World Class Manufacturing

Chapter 1

Faster, Higher, Stronger

In the 1950s through the 1970s, running manufacturing companies became gentlemen’s work. Decisions and policies were made by people twice and thrice removed from the manufacturing arena. Authority was in the hands of staff people who sifted data from other staff people. Venturing out into the plant was, well, venturing. It was prudent to stick around offices and conference rooms and make sure your backside was covered. Excitement in industry was confined to high-tech R&D. Manufacturing was stagnant.

How quickly things change. While the changes have scarcely touched small companies, the well-known manufacturers are caught up in revival, renewal, recovery, and renaissance. A popular term among those caught up is world-class manufacturing or a term like it. World-class manufacturing may sound like Madison Avenue hyperbole, but it is not. The term nicely captures the breadth and the essence of fundamental changes taking place in larger industrial enterprises. A full range of elements of production are affected: management of quality, job classifications, labor relations, training, staff support, sourcing, supplier and customer relations, product design, plant organization, scheduling, inventory management, transport, handling, equipment selection, equipment maintenance, the product line, the accounting system, the role of the computer, automation, and others.

The Goal and the Path

World-class manufacturing (WCM) has an overriding goal and an underlying mind set for achieving it. The overriding goal may be summarized by a slogan suggested to me by a manager at the Steelcraft division of American Standard, where I was presenting a seminar. During the afternoon break, the fellow told me that he had digested all that had been said, and he concluded that the whole thing was like the motto of the Olympic Games: citius, altius, fortius. From the Latin the English translation is faster, higher, stronger. The WCM equivalent is continual and rapid improvement.

A few years ago we didn’t even know the factors of manufacturing that ought to improve. There was little agreement on what excellence in manufacturing is, because we thought in terms of tradeoffs. Plant managers or their corporate overseers picked one set of high-priority targets one year (for example, defect rates and warranty costs) and another, seemingly conflicting, set to work on the next (perhaps overhead costs and customer service rates). The high priorities were where problems seemed most severe. Lacking manufacturing principles, we tackled the problem with trade-off analysis.

Today there is wide agreement among the WCM revisionists that continual improvement in quality, cost, lead time, and customer service is possible, realistic, and necessary. There is now good reason to believe that those goals may be pursued in concert, that they are not in opposition. One more primary goal, improved flexibility, is also a part of the package. While some of our leading manufacturers have trouble avoiding pitfalls that lead to inflexibility (pitfalls that are avoidable), the goal itself is not an issue. With agreement on the goals, the management challenge is reduced to speeding up the pace of improvement.

The improvement journey follows a surprisingly well-defined path. The journey requires clearing away obstacles so that production can be simplified. A fast-growing body of writings (including my own 1982 book, Japanese Manufacturing Techniques: Nine Hidden Lessons in Simplicity)¹ offers lists of obstacles to remove and ways to simplify: fewer suppliers, reduced part counts, focused factories (focused on a narrow line of products or technologies), scheduling to a rate instead of scheduling by lots, fewer racks, more frequent deliveries, smaller plants, shorter distances, less reporting, fewer inspectors, less buffer stock, fewer job classifications.

Beyond the Basics

In the pre-WCM era we thought that production could be managed by the numbers. The numbers would show what to make, what to buy, whom to blame. If, for example, the latest cost report shows a negative cost variance in welding, the onus is on the welding supervisor to cut costs. But how? There are no data on the causes of the cost overage. The supervisor may crack the whip to get more output for the same labor cost. Alternatively, ask industrial engineering or quality engineering to do a study.

The numbers failed to show causes. Mostly they did not even show symptoms of real problems.

Numbers do serve the world-class manufacturer—when they show how good the product and service are, how much improvement is occurring, what problems to attack next, and what the likely causes are. WCM mandates simplification and direct action: Do it, judge it, measure it, diagnose it, fix it, manage it on the factory floor. Don’t wait to find out about it by reading a report later.

Some of that may sound like back to basics. Basics they are, but going back we are not. It is true that some of the emergent WCM techniques were in use in an earlier era—and then forgotten. In the main, however, the good old days in manufacturing never were all that good. Quality concepts were primitive by today’s standards. While some plants had an ethic of continual improvement (applied very selectively), the norm was to transform simplicity into complexity, which sowed the seeds of decline.

Turning Point

There is reseeding going on, and there seems to be a single year that could be called the turning point: the year 1980. In that year a few North American companies (and perhaps some in Europe) began overhauling their manufacturing apparatus. Those first WCM thrusts followed two parallel paths. One was the quality path, and the other was the just-in-time (JIT) production path.

One of the first to try just-in-time in North America was General Electric, which started up two JIT projects in 1980. Kawasaki in Nebraska and Toyota truck in Long Beach, California, began shifting from standard to JIT production in the same year.

The first North American companies to take the quality path, also in 1980 (give or take a few months) were Nashua Corp. in New Hampshire, Tennant Co. in Minneapolis, and IBM. (A bit earlier Matsushita in Franklin Park, Illinois; Sanyo in Forrest City, Arkansas; and Sony in San Diego began their U. S. operations with a quality focus. These may be thought of as imports from Japan rather than as turning points in existing North American companies.) Nashua got its start by bringing in W. Edwards Deming, the American who, along with Joseph Juran, was instrumental in getting Japan’s quality movement going in the 1950s. Tennant and IBM hired Philip Crosby, who was known to a few people as the author of a fine little 1979 book, Quality Is Free.² Tennant provided early support for Crosby to form a quality college in Florida.

Those stirrings in a few companies in 1980 may someday be chronicled as the third major event in the history of manufacturing management. The first two: (1) coordinating the factory through use of standard methods and times, Frederick W. Taylor, Frank Gilbreth, et al., circa 1900; and (2) showing that motivation comes in no small measure from recognition, the Hawthorne Studies at Western Electric circa 1930.

The 5-10-20s

World-class manufacturing could not become the third major event if it were to peter out. The signs that it will not, that WCM is much more than a fad, are persuasive. The list of companies that have already made order-of-magnitude improvements in quality and manufacturing lead time is getting long. For example, I have compiled (and continue to update) a list of the 5-10-20s, which refers to companies, factories, or parts of factories where fivefold, tenfold, or twentyfold reductions in manufacturing lead time have been achieved. The list, with explanatory comments in some cases, is provided as an appendix at the end of this book. Stories about some of the 5-10-20 plants will be told in later chapters.

My 5-10-20 list does not do justice to WCM developments outside of North America, nor is it at all complete for North America. I have conducted seminars and provided consultancy at manufacturing plants in a number of European and Pacific Basin (besides Japan) countries and have found the WCM fever to be globe-spanning.

With so short a history, WCM has not had a chance to mature in all of its natural habitats. What surprises many is the progressive unearthing of more and more natural habitats. I refer not to different continents and countries but to different industries and types of production. That is, what makes a world-class manufacturer in one industry also seems to work in many other industries. Let us see why that should not be surprising.

Streamlined Flow

Consider how a restaurant fills a customer’s order: The cook puts meat from the grill onto the platter, goes to the range to scoop some vegetables, opens the oven to get a baked potato, heads for the salad bar to extract a salad, and so forth. It goes fast, because a kitchen is small and the cook puts only one serving of each food item on the platter.

Stop-and-Go

A machine shop, a sheet metal shop, a printed-circuit-board shop—any shop or factory that makes to order—is just the same. As long as the shop or factory is small, production is usually quite fast. But who wants to stay small? We have plants—for final goods and component parts alike—with thousands of employees and hundreds of thousands of square feet of space. Now the work goes through the plant at a snail’s pace. Plant management has its hands full trying to prevent gridlock.

If a restaurant kitchen grew the way our factories do, the platter would go to the grill area for a piece of meat and then move by slow conveyor to the vegetable area. The meat would get cold—and might even fall to the floor once or twice on the way. At the vegetable area, the massive cookers might be tied up making vegetables other than the kind ordered for the platter, which means waiting until the next batch is cooked.

Growth is not the problem. The problem is the more-of-the-same approach to growth. A restaurant is a little job shop, to use the manufacturing term. It will not work if it becomes a big job shop—where a job (platter) has to traverse vast distances from one shop to another, waiting for one thing or another at most of the shops. Growth must be accompanied by a transformation to preserve speed, to avoid stop-and-go production.

Over the years we came to believe that stop-and-go production was the fate of the job shop. We also believed that job shops were the fate of industry, because customers are fickle; they want the variety that job shops can provide. Job shop people looked enviously at the flow shops, where work just flows down a production line or through pipes continuously (as patrons flow down a cafeteria line).

That view is out of style, because we have learned how to streamline our job shops, to make them behave more like flow shops. Some go so far as to simplify products and regularize schedules, and thereby transform themselves into flow shops. Many others—those that stick with customers who demand variety—will not become flow shops, but they can come close. The chameleon cannot ever be a leaf, but it can look like one. So it is in manufacturing.

Imperfect Flows

What tools and techniques make job shop transformations possible? At the top of the list are the set known as just-in-time production techniques. They were perfected by Toyota in Japan in the 1960s and 1970s. Toyota’s techniques caused work to move through parts fabrication processes fast and get to final assembly just in time for use.

JIT was shaped in the flow shop mold. Continuous-flow industries—the pure flow shops—have been around for a hundred or two hundred years. Examples are bottling, tableting, and canning; extruding and weaving; milling and refining. Some of the processes are tightly coupled. The work leaves one process and flows, perhaps through a pipe, to arrive just in time for the next. In that sense, JIT was around long before the people at Toyota thought of it.

In reality the flows are usually not all that continuous. The grain mills, the food processors, the medicine makers, the cloth producers, and the rest are stop-and-go producers, too. They go for a time on one size, style, model, or chemical formulation, then shut down for a complete changeover in order to run another. Shutdowns for changeover are one concern. The massive quantities that build between changes—the raw and semiprocessed material, and especially the finished goods pushed out well in advance of customer needs—are a greater concern. All are forms of costly waste.

There are dominant WCM precepts for treating the ailment. One is a JIT principle: The smaller the lot size, the better. World-class manufacturers of cars, tractors, and motorcycles have some lot sizes down to one unit by becoming adept at changeovers between models. This permits making some of every model every day, almost like continuous-flow processing. With that capability, they outdo the flow processors they started out trying to copy.

A second precept is the total quality control (TQC) principle: Do it right the first time. In the flow industries this means setting up for a new run so that the first yard of cloth, linear foot of sheet steel, length of hose, can, bottle, or tablet is good.

A third set of precepts is called total preventive maintenance (TPM). Maintain the equipment so often and so thoroughly that it hardly ever breaks down, jams, or misperforms during a production run. There is nothing like an equipment failure to turn a continuous processor into its opposite number.

Mass Production—Just in Time

While the JIT concept (if not the application) is natural in the flow industries, it took Henry Ford and his lieutenants to get JIT worked out in discrete goods manufacturing. Ford has been called the father of mass production. His Highland Park and, later, River Rouge plants mass-produced the parts just in time for assembly, and his assembly lines pulled work forward to next assembly stations just in time, too.

By 1914 the Highland Park facility was unloading a hundred freight cars of materials each day, and the materials flowed through fabrication, subassembly, and final assembly back onto freight cars. The product was the Model T, and the production cycle was twenty-one days. At River Rouge, about 1921, the cycle was only four days, and that included processing ore into steel in the steel mill that Ford built at River Rouge.³

That roughly equals the best Japanese JIT auto manufacturing plants today. But it was much easier for Henry Ford, because his plants followed his now famous dictum, They can have it any color they want, so long as it’s black.

Isn’t it easy to look like a continuous-flow producer, with very short manufacturing lead times, when every unit is the same as every other? Ford’s Tin Lizzies almost could have flowed through a huge pipeline with intersecting pipes bringing in the components at just the right locations and times.

The Model-T factories were what is known as dedicated plants and production lines. Where capacity is cheap (cheap equipment or labor) or volume is high, dedicated JIT lines make sense. Most producers of television sets, radios, videotape recorders, and personal computers today have enough volume to follow the easy dedicated-line path to JIT. Most automobile manufacturing is of lower volume and cannot achieve JIT so easily. Nissan in Oppama, Japan, sets up a dedicated line only if sales volume is 10,000 cars a month or more. Since most models they make fall below that number, other approaches are necessary. Some of the other approaches are examined next.

Making Just What Is Sold—Every Day

Whether making things that pour (the flow industries) or things that are counted in whole units (discrete goods), a WCM precept is to produce some of every type every day and in the quantities sold that day. Making more than can be sold is costly and wasteful, and the cost and waste are magnified manyfold as the resulting lumpiness in the demand pattern ripples back through all prior stages of manufacture, including outside suppliers.

Makers of highly seasonal goods sometimes have sound reason for building at least some stock days or weeks before use or sale. Most of industry’s chronic mismatches between demand rate and production rate are not caused by seasonality, however. Those mismatches are fixable. Companies in the flow industries need to figure out how to change over flow lines so fast that there is no reason for a long production run of one type. Since the flow industries have been investing for years in inflexible equipment that resists quick changeover, it is not an easy fix.

In the assembly industries it tends to be an easy fix. Assembly—of personal computers, washing machines, boats, trucks, furniture, and hundreds of thousands of other products—is still largely manual. Humans are adaptable and can change from one model to another with ease—and efficiency, too. Assembly is efficient, however, only if the work place is orderly, with every part and tool exactly placed. If the assembler has to search, the efficiency is gone.

In Japanese Manufacturing Techniques I told about working for the fastest bricklayer in North Dakota and about how he yelled at me if I didn’t place bricks so that he could reach and find them without looking. That concept—exact placement of all the parts to eliminate search—has enabled the world’s motorcycle manufacturers and some tractor producers to change models after each unit. That is called mixed-model production, and the lot size is one.

Ten years ago, all motorcycle and tractor manufacturers produced in large lots: maybe five hundred of model A; then shut down for a day or two to change over for a run of five hundred of model B; and so on.

Marketing hates this. Marketing might come to manufacturing and say, Which model are you running this week? Manufacturing says, model A.

Oh, that’s too bad. We, ah, overestimated demand for A. In fact, we have a whole warehouse full. When are you going to make E? Manufacturing looks it up in the master schedule: Week 9.

"That’s bad, too. We underestimated demand for E. We’re out and losing sales. Can you possibly move it forward in the schedule? Manufacturing replies, No way. Our suppliers will not begin delivering raw materials until week 8."

Manufacturing then blames marketing for doing a bad job of forecasting. The fault is not marketing’s. Manufacturing gets the blame, because the production schedule pushed the planning horizon out to week 9, and it is impossible to guess (forecast) right that far out.

Now, at Harley-Davidson, Honda, Kawasaki, Yamaha, John Deere tractor, and the others, some of every model is made every day. Marketing therefore has some to sell every day. If marketing comes to manufacturing and says, Can you increase model E by 10 percent and decrease A by 10 percent next week? manufacturing says, Yes, we can. The assemblers are quick-change artists. The makers of the components parts could still be an obstacle, but they can learn quick-change artistry, too.

If a world-class manufacturing effort fails to make it easier for marketing to sell the product, then something is wrong.

High-Variety JIT Production

Some plants or parts of plants seem doomed to have long manufacturing lead times. Western manufacturers of machine tools and thousands of kinds of industrial components, from motors to pumps to hydraulics, take weeks, often months, to produce something. The problem is that those manufacturers are high-variety, low-volume job shops. Ten thousand different part numbers is normal, fifty thousand is not uncommon, and no one knows which ones are going to be needed in the next customer order. There may be four thousand work orders open in the plant at any given time, with a hundred completed and a hundred new ones added every day. How can such a production environment be anything but chaotic?

We know the answer. It is to divide the ten thousand part numbers into families—production families, not marketing families. A production family is a group of parts that follow about the same flow path. Say, for example, that five hundred of those ten thousand parts go from blanking to grinding to drilling to welding to painting. You empty out a square area on the floor and move the following into it in a U-shaped loop: one blanking press, one grinder, one drill press, one welding station, and one paint dip tank. If there is a lot of drilling to be done, move two drill presses into the loop.

The result is a cell, a mini-production line, almost a pipeline that similar parts flow through. The machines are so close together that there is no need for a container, storage rack, or fork-lift truck. An operator, chute, or simple transfer device can move one piece at a time from station to station. Different part types are made in the cell, but all types go through the same machines (a few part numbers may skip one or more of the stations). Also, the parts in the family have similar setup times, cycle times, tool and fixture requirements, and needs for inspection. While the cells do not make the same part over and over again, they make the same family of parts over and over again, hence the term family repetitive production in Figure 1-1. The figure also shows the three other modes of production, discussed above, that are valid for the world-class manufacturer.

Next, find another family and move the needed machines and work stations into cell 2. Then create cell 3, and so on. Engineers sometimes call this approach group technology, although many prefer to use the more descriptive term cellular manufacturing.

The approach is much more than industrial engineering and plant layout, however. Cells create responsibility centers where none existed before. A single supervisor or cell leader is in charge of matters that used to be fragmented among several shop managers. The leader and the work group may be charged with making improvements in quality, cost, delays, flexibility, worker skills, lead time, inventory performance, scrap, equipment up time, and a host of other factors that distinguish the world-class manufacturer.

Large numbers of Western manufacturers are following this path in their quest to become world-class. The machine-tool, aerospace, and shipbuilding industries are especially active in reorganizing their plants into cells. That is natural in view of the mind-boggling numbers of parts that go into large machines, ships, aircraft, rockets, and tanks.

Figure 1-1. Repetitive Manufacturing

General Electric has transformed its dishwasher plant in Louisville, Kentucky, into a WCM showcase, and moving machines into cells was a basic step. Punch presses that had been in punch-press shops were dispersed to form cells or spurs off other cells or production lines. Figure 1-2 is a photo of one of the moved presses. A sign tacked to it proudly proclaims point-of-use manufacturing. Other presses and other machines around the plant have the same kinds of signs. GE’s success in transforming the dishwasher plant has served as a model for the rest of GE’s Appliance Park in Louisville. Refrigerator, range, and washer plants are being converted the same way.

Universals of Manufacturing

The metal fabrication industries have no prior claim on cellular manufacturing. It is emerging as a prescription for much of the world of work, on a par with Do it right the first time. Most of our plants’ facilities and people are organized with giant barriers to problem-solving, and the same goes for most of our offices. No one is in charge. Distances between processes are too long for decent coordination. Flow times are too long for us to reconstruct chains of causes and effects when things go wrong—and they go wrong so often.

The immensity of the task would be daunting if we were unsure of what paths to take. We know what paths to take, because there are many role models. Western manufacturers that have executed the WCM formula have been getting the same spectacular results that Japanese manufacturers did a bit earlier: product defects down from several percentage points to just a few per million pieces, and lead times cut by orders of magnitude. Knowing what it takes to get such results turns on the adrenalin pumps. The competitor whose pump does not get primed is the loser.

Figure 1-2. Point-of-Use Manufacturing

That is not to say that the company or plant involved in the WCM quest is completely surefooted. How, for example, can progress be measured? How do the movers get reinforcement so that they stay inspired? The answer is to choose the right goals of improvement and to organize the enterprise for continual progress against those goals. A host of WCM subgoals can be contained within two overriding goals. One is reduction of deviation, and the other is reduction of variability.

Deviation Reduction

Deviation