could get on the road. None of the center’s call and operations staff could get to work that day. The datacenter’s functions were limited, and the company lost a lot of money.
The odds of this happening again were quite slim. However, the datacenter manager decided that he would never let this happen again and mandated that all new hires who were part of the call and operations staff had to live within a mile of the datacenter. He misidentified special-cause variation as something that should be predictable (i.e., common-cause variation).
Firing Bob would have been applying special-cause logic to a common-cause situation. However, the datacenter manager requiring all new hires to live within a mile of the place was applying common-cause logic to a special-cause situation.
According to Shewhart’s Statistical Process Control, managers shouldn’t waste their time trying to fix every single problem. Instead, they should identify which ones can be predicted and fix them. Identify the ones that will likely never happen again and don’t make knee-jerk decisions. As a result, managers can spend their time on things they can control and waste very little of their time on things they can’t.
When Ed was introduced to Shewhart years later, after he had left Hawthorne, the lightbulb went off. Ed already knew variation was a fact of life; as a mathematical physicist, he was intimately familiar with thinking in statistics and probabilities. Shewhart’s insights quantified this variability in manufacturing.
Since the randomness could be predicted, Ed understood that defects weren’t due to the workers but to how the manufacturing process was designed and operated. As we’ll see, this understanding—that workers weren’t the problem in any given system—would go on to become one of Ed’s most important ideas (the Theory of Variation) in Profound Knowledge. But that wouldn’t be for many years to come.
Deming’s Journey to Profound Knowledge - How Deming Helped Win a War, Altered the Face of Industry, and Holds the Key to Our Future - Part 1 - Chapter 5: Pragmatist
Shewhart imparted one more foundational concept to Ed: the philosophy of pragmatism, what Deming would later call the Theory of Knowledge, the first element in the System of Profound Knowledge. And this uniquely American philosophy began with an unliked and unlikely character who just wanted to measure things. This character was fascinated with a problem that had plagued humanity almost ever since Lucy began making tools. That is, the fundamental challenge of accuracy and standardization in measurement as well as the dichotomy of the pursuit of perfection over the practical.
How Do You Know What You Know?
Le Grand K. Sounds like the stage name of a French rap artist. It is a stage name of sorts, not for a person but a piece of platinum-iridium sitting in three vacuum-sealed bell jars in an environmentally controlled, underground, triple-locked vault outside of Paris since 1879. France’s version of Fort Knox, if you will. It’s so precious that it’s been “out” only three times: 1899, 1939, and 1988.
This antique oddity is dense, twice as dense as lead. The golf ball–sized cylinder is incredibly strong and will never rust. Despite its hardiness, it’s handled with kid gloves. Well, better, actually: its guardians are too afraid to touch it even using gloves. They use a special tool wrapped in filter paper to avoid even the most minute of scratches. Perish the thought of a fingerprint!
Such care is usually only given to highly valuable works of art, like a Fabergé egg. Not even the British crown jewels get this kind of treatment. Curiously, this unremarkable piece of metal is worth only about $43,000. What gives? Why does this smooth little alloy cylinder warrant so much security and care? What’s its significance? Perhaps at risk of sounding melodramatic, this piece of metal is how we measure the world. Or most of it, anyway.
Starting in 1879, this small metallic chunk has served as the international prototype of the kilogram. It doesn’t weigh a kilogram; rather, a kilogram is whatever it weighs. There are six official copies; those, plus the original, are stored at the headquarters of the International Bureau of Weights and Measures in Saint-Cloud, France. Forty replicas were created in 1884 and distributed to a handful of nations. The US, for example, received two (named K4 and K20). All the official measurements for anything measured in kilograms in each country are calibrated to these copies of Le Grand K. Pharmaceutical scales, aerospace calibrators, surgical equipment, you name it: they are all derived from the same international standard. Every forty or so years, the copies are flown to Saint-Cloud to be compared to Le Grand K and recertified as being exactly one kilogram.
The problem is that a kilogram isn’t always a kilogram. Like my own diets, K20 initially “lost” weight in the recertification of 1948 but by its next weigh-in had gained it all back. (Unlike my diets, the differences were measured in micrograms.) K4, like those I envy, consistently lost weight with every check-in. Some of this ebb and flow had to do with differences in how the weights were cleaned and stored, and even how they absorbed atoms floating in the air.
All jokes aside, this was a serious problem. If I’m flying in a thirty-year-old airplane, I want to know that the calipers the engineers used to design the plane are the exact same size as the ones the maintenance people used to service the plane this morning. We can’t have wandering measurements, and many of our global measurements are derived from the kilogram (such as the newton and ampere) as well as derivatives of those measures (such as the pascal and joule) and those measures’ measures (such as the watt, volt, tesla, and lumen).
Imagine how much worse it was before scientists began to standardize measurements. A “foot” was the length of a human foot. Can you imagine King James—LeBron, that is—asking a cobbler for a pair of shoes one foot long?
Everyone used different measures. In the ninth century, Charlemagne decided his foot would be the standard foot everyone in his empire should use. In the twelfth century, King Henry I declared a foot to be one-third the length of his own arm. His arm was thirty-six inches; thus, the twelve-inch foot used only by the US, Myanmar, and Liberia still adhering to the imperial system. In the thirteenth century, King Edward II decided that three grains of barley, or a “barleycorn,” would be used for shoe size measurement, with thirty-six barley-corns equaling one foot.
So many local and regional measures were used that at one point there were something like a quarter-million different measures of length, weight, etc. A pound of lead in one part of the world could be lighter than a pound of lead in another.
The Metre Convention on May 20, 1875, established the metric system, including the kilogram (cue the first Le Grand K) and the meter. The thinkers behind it wanted a system of measurements “for all times; for all people.” The Treaty of the Meter defined a meter as one-half of one-ten millionth the distance between the North Pole and the equator. Seeing as how that was a somewhat inconvenient thing to measure, the more practically minded thinkers agreed on a handy substitute: a meter would be defined as the cord length needed for a clock’s pendulum to travel one swing per second. This was about as close to a universal constant as they could agree on.
But no matter how hard you try to get the perfect measurement, there’s always going to be some troublemaker who comes along and picks holes in it.
In this case, Charles Sanders Pierce. C. S. Pierce was a bona fide member of the Boston Brahmins, the elite of society. It was said that before immigrating to America, the Brahmins sent their servants ahead on the Mayflower to prepare the summer cottage—a blue blood in every sense of the term. He was, however, brilliant.
He was particularly fascinated by weather (he would be employed off and on with the US Weather Service) and precision measurements—pendulums, in particular. He recognized early on that pendulums of different lengths swung a two-second cycle depending on local variations in the Earth’s gravity.
He became obsessed with improving pendulums’ precision. But if pendulums everywhere in the world needed to be of slightly different lengths for a uniform cycle, then a meter in Britain wouldn’t equal a meter in Boston. Not precisely.
In 1872, Pierce, along with the famous Supreme Court Justice Oliver Wendell Holmes and others, founded the Metaphysical Club to discuss philosophy. These budding philosophers rejected the deterministic worldview of their European counterparts who espoused Enlightenment. Pierce called their new idea pragmatism.
A proponent of Pierce’s, a man by the name of C. I. Lewis (not to be confused with C. S. Lewis of Narnian fame), authored a book called Mind and the World-Order.
The ideas it contained were fundamental to Shewhart’s creation was rooted in the thinking of Aristotle. In the Enlightenment’s way of thinking, you could know something without needing evidence to prove it. Descartes wrote, “I think; therefore, I am.” He knew something without needing any kind of outside validation or evidence—a priori knowledge. Another example would be knowing that one plus one equals two. Philosophers don’t need to run an experiment to know the answer is two; they just do.
A posteriori knowledge is when you know something because, and only because, you have the evidence to prove it. A priori thinking would say, “Add a gallon of milk to a gallon of milk and you’ll have two gallons of milk.” Pierce’s new branch of philosophy would say, “I know I have two gallons of milk because I added one gallon of milk to a gallon I already had.” One theorizes; the other experiments and observes.
A priori thinkers believed they could reason their way out of anything inside their own heads; a posteriori thinkers, or “pragmatists,” believed they could reason their way through something only by doing it. Put another way: They believed experience was the best teacher.
With his pendulums, for example, Pierce realized it was possible to create a perfectly precise pendulum. Theoretically. Practically speaking, there was a point where creating an ever-better pendulum simply wasn’t worth it.
Mathematically, he had reached a point of diminishing returns: investing more time and money simply didn’t make any sense. Philosophically, he reasoned that using a pendulum’s swing as the basic unit of measurement simply wasn’t practical.
That’s why in 1877, just two years after the world’s top minds agreed on using grandfather clocks as the basis for all physical measurements of distance in the universe, Pierce wrote that the meter should be tied to an unalterable, absolute unit of measurement: a certain number of wavelengths of light at a certain frequency.
His ideas were taken up and improved upon by Albert Michelson, who would win the 1907 Nobel Prize in Physics for measuring the prototype of a meter to within one-tenth of a wavelength. In 2019, one hundred and forty-four years to the day, the successors of the Treaty of the Meter redefined the meter in terms of the speed of light. They also redefined the kilogram from being measured by an expensive paperweight to being a derivative of Planck’s universal constant.
The story of accuracy and standardization is, in reality, a story of pragmatism. At any given time, a standard is a measurement that suffices and that everybody agrees upon. Even Planck’s universal constant is a pragmatic approximation of the true value. The idea of a pragmatic approach to standards coupled with the accuracy of measurement (described in Chapters 3 and 4) is a great example of why Profound Knowledge is so revelatory. If we focused only on the Theory of Variation without applying the Theory of Knowledge, we’d miss the opportunity to improve upon standards (Pierce, for example, finding a better way to measure the meter).
America’s Jazz Philosophy
Jazz may be America’s one truly original musical art form. Pragmatism is America’s one truly original contribution to philosophy.
Consider the number pi. We learn about the basic concept in high-school algebra. It’s simply a circle’s circumference divided by its radius. When we do the math, however, we quickly discover that the division problem just keeps going and going. Our teachers tell us to round it to two decimal places and simply use 3.14.
A pragmatist says, “Okay, 3.14 is good enough for what we’re doing here—it doesn’t make sense to find its absolute value. We’ve got too many other things to do.” While the determinist would spend his life calculating all the digits in pi, the pragmatist would say, “I can’t spend my life doing a math problem—I’ve got real problems to solve. Two decimal places is close enough.”
While this simple example sounds like common sense, most people are still stuck in the absolute deterministic approach. They believe that perfection can—and should!—be reached.
Pragmatists begin with observations and empirical data—in other words, hard evidence—then work their way backward. This might sound simple, but it’s actually quite rare for people to think like this. Most rely on “common knowledge,” knee-jerk reactions, and going along with “the way we’ve always done it around here.”
In Taylor’s and Ford’s time, quality inspection was essentially a question of whether a product was close enough to perfect to pass. But Shewhart, bringing a pragmatic outlook, asked, why wasn’t it perfect? In so many words, he said, “Look, perfection is an illusion. We’ll never reach it. But if we use a posteriori thinking, we can systematically improve how we manage what we manufacture.”
Let me provide a modern example of the perfection mentality. There was a time when the owners and managers of banks, insurance companies, and major retailers wanted their websites to have 100% uptime; they didn’t want their websites to be down. Ever. That may have been the goal, but the reality was too chaotic to adhere to corporate policy. Well, if the suits couldn’t have 100%, how about 99%? Or 99.9%? Could the IT guys get it up to 99.99%? Of course, five nines would be even better! Just how many extra nines could the IT department get to? This went on for years (and still goes on today).