Lean Thinking

by Daniel T Jones

  In 1925, aircraft engine design was still a trial-and-error process of building a prototype and testing it to failure, then strengthening the part that failed and testing the design again. Rentschler knew that the key to success was to attract the most experienced engineers in the industry and to quickly create a scaled-up version of the Wright Whirlwind that would work well on the first try. He soon convinced several senior engineers from Wright to join him at Pratt, and his new design team made spectacular progress.

  In only nine months, Pratt’s six engineers and twenty craftsmen (out of a total payroll of thirty including Rentschler) were able to design the new Wasp engine (with about two thousand parts), incorporate a key processing innovation to save weight, 9 build three prototypes, and have them ready for testing by potential buyers. When tested, the Wasp produced 50 percent more power (425 horsepower) than the Wright Whirlwind air-cooled engine and weighed only 650 pounds compared with the 1,650-pound Curtiss Liberty liquid-cooled engine producing the same horsepower. (The latter engine was the standard design used by the U.S. military at that time.)

  Orders from both military and commercial customers poured in, and by 1929, Pratt & Whitney was the world leader in the tiny but rapidly growing aircraft engine business. The Pratt engine quickly established a reputation for reliability and was chosen for the next generation of commercial transports, beginning with the Ford Tri-motor. (The corporate logo—an American Eagle encircled by the words “Pratt & Whitney—Dependable Engines”—was affixed to every engine from the beginning and became familiar to airline passengers around the world.) In 1929, Rentschler was able to buy out the Pratt & Whitney machine tool company’s interest and build a new headquarters and vast production facility in East Hartford. 10

  In the beginning, the three key activities of Pratt & Whitney—design of new products, order-taking, and production—could be accomplished effectively in an utterly simple organization. Indeed, the initial production run of two hundred Wasp engines for the U.S. Navy was designed and then built in one large room by a group of highly skilled machinists directly interacting with the tiny group of product engineers.

  By the early 1930s, as production volumes grew from dozens of engines to hundreds, organizational differentiation like that undertaken by Lantech seemed to be required. Departments were created for each major activity—sales, engineering, prototype building and testing, quality control, purchasing, production, and service. Shops were created inside each department for specialized activities; for example, heat treatment, paint, and final assembly shops were established within the Production Department. As long as Pratt had only one product in development (the Hornet, which followed the Wasp and increased horsepower to 500) and only the Wasp in production, this system worked well without the need for cross-functional management.

  However, by the mid-1930s, as Pratt expanded its product offerings to include the 300-horsepower Wasp Junior and the 800-horsepower Twin Wasp, and conducted experiments with a range of new engine configurations, something more was needed. A new position was created, the “project engineer” reporting to the heads of engineering and production. The project engineer was given the job of coordinating all of the activities involved in the design, production, and installation in the customer’s airplane of a specific product line (such as the Wasp) as it moved through the host of departments and shops. 11 The project engineer was only a coordinator with no dedicated employees or resources—in today’s terminology, a “lightweight” program manager 12 —but a startling conceptual leap had been taken, going far beyond a purely functional organization and common management practices at that time. Indeed, the concept of a project engineer to oversee the entire value stream foreshadows the lean principles described in this book.

  As Pratt grew in the 1930s, changes were required in the factory as well. Initially, all of Pratt’s metal-cutting tools had been relatively small machines—lathes, drills, millers, borers, etcetera—which could be lined up in the actual sequence of work flow. 13 For example, in 1936, the Cylinder Shop in the East Hartford plant was organized as follows:

  “… the first shop … immediately following the raw material inspection and the Experimental Department is the Cylinder unit. On one side of the main aisle are produced all of the steel cylinder barrels. On the other side are produced all of the aluminum alloy heads and in addition, the barrels are assembled to the heads, together with valve seats, bushings, valve guides and other minor parts, so that when the cylinder is ready to leave the department, it … proceeds directly to the Finished Stores Department. … with spare parts requirements, there are approximately 50 separate active cylinder designs. The equipment has been laid out so that the machines are in a sequence and the raw material proceeds in a straight line. Naturally, not all machines are required for any given cylinder.” 14

  Similar shops had been created for master links and rods, crankcases, crankshafts, pistons, rocker shafts and valve guides, and cams. These sound remarkably like the work cells for complete components we have encountered throughout this volume and it is clear that Pratt’s operations managers at that time had at least a rudimentary notion of flow: “…the scheme of production is a relatively simple one. Raw material is received by rail or truck through the front of the shop [factory] and then flows through the various manufacturing departments to the Finished storeroom at the rear.” 15

  However, it is also clear that continuous flow was strictly limited to assembly and those activities which could be performed with simple machines. Special shops were created for machining parts made from magnesium and hard steel alloys as well as for heat treating, painting, and polishing. Because most of the parts in each completed component needed at least some of these treatments, much material was moved back and forth from shop to shop.

  In addition, an elaborate system of centralized storage areas, tool cribs, and inspection stations was in place. It was taken as a given that quality inspections must be done independently of the primary workers by technicians reporting to the head of that function, not to the head of production, and that production could be more tightly controlled by storing tools, fixtures, and parts-in-process in a central location. These decisions meant that every part and every worker moved to a central storage area between each major production stage and during setups for the next job.

  Finally, the philosophy of the company was that many defects could only be detected by test running of completed engines. Therefore, a row of test cells went across the entire rear section of the factory. Each engine was run for eight to thirteen hours, then completely disassembled. The parts were inspected, replaced as necessary, and reassembled. The engine was run for five to twelve more hours and then, in the event no problems were found, it was shipped. 16 As we will see, this final safety net created an “assemble it, then tinker until we get it right” mentality which persisted at Pratt until 1994.

  Even with a relatively simple plant layout and product line, it is clear that, by 1936, Pratt was having to work very hard to get products through the system. An organized system of “shortage lists” and “follow-up” (read “hot lists” and “expediting”) was in place and the assistant general manager was eager to tell an audience of peers that “high-tech” help with these tasks was already in place:

  It might also be of interest to point out that all the shortage lists and schedule sheets are made on electrically operated Hollorith machines 17 from punch cards made in the store room office so that these lists are printed and supplied to the Schedule Dept. and Follow-up Dept. in a neatly printed segregated form and without any delay. This is a major factor in the efficient control of shop production. 18

  In short, Pratt & Whitney was for the second time moving down the path from a lean workshop to a massive mass producer. The major innovation during the second transition was that the growing emphasis on complex tools housed in specialized departments could be supported by automated information management to shepherd products from raw materials to finished goods.

>   What should have been the major organizational innovation, the project engineer, never worked as planned. By 1939, Chief Engineer L. S. Hobbs was writing to his superiors, “It has been fairly obvious from the time of our institution of the Project Engineer system that in reality the system has not functioned as such.” 19 Instead, the project engineer was a lightweight manager within the product development organization and products moved through sales, scheduling, production, and installation as best they could, with expediting by the centralized information management system but with no individual or team fully responsible for their progress.

  World War II as the Engine of Mass Production

  When the flow of orders increased from hundreds to hundreds of thousands in World War II, 20 Pratt made the final leap to mass production in the factory. A shortage of skilled workers meant that the new machine tools for the war effort were designed for very specialized tasks with only modest skill requirements by the operator. The number of shops, each assigned a narrow task, grew dramatically as the division of labor continued. What’s more, the volume of orders meant it was often feasible to dedicate a given machine to a given part, perhaps for years at a time, so the need to do frequent setups was reduced. Work-in-process, travel within the production system, rework in the test department at the end of production, and managerial complexity all increased but engine output increased even more, and the latter was the only important consideration during the war.

  Not surprisingly, by the end of the war, the mentality of the workforce had changed. Rather then being highly skilled, semi-independent craftsmen, the new workforce was much more narrowly trained, assigned to largely interchangeable jobs, and under much tighter management control. A conventional union had little appeal to the initial generation of craftsmen at Pratt, but by 1945, a different mentality and a different shop-floor reality created an environment in which the International Association of Machinists easily won an election to unionize the workforce. 21 A maze of work rules and grievance procedures soon followed as a mirror image of the division of labor instituted by management.

  The second important consequence of World War II was in product development, where the growing complexity of designs and the need to extract ever more power from the basic radial engine configuration created the need for very deep technical functions. The key disciplines were materials scientists to develop new materials, structural engineers to address weight and durability problems, aerodynamicists to tackle the problem of airflow and drag through and around the engine, and mechanical engineers able to design and link together the thousands of individual parts required for each engine. Each of these specialties gained its own department within the vast Pratt & Whitney Engineering Division.

  By the end of the war, Pratt’s Wasp Major engine had thirty-six cylinders in four rows turning a single crankshaft. It was both supercharged and turbocharged to yield 4,600 horsepower (compared with the 425 horsepower of Pratt’s original nine-cylinder Wasp). Along with the turbo-compound engine being developed at the same time by the Curtiss-Wright Company (the merged successor firm to Wright Aeronautical and Curtiss), the Wasp Major was one of the most complex pieces of purely mechanical apparatus ever devised. 22

  The Jet-Propelled Eagle

  During World War II, the U.S. government directed Pratt and Curtiss Wright to stick to what they knew: designing and building reciprocating piston engines. Other American firms with no previous experience in aircraft engine building (General Electric, Westinghouse, and Allison) took the lead in jet engines, and by war’s end Pratt was the clear world leader in a technology with no future. What was worse, it was nowhere with the technology that did have a future—the jet turbine.

  In 1946, P&W took a tremendous but unavoidable gamble by abandoning research on piston engines. It attempted to leapfrog its new jet-age competitors with a two-shaft, axial-flow jet engine considerably larger and more complex than any previously envisioned. Curtiss-Wright, by contrast, continued to elaborate the piston engine with its turbo-compound version for the Douglas DC-7 and the Lockheed Super Constellation in the early 1950s. C-W exited the industry when jet aircraft quickly supplanted these final iterations of the piston-engine airplane.

  Jet engines were based on different principles but required many of the same technical skills in Pratt’s existing engineering functions. The materials scientists were now concerned with managing the extreme heat in the hot parts of the engine. The structural engineers were concerned with vibration in complex turbo-machinery. The aerodynamicists were concerned with airflow past the compressor and turbine blades. The mechanical engineers were still concerned with detailed design of the thousands of parts, now rotating rather than reciprocating, aggregating to a complete engine. The big difference was that the nature of the knowledge was now highly scientific and the amount of effort required was much greater. 23

  Pratt’s technical functions became deeper and more silolike as the nature of the necessary knowledge became more arcane. The project engineer system within product development groaned as the walls between functions thickened, giving rise to the “Pratt Salute” of arms crossed and pointing in opposite directions to assign fault to other departments for all design and manufacturing problems.

  The production system, for its part, was remarkably unaffected by the jet age. Highly specialized machine tools—joined in the 1970s by truly massive special-purpose devices such as electron-beam and fusion welders—were located together in shops inside departments to feed batches of parts to a bench-assembly operation creating the finished engine. Every engine was then extensively tested and “tuned” (reworked) before shipment. The common joke was that the average part traveled farther inside Pratt’s plants during production than it did in airline service. But there seemed to be no better way.

  Pratt’s leap to jet propulsion in 1946 produced a technical and commercial triumph by 1952. The P&W J-57 engine powered the American eight-engine B-52 bomber first flown in that year. Slightly modified and renamed the JT3, this engine captured 100 percent of sales for the initial versions of the four-engine Boeing 707 and Douglas DC-8 by the end of the decade. P&W quickly followed up with an entirely new engine, the JT8D, to power the entire world fleet of Boeing three-engine 727s and two-engine Douglas DC-9s and the initial versions of the two-engine Boeing 737. When the American military awarded Pratt a contract in 1970 as the sole source of the F100 engine for the F15 and F16 fighter planes, the company totally dominated the global aircraft engine business. Indeed, at the end of the 1960s, Pratt held a staggering 95 percent share of the world’s commercial jet engine market (outside the Russian bloc) and nearly a 50 percent share of American military orders.

  In the process of reaching industry dominance, Pratt and its organization fine-tuned and hardened the standard features of a mass producer. Tasks were finely divided in physical production, with specialized machines making batches of parts with long lead times. During product development lightweight team leaders coordinated engineering efforts across thick functional walls.

  In fact, this system was adequate if not perfect for its environment. For decades aircraft engines were ordered by regulated airlines—competing on service but not on price—and by the military—interested in wartime performance with purchase price only a secondary consideration. In addition, advances in materials science and aerodynamic analysis meant that each new generation of product could achieve substantial performance improvements. As long as Pratt’s technical depth could produce products which performed better than competing products, the fact that they took unnecessary time to design and manufacture, cost more than they needed to, and sometimes failed to perform properly when first launched in service could all be overlooked.

  During this golden age the specification of new products at Pratt tended to work backwards. The senior engineers decided what technologies were ready for introduction in the next product generation and specified the engine configuration needed to utilize them. They then calculated the production cost and
selling price as a sort of resultant. Once in production, costs were not rigorously tracked, but instead rolled upin the profit-and-loss statement in the president’s office, by which point it was too late to do much about them.

  By the 1980s, as airframe makers began to offer a choice of two or three engines (from Pratt, GE, and Rolls) for each wide-body aircraft type, the issue of production costs was confused further by the industry practice of progressively discounting prices for new engines, eventually to far below costs. 24 This was done in the hope that profits could be recovered from sales of spare parts, in particular turbine blades, where the engine makers had a monopoly. For example, the spares purchased by an airline during the operating life of a JT8D were likely to equal five times the initial purchase price of the engine. In this environment, the manufacturing side of the jet engine companies could easily get confused about the importance of costs—after all, the engines were being sold for prices far below any imaginable production cost.

  The final feature of this mature mass-production system was its peculiar method of order-taking. The twenty-four-month lead times needed to physically produce an engine conspired with the three-year lead time needed to produce the complete airplane to create gigantic waves of orders for jet aircraft in the postwar era, 25 as shown in Figure 8.1 .

  F IGURE 8.1: C OMMERCIAL J ET O RDERS

  As the airline industry emerged from recessions, aircraft customers signed up for planes and engines they might not need in order to ensure themselves a place in the production queue, while sales departments often made special deals for large orders even when sales were booming in order to hold market share and protect the spares base. These orders could suddenly evaporate when the economy slumped, but waves of military orders often offset slumps in civilian demand and spare-parts purchases often went up when new engine deliveries slumped after 1980, as shown in Figure 8.2 .