Washington — During his 19th trip to Japan, W. Edwards Deming -- the only American besides Douglas MacArthur to receive the Medal of the Sacred Treasure from the Japanese Emperor -- had to make a train connection in Hakata. The train he was riding on arrived at the station at 7:23; his connecting train left at 7:24. He wasn't worried about making the connection. "I knew it wouldn't take me a full minute to cross the platform," he explains.
Comparing that train ride to delays he has experienced in the United States "shows how much we accept defects as a way of life," the statistician says.
His mission over the last 45 years has been to demolish the acceptance of mediocrity, mistakes, and delays -- a mission that has taken hold in Japan with religious fervor, and is starting to have "a leavening effect on American management," he says.
"No other statistician in history," he says warmly, "has seen his work bear such fruit."
A tall, crewcut-topped, crusty old gentleman, Dr. Deming planted his first seeds in Japan at the end of World War II when, as a statistician working briefly for the US War Department, he met with his Japanese peers and "told them they could lead their country toward quality. They didn't understand, of course , but it was a start."
Meanwhile, a group of engineers and statisticians there gathered under the title of the Japanese Union of Scientists and Engineers toward the end of the war help the war effort. They held themselves together after the war with the aim of helping the reconstruction of Japan.
One member of the union, Koji Kobayashi of the Nippon Electric Company, had read a paper on the statistical control of quality by Dr. Walter Shewhart, a friend of Dr. Deming's. "Mr. Kobayashi had tried to make it go -- learning by doing," Dr. Deming says, "but he did not meet with success."
The method described by Dr. Shewhart intrigued Mr. Kobayashi, though, and was spread byhim to other members of the union, who decided to invite Dr. Deming to return to Japan and instruct them.
"I told them that Japanese industry could develop in a short time. I told them that they could invade the markets of the world and have manufaturers screaming for protection in five years," he says with a twinkling grin. "I was, in 1950, the only man in Japan who believed that."
Statistical methods, filtered through the 500-plus managers who attended those first seminars, caught fire in Japan, and within two years started to show significant results. Now, the Japanese reputation for quality is worldwide.
Former Ambassador Walter Annenberg, in a recent Wall Street Journal editorial , spelled out the results: "Japan, a country that is 1/25th our size in land area and that must import 90 percent of its raw materials and energy, now produces more television sets, as much steel, and as many automobiles as we do," he says. What is more, "the little country has productivity growth rate of 11.4 percent compared with about zero growth in the US."
This difference has been attributed to everything from government regulations in America to the energy levels of the Japanese workers, but Dr. Deming sees quality behind it all. The Japanese do not waste time creating -- and then correcting -- defective products, he appoints out, and their reputation for quality guarantees productivity, competitive position, and sales.
Others are not so sure. "The statistical control of quality can't hurtm ," Gregory Gruska, engaged in quality control in the US automotive industry, says. "But just hiring a statistician is not going to turn your business around overnight."
Dr. Deming responds brusquely: "You must engage a statistician of competence and maturity, not just a statistician."
Quality control is a religion to Japanese industry. Through the Japanese Union of Scientists and Engineersd, nearly 8 million workers have undergone quality training, and one longtime observer of Japanese industry believes that nearly every worker in Japan has had some company indoctrination in quality control.
The Japanese certainly see it as the source for their success. In fact, the most coveted industrial award in Japan is given to the company and individual who have done the most in quality control for the past year. Named after their prophet, it is called the "Deming Award."
In his modest basement office in Washington which is filled with stacks of paper, dusty books, and shelves strewn with awards and Japanese knickknacks, Dr. Deming seems less the successful leader and more the voice crying in the American wilderness.
The voice -- c combination of barked declerations, well- modulated musings, and explosions of belly-deep laughter -- has been crying since 1942, when at Dr. Deming's suggestion a series of 10-day crash courses on statistical methods were initiated by Stanford University and continued by the War Department. "Brilliant applications [of the statistical control of quality in the US] burned , sputtered, fizzled, and died out," he laments.
There are those who think that the idea lived on. "Most large businesses have some sort of quality control," says Robert Abbott, technical director of the American Society for Quality Control in Milwaukee, Wis.
"Absolutely not!" Dr. Deming storms."They might have inspectors,m sorting out the good products from the bad, but that's just burning toast and scraping it, burning toast and scraping it. They planm for defects, when they could be preventingm defects!"
Quality control, as Dr. Deming envisions it, consists of methods for improving the performance of worker and machine. These methods, he explains, must become a way of life. Covering everything in a plant from the quality of the materials used, through the porduction, to the customer's use, it makes sure that every part is correct and that the process improves day by day.
At its simplest base, statistical control is a series of definitions which set out the specifics on what each part and combination of parts should be.
If materials is to meet a certain standard, Dr. Deming reasons, the inspector should know what that standard is and be able to measure it, and the people making the product should also know exactly what it is they are expected to make.
Companies that provide the parts are also included in this process: They should know what they are expected to supply. And those who design and redesign the product should know both how the customer is using it (so they can see what needs improving) and what the system is capable of producing (so they don't design something that cannot be produced economically).
Once these standards are set, they can be used to measure the parts and process at various points along the line to find what is causing the faults and eliminate the causes. Statistical methods pinpoint problems on "control charts" -- a simple means of graphically plotting the measurements to reveal pitfalls, invented by Dr. Shewhart.
By use of controls charts, a manager is able to see specific points in production where defects occur and determine whether the cause is a fault that a specific worker or group can correct and avoid hereafter, or a common cause, a fault of the system, correctible only by the management -- such as using the wrong thread in a battery of machines, or inadequate lighting, or even poor food in the cafeteria.
"A schoolboy can understand these principles," Dr. Deming growls, "but American management doesn't understand them."
He blames American managements for "80 percent of the production problems," claiming that "a good 50 percent happen because the worker does not know what he is supposed to do, so he does it wrong. And who gets blamed -- the manager, who didn't train him correctly? No, it's the poor worker, every time."
To be successful, the method appears to rely on open and extensive communication between management and workers to unravel the snags of production. "We made a mistake when we started quality control in this country of not involving top management," Dr. Deming says, "so I made sure that we did not make the same mistake in Japan."
In the US, he refuses to work with any company that does not include all of top management in his four-day seminars, saying, "If they don't know that they're in trouble, I can't tell them."
Although his most fruitful work was done in Japan, Dr. Deming's consulting schedule has always been crammed with engagements throughtout the US and Europe ("I'm 140 percent busy -- always have been").
And although the longtime Washington, D.C., resident speaks of Japan with affection and a touch of pride, it is clear that Dr. Deming is all-American from his three-piece, blue serge suit to his 12-year-old Lincoln Continental.
Perhaps that is why he stolidly refuses to believe those who credit JApanese culture and government as the source of their dizzying surge in productivity.
"Excuses!" he shouts, citing the 200 or so Japanese companies that run US branches with American workers -- and excellent results.
But he concedes that the key to involving top management fits neatly into Japanese culture, where emphasis is placed on the ability to work as a group. There, individual workers on a particular production line often meet in "quality circles" (QC) to discuss ways to improve that line's quality.
The QC is seen by some as one of the most fundamental reasons for increased productivity, and it is rapidly being immitated here in the US. But Dr. Deming is not convinced that such imitations will work with Americans "because the group spirit does not exist here."
Lacking the group spirit is one drawback to American productivity. But Dr. Deming cities another, more basic reason: fear. "If an Americans worker sees a problem on the line, he takes it to his foreman, and nothing is done about it," Dr. Deming explains. "He talks to the foreman again, and still nothing is done.
"He talks to him a third time, and the worker is labeled as a troublemaker," the doctor scowls. "That's a hard reputation to shake in this country -- once you're known as a troublemaker, you can't say or do anything right."
So the American worker hesitates to "rock the boat," he says, and remains frustrated when solutions that seem obvious to him are not applied.
"A man came to me for help with one of his production lines, which had a terrific turnover," he illustrates. "I found that the production on that line was 30 percent defective -- they were trying to make two materials stick together with the wrong glue. No wonder the workers quit!"
It is this difference in attitude between American and Japanese management, Dr. Deming claims, that is responsible for the difference in productivity. But the Japanese attitude, he says, has been repeated with success in this country. "There are some very fine firms; I like to think I've worked with the best of them."
Aside from the "few fine firms," Dr. Deming says it will take "some high-class funerals" to solve the problems of American management. "Even then you won't solve it, though," he muses, "because there's no one coming up who knows any better than the lot they have now."
Softening a little, he admits that "respect for the statistician has increased in the US" over the last 45 years, though he is not convinced that American management has any real idea of what a statistician is supposed to do in a company.
"A man called me the other day and said, 'Please come down here for a day, and do for my company what you did for Japan. We, too, wish to be saved.'" He reflects on this with a twinkling eye, and explodes into laughter. "It is not so simple," he says.
In the book he is writing on quality control, however, he gives a more thorough description of the uses to which a statistician can be put, including:
* Teaching statistics to the managers, engineers, foreman, workers, purchasers, and consumer researchers.
* Writing specifications for the incoming material, taking into account how that material will be used on the production line.
* Setting up cooperatives tests with each supplier to make sure that his tests agree with yours, and working with the supplier so that you do not end up accepting defective parts or materials.
* Making sure that workers know what their jobs are by defining satisfactory performance for assemblies, prototypes, and final products.
* Carrying out inspections on incoming materials or data, process, and final products, and making sure inspectors agree with one another.
* Using data from the inspections to (a) detect by control charts a special cause of variation, and (b) learn what proportion of the problems of rejection, waste, and productivity belong to the system (management), and what proportion can be attributed to the individuals workers.
* Measuring the loss from handling damage, and designing a system to prevent such damage.
* Coordinating the redesign of the product with (a) information about the problems of competitive products in service, (b) information about the problems of the customer, and (c) information about the costs of the incoming materials and the capabilities of the process.
Using a statistician in this way, Dr. Deming believes, will lead American management back into the faith of statistical quality control and result in significant productivity increases during the 1980s.
Will the US productivity rate then surpass the Japanese?
"I expect the gap will actually get wider," he muses. "The Japanese aren't standing still, you know, waiting for us to play tag."