The original small-lot, quick-change techniques pioneered at Toyota in the 1960s are a striking achievement, but we caution readers not to take quick-change machines still producing batches, however small, as an end in themselves. Any changeover requiring any loss in production time and any machine which must run at a rate far out of step with the rest of the production sequence can still create muda. The end objective of flow thinking is to totally eliminate all stoppages in an entire production process and not to rest in the area of tool design until this has been achieved.
Let’s tie all of these techniques together by showing what a lean bicycle production process looks like, as shown in Figure 3.2 . First, note that the same number of bikes are being produced but that the plant is more than half empty, in large part because all of the in-process storage areas have disappeared. Although the diagram cannot show this, the human effort needed to produce a bicycle has been cut in half as well, and time through the system has been reduced from four weeks to four hours. (We’ll talk in Part II about what to do with people no longer needed for their traditional tasks as muda is eliminated. Protecting their jobs by finding them other productive tasks is a central part of any successful lean transition.)
F IGURE 3.2: L EAN B ICYCLE P LANT L AYOUT AND F LOW
The diagram does show that single large machines have been broken down into multiple small machines, in particular the washing systems and paint booths, so that bikes can proceed continuously, one at a time, from tube cutting to mitering to bending to welding to washing to painting to final assembly without ever stopping. In this arrangement the inventory between workstations can be zero and the size of the work team can be geared to the production volume of the cell, with high-volume cells having more workers than low-volume cells. Finally, note that the track assembly operations have been eliminated. When production is broken into product families, it is often the case that no family accounts for the kind of volume needed for track assembly. Remarkably, manual advancing of the product through assembly is often cheaper .
Because the work flow has been so drastically simplified, the MRP system and the accompanying expediters are no longer needed to get parts from step to step. (MRP still has a use for long-term capacity planning for the assembler firm and its suppliers.) When the sequence is initiated at the end of final assembly, work progresses from each station to the next in accordance with takt time and at the same rate as final assembly.
The entire product team including the team leader, the production engineer, the planner/buyer, the TPM/maintenance expert, and the operators (collectively the heart of the lean enterprise) can be located immediately adjacent to the machinery for each product cell. Because the process machinery currently available for these operations in the bicycle industry either makes very little noise inherently—for example, paint—or can be shielded so that very little noise escapes into the team area—the mitering step—it’s possible to lay out activities so everyone can see the whole operation and its status at a quick glance.
A final point about the cells which is hard to illustrate with a diagram is that the work in each step has been very carefully balanced with the work in every other step so that everyone is working to a cycle time equal to takt time. When it’s necessary to speed up or slow down production, the size of the team may be increased or shrunk (contracting or expanding job scope), but the actual pace of physical effort is never changed. And when the specification of the product changes, the right-sized machines can be added or subtracted and adjusted or rearranged so that continuous flow is always maintained.
RIGHT LOCATION
Only one more flow technique needs mentioning, which is to locate both design and physical production in the appropriate place to serve the customer. Just as many manufacturers have concentrated on installing larger and faster machines to eliminate direct labor, they’ve also gone toward massive centralized facilities for product families (sometimes called “focused factories”) while outsourcing more and more of the actual component part making to other centralized facilities serving many final assemblers. To make matters worse, these are often located on the wrong side of the world from both their engineering operations and their customers (Taiwan in the bicycle case) to reduce the cost per hour of labor.
The production process in these remotely located, high-scale facilities may even be in some form of flow, but launching products and improving the process machinery is much harder (because the core engineering skills are on the other side of the world), and the flow of the product stops at the end of the plant. In the case of bikes, it’s a matter of letting the finished product sit while a whole sea container for a given final assembler’s warehouse in North America is filled, then sending the filled containers to the port, where they sit some more while waiting for a giant container ship. After a few weeks on the ocean, the containers go by truck to one of the bike firm’s regional warehouses, where the bikes wait until a specific customer order needs filling, often followed by shipment to the customer’s warehouse for more waiting. In other words, there’s no flow except along a tiny stretch of the total value stream inside one isolated plant.
The result is high logistics costs and massive finished unit inventories in transit and at retailer warehouses. Another consequence is obsolete goods, eventually sold at large discounts, created by the need to place orders based on forecasts months in advance of demonstrated demand. When carefully analyzed, these costs and revenue losses are often found to more than offset the savings in production costs from low wages, savings which can be obtained in any case by locating smaller flow facilities incorporating more of the total production steps much closer to the customer. (We’ll return to this point in Chapter 10 on Japan, because wrong location, rather than high wages, lies at the heart of Japan’s current competitive dilemma.)
Applying Flow Thinking to Any Activity
Flow thinking is easiest to see in conventional, discrete-product manufacturing, which is where flow techniques were pioneered. However, once managers learn to see it, it’s possible to introduce flow in any activity and the principles are in every case the same: Concentrate on managing the value stream for the specific service or good, eliminate organizational barriers by creating a lean enterprise, relocate and right-size tools, and apply the full complement of lean techniques so that value can flow continuously. At the end of this volume, in Chapter 13 , we’ll apply lean thinking to a wide range of activities besides traditional manufacturing.
Flow in Work; Work as Flow
So far, we have been talking about the flow of value as if the needs of the customer and the investor are the only ones which count. However, we all know from our daily lives that our experience as producers (that is, as employees and workers) is often far more significant than our activities as consumers or investors. What does the transition to flow mean for the experience of work ?
Let’s begin with a brief look at the recent research findings of the Polish-born psychologist Mihaly Csikszentmihalyi, now at the University of Chicago. He has spent the last twenty-five years reversing the usual focus of psychology. Instead of asking what makes people feel bad (and how to change it) he has explored what makes people feel good, so that positive attributes of experience can be built into daily life.
His method has been to attach beepers, which sound at random intervals, to his research subjects. When the beeper sounds, the subject is asked to record in a notebook what she or he was doing and how they were feeling. After sifting decades of notebook data from thousands of subjects around the world, he has reached some very simple conclusions.
The types of activities which people all over the world consistently report as most rewarding—that is, which make them feel best—involve a clear objective, a need for concentration so intense that no attention is left over, a lack of interruptions and distractions, clear and immediate feedback on progress toward the objective, and a sense of challenge—the perception that one’s skills are adequate, but just adequate, to cope with the task
at hand.
When people find themselves in these conditions they lose their self-consciousness and sense of time. They report that the task itself becomes the end rather than a means to something more satisfying, like money or prestige. Indeed, and very conveniently for us, Csikszentmihalyi reports that people experiencing these conditions are in a highly satisfying psychological state of flow. 10
Csikszentmihalyi’s classic flow experience is rock climbing, where the need for concentration is obvious and the task itself is clearly the end, not a means. Participation sports less dangerous than rock climbing, interactive games, and focused intellectual tasks (such as writing books!) are often mentioned by Csikszentmihalyi’s respondents as flow experiences. However, traditional work-related tasks are only rarely mentioned despite the fact that work is rated the most important overall life activity. This is for a good reason. Classic batch-and-queue work conditions are hardly conducive to psychological flow. The worker can see only a small part of the task, there is often no feedback (much less immediate feedback), the task requires only a small portion of one’s concentration and skills, and there are constant interruptions to deal with other tasks in one’s area of responsibility.
By contrast, work in an organization where value is made to flow continuously also creates the conditions for psychological flow. Every employee has immediate knowledge of whether the job has been done right and can see the status of the entire system. Keeping the system flowing smoothly with no interruptions is a constant challenge, and a very difficult one, but the product team has the skills and a way of thinking which is equal to the challenge. And because of the focus on perfection, to be further explored in Chapter 5 , the whole system is maintained in a permanent creative tension which demands concentration.
Flow Is Not Enough
We’ve now seen striking examples of what happens when the value stream flows smoothly. What’s more, there is absolutely no magic involved. Any organization can introduce flow in any activity. However, if an organization uses lean techniques only to make unwanted goods flow faster, nothing but muda results. How can you be sure you are providing the services and goods people really want when they really want them? And how can you tie all the parts of a whole value stream together when they can’t be conducted in one continuous-flow cell in one room? Next you need to learn how to pull.
CHAPTER 4
Pull
Pull in simplest terms means that no one upstream should produce a good or service until the customer downstream asks for it, but actually following this rule in practice is a bit more complicated. The best way to understand the logic and challenge of pull thinking is to start with a real customer expressing a demand for a real product and to work backwards through all the steps required to bring the desired product to the customer. Bob Scott’s bumper for his out-of-production 1990 Toyota pickup truck provides a mundane but perfectly typical example.
In August 1995, Bob Scott backed his pickup into a pole near Glenside, Pennsylvania, and bent his rear bumper to a point where it couldn’t be straightened. He was determined that his truck look sharp—it was originally ordered with the “deluxe” chrome bumper at extra cost—and the severity of the dent also meant that the trailer hitch on the bumper was no longer safe to use. He needed a new bumper.
When Bob Scott took his pickup to Sloane Toyota in Glenside to get a new bumper installed, he touched off a pull sequence just at the point Toyota was taking a major step in its decade-long effort to synchronize the effort of its dealers, its parts distribution system, and its suppliers so customers could truly pull the flow of value all the way through a highly complex production and service system.
The Bad Old Days of Production
If Bob Scott had wrecked his bumper a year or two earlier, nothing would have happened immediately. When he tried to pull, Sloane Toyota wouldn’t have had the right bumper on hand for his out-of-production vehicle. Using a traditional stocking system, it’s simply impractical for a car dealer to keep on hand a wide range of replacement parts for older vehicles. With about ten thousand part numbers per vehicle, the carrying cost of the inventory would be staggering .
Instead, Sloane Toyota would have needed several days to get a bumper shipped by truck from a Toyota parts warehouse or used expensive overnight freight in order to get it delivered the next day. Bob Scott would have lost the use of his vehicle for some period of time or paid a premium if he wanted it the next day, and in either case would have been an unhappy customer.
Yet even as he waited, there would have been stacks of the precise bumper needed, indeed mountains of them, in Toyota’s parts warehouses and at the bumper maker because no satisfactory method was in place for pulling. To see why this was so and to understand what is being done to implement a true pull system all the way along the value stream, let’s go back in time and very nearly to the headwaters of the stream, to the Bumper Works factory in Danville, Illinois, which made the bumper Bob Scott cracked.
Shahid Khan, the president of Bumper Works’s parent firm, Flex-N-Gate Corporation, is practically a cliché of the American dream. He came to the United States from Pakistan when he was sixteen to go to engineering school at the University of Illinois in Urbana. To put himself through school, he got a job running a massive stamping press in the down-and-out Bumper Works factory in nearby Danville. When he graduated he became the engineering director of Bumper Works and then, by the time he was twenty-eight, he had raised the funds to buy the company.
When Khan entered Bumper Works in 1970, he also entered the world of batch-and-queue. Bumper Works made chromed and painted steel bumpers in a variety of styles for customizing pickup trucks at the car dealership. It made large batches of each type of bumper—typically a month’s worth—before shifting production to the next model and sold the bumpers to newcar dealers and crash-repair body shops through a complex wholesale distribution system.
Because large batches were considered normal in this world, it was not important that it took sixteen hours to change over Bumper Works’s stamping presses. Because large batches of raw materials were considered unavoidable, Bumper Works had a warehouse at the end of its plant to receive flat sheets of steel by the ton from the steel company. And because the chroming company performing the key step in the middle of the production process also worked in a batch mode, Bumper Works piled up partially made bumpers in its intermediate goods warehouse until there was an enormous batch and then shipped them to the chromer all at once.
When the chromer shipped them back, all in a batch, they were run through a final assembly operation (to install inner reinforcing bars, attachment brackets, and cosmetic coverings), stored once more in a finished goods warehouse, and sent in a batch to the customer according to a predetermined schedule .
As Shahid Khan grew his business in the 1980s, he began to supply replacement bumpers to the service parts organizations of the American Big Three auto companies and he did very well. His batch thinking and their batch thinking were a match. However, Kahn had always set his standards very high, so in 1984 he approached Toyota about supplying bumpers for the pickups they were importing from Japan. This would give him their “crash” parts business as well.
In 1985, Bumper Works was signed on as a supplier for a small volume of Toyota business, and in 1987 won a sole-source contract for the bumpers on the new version of Toyota’s small pickup (the model Bob Scott bought). By 1989, Bumper Works was Toyota’s sole bumper supplier for North American needs.
There was only one problem: Bumper Works’s production system was still a classic case of batch-and-queue. Toyota took Shahid Khan and his senior managers on their first trip to Japan late in 1989 and walked them through showcase lean suppliers, but as Khan remembers, “The light didn’t come on; I really couldn’t figure out how they could stay in business using the strange practices I saw.” So in May 1990, Toyota told Khan they were dispatching a lean sensei, 1 a master of the Toyota system, as Khan’s personal tutor.
In fact,
Toyota sent a number of sensei from its Operations Management Consulting Division, the group established in 1969 by Taiichi Ohno to promote lean thinking within Toyota and in the firms in its supplier group. 2 They stayed for months at a time, and by the end of 1992, they had totally transformed Bumper Works—a unionized, grimy operation using old tools in old facilities—into one of the best examples of lean production in North America.
Lean Production for Pull
The first thing the Toyota sensei noted at Bumper Works was the massive inventories and batches. Nothing flowed. Immediately right-sizing the massive stamping presses to permit single-piece flow was not possible, so the only solution was to drastically reduce their changeover times and shrink batch sizes. Changeover times were already down from sixteen hours in the mid-1980s to around two hours, but this was not nearly enough.
The Toyota sensei applied their standard formula that machines should be available for production about 90 percent of the time and down for changeovers about 10 percent of the time. Then they looked at the range of products Bumper Works would need to make every day. They concluded that the large presses would need to be changed over in twenty-two minutes or less and the small presses in ten minutes or less. (In fact, the numbers were soon down to sixteen minutes and five minutes, respectively.)
Next, the plant was physically reorganized so flat sheets flowed directly from the receiving dock to the blanking machine, which cut the steel into rectangular shapes just larger than a bumper. The blanks then went immediately to the adjacent cell of three stamping presses, where they were given their shape. Next, they were shipped at frequent intervals to the outside-the-plant chroming operation and returned to the welding shop adjacent to the stamping presses. There, the inner and outer parts of the bumper plus the brackets for attaching the bumper to the vehicle were welded together. Finally, the bumpers went straight to the shipping dock just in time for scheduled shipment. But they flowed only when pulled by the next step. That is, the blanking machine did nothing until it received a signal from the stamping machines and the stamping machines made nothing until instructed to do so by the welding booth. Each activity pulled the next. The shipping schedule and takt time became the pacemaker for the entire operation.
Lean Thinking Page 8