Perhaps "the most ferocious foe of waste human history has produced" was Taiichi Ohno (1912-90). Ohno-sensei was the father of the Toyota Production System, which is the conceptual foundation of the world's premier manufacturing organization, and one of the pivotal innovators in industrial history. His approach, though adopted successfully by Toyota, remains rare in Japan. However, it has shown remarkable results in America and elsewhere in the West, and is poised for rapid expansion now that it has been systematized by industrial experts Dr. James Womack and Professor Daniel Jones. With their kind permission, we gratefully quote and paraphrase their book, Lean Thinking, in the hope that more business leaders will read it in full.
Ohno created an intellectual and cultural framework for eliminating waste, which he defined as "any human activity which absorbs resources but creates no value." He opposed every form of waste. Womack and Jones restated thus his classification of the forms of waste: "mistakes which require rectification, production of items no one wants so that inventories and remaindered goods pile up, processing steps which aren't actually needed, movement of employees and transport of goods from one place to another without any purpose, groups of people in a downstream activity standing around waiting because an upstream activity has not delivered on time, and goods and services which don't meet the needs of the customer." Ohno called these muda, which is Japanese for "waste," "futility," or "purposelessness." Each of these classes of muda involves a whole family of blunders, which range from activities like having to inspect a product to see if it has the quality it should have had in the first place (an unneeded process step) to filling a new-car lot with vehicles that meet no specific demand, if the cars were wanted, customers would have bought them already, and then discounting them enough to sell them. Ohno's and his students' vast practical experience helped them to develop penetrating modes of perception, mental "muda spectacles", that reveal the previously invisible waste all around us.
So where is all this muda? Start, say, by visiting a job site where builders are constructing a custom house. You'll notice periods of recurrent inactivity. But these lags aren't taking place because the workers are lazy. Builder Doyle Wilson discovered that five-sixths of the typical custom-house construction schedule is spent in waiting for specialized activities to be completed and fitted into a complex schedule, or in reworking, tearing out and redoing work that was technically wrong or that failed to meet the customer's needs and expectations. Eliminating even part of that wasted time can create a huge competitive advantage for a savvy construction firm.
Or take a much more familiar experience: air travel. Often you can't get a direct flight to where you want to go. Instead, you must somehow get to a major airport, fly in a large airplane to a transfer point quite different from your actual destination, become "self-sorting cargo" in a huge terminal complex once you arrive there, and board another large plane going to the destination you originally wanted. Most travelers tolerate this because they are told that it's a highly efficient system that fully utilizes expensive airplanes and airports. Wrong. It looks efficient only for the tautological reason that the airplanes are sized for those large hubs, which are designed less for efficiency than to monopolize gates and air-traffic slots, thus reducing competition and economic efficiency as well as convenience.
Much if not most air travel would cost less, use less fuel, produce less total noise, and be about twice as fast point-to-point by using much smaller and more numerous planes that go directly from a departure city to a destination. That concept, reinforced by turning around planes in fifteen instead of thirty minutes, is the secret of Southwest Airlines' profits. In contrast, most other airlines have established systems designed to transfer idleness from capital to customers. These systems are so riddled with waste that Jones once found nearly half the door-to-door time of a typical intra-European air trip to have been spent in waiting in ten different lines, seven baggage-handling operations, eight inspections asking the same questions, and twenty-three processing steps performed by nineteen organizations. Each was specialized to perform its own narrowly defined task "efficiently", in a way that ultimately added up to dreadful inefficiency for the customer. Removing inefficiencies like these through whole-system engineering of the firm is the next great frontier of business redesign.
The nearly universal antidote to such wasteful practices is what Womack and Jones call "lean thinking," a method that has four interlinked elements: the continuous flow of value, as defined by the customer, at the pull of the customer, in search of perfection (which is in the end the elimination of muda). All four elements are essential to lean thinking: For example, "if an organization adopts lean techniques but only to make unwanted goods flow faster, muda is still the result." The parts of the definition also functionally reinforce one another. "Getting value to flow faster always exposes hidden muda in the value stream. And the harder you pull, the more the impediments to flow are revealed so they can be removed. Dedicated product teams in direct dialogue with customers always find ways to specify value more accurately[,] and often learn of ways to enhance flow and pull as well."
Value that flows continuously at the pull of the customer--that is, nothing is produced upstream until someone downstream requests it--is the opposite of "batch-and-queue" thinking, which mass-produces large inventories in advance based on forecast demand. Yet so ingrained is batch-and-queue, and so deeply embedded is the habit of organizing by functional departments with specialized tasks, that Womack and Jones caution: "[P]lease be warned that [lean thinking] requires a complete rearrangement of your mental furniture." Their basic conclusion, from scores of practical case studies, is that specialized, large-scale, high-speed, highly efficient production departments and equipment are the key to inefficiency and uncompetitiveness, and that maximizing the utilization of productive capacity, the pride of MBAs, is nearly always a mistake.
Consider the typical production of glass windshields for cars. Economies-of-scale thinking says that the giant float-glass furnace should be as large as possible: a theoretically ideal situation would be if all the flat glass in the world could be made in a single plant. Big, flat sheets of glass emerge from the furnace and are cut into pieces somewhat larger than a windshield. The glass is cooled, packed, crated, and shipped 500 miles to the fabricator. There, 47 days later, it's unpacked and cut to shape, losing 25 percent in the process. It is then reheated and drooped or pressed into the right curving shape. (Because each car model has different specifications, huge batches of windshields are shaped at once while a given set of dies is installed.) Then the glass is cooled, repackaged, and shipped 430 miles to the glass encapsulator. There, 41 days later, it's unpacked, fitted with the right edge seals and other refinements, repacked, and shipped another 560 miles to the car factory. There, 12 days later, it's unpacked and installed in the car. Over 100 days have elapsed and the glass has traveled nearly 1,500 miles, almost none of which contributes to customer value.
Each part of this sequence may look efficient to its proprietor, but in fact the cooling, reheating, unpacking, repacking, shipping, and associated breakage is all muda. An efficient system for manufacturing windshields would build a small plant at the same place as the car factory, and carry out all the steps in the production process in immediate succession under one roof, even though several machines and companies might be involved. The machinery would be sized to deliver windshields only as fast as the automotive assembly line "pulls" them in.
Traditional substitutions of complex machines for people can backfire, as Pratt & Whitney discovered. The world's largest maker of jet engines for aircraft had paid $80 million for a "monument", state-of-the-art German robotic grinders to make turbine blades. The grinders were wonderfully fast, but their complex computer controls required about as many technicians as the old manual production system had required machinists. Moreover, the fast grinders required supporting processes that were costly and polluting. Since the fast grinders were meant to produce big, uniform batches of product, but Pratt & Whitney needed agile production of small, diverse batches, the twelve fancy grinders were replaced with eight simple ones costing one-fourth as much. Grinding time increased from 3 to 75 minutes, but the throughput time for the entire process decreased from 10 days to 75 minutes because the nasty supporting processes were eliminated. Viewed from the whole-system perspective of the complete production process, not just the grinding step, the big machines had been so fast that they slowed down the process too much, and so automated that they required too many workers. The revised production system, using a high-wage traditional workforce and simple machines, produced $1 billion of annual value in a single room easily surveyable from a doorway. It cost half as much, worked 100 times faster, cut changeover time from 8 hours to 100 seconds, and would have repaid its conversion costs in a year even if the sophisticated grinders were simply scrapped.
Just as unwanted weight in a car or unwanted heat in a building is prone to compound and multiply, muda tends to amplify itself, because excessive scale or speed at any stage of production turns the smooth flow of materials into turbulent eddies and undertows that suck down earnings and submerge whole industries. Remember chapter 3's saga of the aluminum cola can? It takes 319 days in production to get to the customer's hand, then minutes to reach the trash bin. This is 99 and 96/100ths percent pure muda. For such a massive batch-and-queue system to produce what the customer perceives as an uninterrupted supply of cola requires huge inventories at every upstream stage to deal with unforeseen fluctuations in demand or delays in supply. Wherever there's a bottleneck, the supplier adds buffer stocks to try to overcome it, thereby counterintuitively making the stop-and-go traffic of the materials flow even worse.
All this results from the mismatch between a very small-scale operation--drinking a can of cola--and a very large-scale one, producing it. The production process is designed to run in enormous batches, at very high speeds, with very high changeover costs. But that logic is the result of applying to business organization precisely the same design flaw, discussed in the previous chapter at the level of components, namely, optimizing one element in isolation from others and thereby pessimizing the entire system. Buying the world's fastest canning machine to achieve the world's lowest fill cost per can presumably looks like an efficient strategy to the canner. But it doesn't create customer value at least cost, because of such expenses as indirect labor (in such forms as technical support), the inventories throughout the value chain, and the pervasive costs and losses of handling, transport, and storage between all the elephantine parts of the production process. Just as Pratt & Whitney's grinders looked fast and cheap per grind but were slow and costly per finished blade, from a whole-system perspective, the giant cola-canning machine may well cost more per delivered can than a small, slow, unsophisticated machine that produces the cans of cola locally and immediately on receiving an order from the retailer.
The essence of the lean approach is that in almost all modern manufacturing, the combined and often synergistic benefits of the lower capital investment, greater flexibility, often higher reliability, lower inventory cost, and lower shipping cost of much smaller and more localized production equipment will far outweigh any modest decreases in its narrowly defined "efficiency" per process step. It's more efficient overall, in resources and time and money, to scale production properly, using flexible machines that can quickly shift between products. By doing so, all the different processing steps can be carried out immediately adjacent to one another with the product kept in continuous flow. The goal is to have no stops, no delays, no backflows, no inventories, no expediting, no bottlenecks, no buffer stocks, and no muda. Surprisingly, this is as true for small- as for large-scale production.
(End of excerpt)
Download the entire chapter (PDF-144k)