In 1969, Boeing assembled a New Airplane Program (NAP) study group. Its goal was not to develop a new plane, but to review the company's past experiences with each of its major programs—the 707, 727, 737, and 747—so that problems, such as those incurred by the 737 and 747 programs, would not be repeated. As Neil Standal, a member of the NAP group who later became the 767 program manager, observed:
We knew that we were going to have another commercial airplane. But we didn't know what, or when, it was going to be. Our objective was to provide lessons for the future, to look at our history and decide what we had done right and what we had done wrong.
This process, called Project Homework, took three years and produced a long list of "lessons learned," as well as a reasonable idea of the costs of developing the next-generation airplane.
Meanwhile, pressures were beginning to mount within Boeing to launch a new airplane program. Salespeople were especially insistent, as T. A. ("T") Wilson, Boeing's chairman, recalled:
Our salespeople kept saying, "We need a new product." They didn't really care what it was, as long as it was new.
Because the company's last new plane, the 747, had been launched in 1966, there was also concern among the board of directors that Boeing's next generation of leaders was not being trained in the best way possible: by developing a new plane of their own.
In 1973, at Wilson's behest, Boeing initiated a new airplane study, naming it the 7X7 (X stood for development model). Key team members, including J. F. Sutter, the program's first leader, and Dean Thornton, who replaced Sutter after he was promoted to vice president of operations and development, were handpicked by Wilson. The team was given a broad charter: to define and, if approved, to develop, Boeing's next generation airplane.
The first stage of the process, called program definition, extended from May 1973 to December 1977 (see Exhibit 10-5). During this period, Boeing worked the puzzle of market, technology, and cost. Team members projected airline needs into the future to see if there were holes in the market not met by existing planes; considered alternative plane configurations; examined new technologies to see what might be available within the next few years; and estimated, in a preliminary fashion, likely development and production costs.
Forecasting the airframe market for the 1980s and 1990s was a complex and challenging risk. Market analysts began by talking directly with the major airlines to get their estimates of future needs. That information was then combined with econometric models to generate three forecasts— optimistic, conservative, and expected—for each market segment. Segments were defined by range of travel—short (less than 1,500 nautical miles), medium (1,500-3,000 nautical miles), and long (greater than 3,000 nautical miles)—and all forecasts were based on the following assumptions: continued regulation of the airline industry; continued airline preferences for routes that directly linked pairs of major cities; steadily rising fuel prices; and no new competition from other airframe manufacturers in the medium-range market. Complete forecasts were run annually and readjusted quarterly.
Boeing's expected forecast for 1990 was a total market of $100 billion. The critical medium range segment—the expected target of the new airplane—was estimated at $19 billion. In that segment, Boeing expected to capture 100 percent of domestic sales. Continued production of the 727 would meet most replacement needs, and the 7X7 would be positioned for market growth.
While these forecasts were being developed, another group was working on design specifications. After a year or two of study, the basics were de-
Exhibit 10-5. Critical Program Decisions and Reviews
Exhibit 10-5. Critical Program Decisions and Reviews
cided. Market research indicated that the new plane should carry approximately 200 passengers; have a one-stop, U.S. transcontinental range; and offer minimal fuel burn. The last requirement was regarded as especially important. With the rise in oil prices that followed the 1973 Arab oil embargo, fuel costs had become an ever-larger portion of airlines' operating expenses. Moreover, airline preferences were changing, as Frank Shrontz, president and CEO, observed:
In the old days, airlines were infatuated with technology for its own sake. Today the rationale for purchasing a new plane is cost savings and profitability.
Market needs were thus reasonably clear, at least within broad outlines. Designers, however, still faced a number of critical choices. All involved some aspect of the plane's basic shape.
The most vexing question was whether to design the 7X7 with two or three engines. A two-engine version would be lighter and more fuel efficient; a three-engine version would offer greater range. But exactly what were the trade-offs? And how far was engine technology likely to advance in the next few years? Boeing, after all, did not build its own engines, but bought them from one of three manufacturers: General Electric, Pratt & Whitney, and Rolls Royce. Airlines paid separately for airframes and engines; however, they could only choose engines that were offered for the airplane. (This was necessary because Boeing guaranteed the performance of every plane it sold.) Early in the 7X7 program, managers chose to offer engines from both General Electric and Pratt & Whitney, despite the additional time and expense that Boeing would incur. This decision was a direct outgrowth of the company's experiences with the 747. Managers felt that continued competition among engine manufacturers was essential to moderate costs. Equally important, competition was expected to provide a steady stream of improvements in engine technology.
The certification decision proved to be far easier than the choice between a two - and three-engine plane. In fact, for most of the program definition phase, the 7X7 team worked simultaneously on two- and three-engine models. Eventually, fuel efficiency won out—as one manager put it, "in those days, an engineer would shoot his mother-in-law for a tenth of a percent improvement in fuel savings"—and the two-engine version was selected.
Other key configuration decisions involved the wings and tail. Both decisions showed the family of planes concept in action, and the need for designs that were adaptable to future needs. The 7X7 was conceived originally as a medium-range aircraft; however, later additions to the 7X7 family were expected to target longer-range flights. Engineers therefore selected a wing size—3,000 square feet—that was larger than necessary for short- and medium-range flights. It added weight to the basic design, with some loss of fuel efficiency. But the design was highly adaptable: It could be used, without modification, on longer-range versions and stretched models with greater carrying capacity.
Because they were so complex, configuration decisions required the close coordination of marketing, engineering, and production personnel. The airlines were also intimately involved. After a new configuration was developed, Boeing's marketing managers brought it to the airlines, who reviewed, among other things, its flight characteristics, range, cruising speed, interior, cockpit, systems and operating costs. Their reactions were then fed back to designers, and the process was repeated. Haas observed:
Designing airplanes to best meet the unique requirements of customers is a difficult process. Each airline would prefer that it was designed a bit differently—a little longer, a little shorter, a few more people, a few less. Therefore, the configuration changes constantly.
Configuration decisions could not be made without assessing the technology that was then available. What was desired by the market might not be possible or economical given the current state of knowledge.
Technology development was an ongoing process at Boeing, and included such areas as structures, flight systems, aircraft systems (hydraulic and electrical), and aerodynamics. Each area had its own chief engineer, who was responsible for overseeing research, development, and application of the technology. The last requirement was regarded as especially critical, as David Norton, chief of technology, pointed out:
There is nothing that brings me up quicker than thinking of how long we have to live with our decisions. At Boeing, applying a new technology is as important as developing it. We had better be right.
When a new plane was proposed, engineers first reviewed all existing technology projects to see if any were appropriate. They asked three questions of every project:
(1) What is its ultimate value to the customer? (2) Is it an acceptable technological risk? and (3) Can it be incorporated within schedule and cost? Responsibility for answering these questions was divided among the chief engineers of each technology and a chief engineer in charge of the plane program. Line engineers therefore reported through a matrix, and were accountable to two bosses: the chief engineer of their technology and the chief engineer of the program. The former was more concerned with technical questions (e.g., What is the most efficient approach? Will we have a technologically superior product?), while the latter had more practical concerns (e.g., What will the airlines think of the new technology? How will its initial costs compare with the reduced maintenance costs expected over the plane's lifetime? What will be the program's cost and schedule?).
A number of the "new" technologies considered for the 7X7 had, in fact, already been employed elsewhere, primarily on space vehicles. They were therefore regarded as proven, with few technological risks. For example, digital avionics prototype systems in the cockpit, which replaced the traditional analog systems, had originally been developed for the SST program in 1969. Because it offered improved reliability, more accurate flight paths, lower maintenance costs, and the potential for a two-person cockpit, it was incorporated into the 7X7 with little debate.
Decisions involving unproven technologies were considerably more difficult. As Everette Webb, the 7X7's chief engineer, pointed out: "In such cases, deciding what is an acceptable risk is largely a judgment call." Composites provide an example of Boeing's approach.
Composites are complex materials, formed by combining two or more complementary substances. They appeal to airframe manufacturers because they combine great strength with light weight. In the 1960s and 1970s, Boeing engineers conducted a number of laboratory tests on large, composite panels; eventually, they found a promising material, a mixture of graphite and kevlar. Laboratory tests, however, were not regarded as representative of the "real-world airline environment." To gather such data, Boeing worked with a small number of airlines and conducted limited, inservice tests. Boeing fabricated structural parts, such as wing control surfaces or spoiler panels, using composites; had them installed on a plane then in production; and monitored the material's performance as the plane underwent normal airline use. These tests soon indicated a problem with water absorption in environments of high heat and humidity, such as Brazil. A layer of fiberglass was added to the composite panels to solve the problem, and tests continued through the early 1970s. Yet, despite the tests, engineers decided against using composites for the 7X7's primary structure, and recommended instead that they be used only for secondary parts, where the safety risks were lower. Norton explained: "We push technology very hard, but we're conservative about implementation."
Audit teams were also active during the program definition phase, starting in September 1976. Teams were staffed by experienced Boeing managers, and were assigned to review every significant element of the 7X7 program, including technology, finance, manufacturing, and management. Teams acted as "devil's advocates," and a typical audit took three months. According to Standal:
In the past, we occasionally used outside consultants as auditors. But we found that, for the most part, we do a better job with our own people. We isolate them organizationally and give them a separate reporting line straight to T. Wilson.
In September 1977, the 7X7 program was renamed the 767, and in January 1978, the cost definition phase began (see Exhibit 10-5). This shift was a major step: It indicated escalating program commitment and required the authorization of the president of the Boeing Commercial Airplane Company. Approximately $100 million had already been spent on the 7X7; most of it, however, was regarded as part of ongoing research and development. Now the critical decision was at hand: Would Boeing commit to building a new plane and, in the process, incur up-front costs of several billion dollars?
Only the board of directors could make such a decision. First, however, detailed cost estimates were necessary; they, in turn, had to be based on a single configuration. Cost definition forced engineers and marketing managers to stand up and say, "We want to offer this airplane." The 767's basic design, including the long-delayed choice between two and three engines, was finally frozen in place in May 1978 (see Exhibit 10-6).
Once the basic design was established, costs could be estimated using a parametric estimating technique. This method, adapted by Boeing, had been developed by the New Airplane Program study group from comparisons of the 707, 727, 737, and 747. It predicted the costs of a new plane from design characteristics, such as weight, speed, and length, and historical relationships, such as the number of parts per airplane, that were known well in advance of production.
The critical calculation involved assembly labor hours. Managers began with data from a benchmark (and profitable) program, the 727, and noted, for every major section of the plane, the number of labor hours per pound required to build the first unit. That number was then multiplied by the expected weight of the same section of the 767; this result, in turn, was multiplied by a factor that reflected Boeing's historical experience in improving the relationship between labor hours and weight as it moved to the next-generation airplane. Totaling the results for all plane sections provided an estimate of the labor hours required to build the first 767. A learning curve was then applied to estimate the number of labor hours required to build subsequent planes.
Engineers believed that the historical relationships underlying these calculations remained valid for long periods. According to Dennis Wilson, manager of scheduling for the 767:
Unless we drastically change the way we do business, we will be able to use the same parametrics to compare programs. After all, an airplane is an airplane.
Parametric estimates were, however, carefully fine-tuned to account for differences in plane programs. Adjustments could go in either direction. Improved equipment and management control systems, as enforced reduction in engineering change orders, and heavy use of Computer Aided Design and Computer Aided Manufacturing (CAD/CAM) suggested that the 767 would require fewer hours than predicted by parametrics derived from the 727; increased product complexity and a larger variety of customers suggested that more hours would be required. These factors were combined to form a final, adjusted estimate of total assembly hours.
A similar process was used to develop the Master Phasing Plan, which established the program schedule and identified major milestones (see Exhibit 10-7). The
Exhibit 10-6. Airline Configuration
Exhibit 10-6. Airline Configuration
critical task was linking the schedules of interdependent groups, such as engineering and production, to avoid schedule compression or delays. Parametrics were used for that purpose. For example, comparisons of the 727 and 747 programs suggested that, if problems were to be avoided, fabrication should not begin until 25 percent of structural engineering drawings were complete, and that major assembly should not begin until 90 percent of engineering drawings were complete. Such values became the baseline for the 767's Master Phasing Plan. The initial plan was completed in October 1977, and was revised repeatedly as more up-to-date information became available.
The Go/No -Go Decision
In February 1978, Boeing's board of directors was asked to commit to the 767. Prior to that time, Wilson and the 767 team had briefed them, reviewing all aspects of the program. The board agreed to authorize the new plane, but only if two conditions were met: commitments to purchase were received from one foreign and two domestic airlines, and preproduction orders totaled at least 100 planes.
On July 14, 1978, United Airlines placed a $1 billion order for thirty 767s, making it Boeing's first customer. Being the first customer had certain tasks—the offer to sell was conditional, and could be canceled at a later date—but offered advantages as well. Prices were lower, and the first buyer had an opportunity to help shape the plane's final configuration. By November 1978, American and Delta Airlines had also placed orders, bringing the total to eighty planes, with an additional seventy-nine on option. The board then committed Boeing to full production of the 767. The cost definition phase had ended in July 1978; meanwhile, teams began to flesh out the details of supplier and production management.
A complete 767 consisted of 3.1 million parts, which were supplied by 1,300 vendors. Of these, the most important were the two program participants and four major subcontractors, who built such critical parts as body structures, tail sections, and landing gear. Program participants were, in effect, risk-sharing partners who bore a portion of the costs of design, development, and tooling; major subcontractors were similar, but took on a smaller share of the work. Both were necessary because new airplane programs had become too big for Boeing, or any other single company, to handle alone. On the 767, Aeritalia, the Italian aircraft manufacturer, and the Japan Aircraft Development Company (JADC), a consortium made up of Mitsubishi, Kawasaki, and Fuji Industries, were the two program participants. Both were contracted with in September 1978.
In the late 1960s and 1970s, Aeritalia had worked with Boeing on several proposed airplane designs, including one plane with short-field takeoff and landing capacity. Based on that experience, Aeritalia asked to participate in future work with Boeing. Cerf recalled:
Boeing honored Aeritalia's request. We decided that they would produce the 767's wing control surfaces and tail, parts which were considered to be significant but which were less critical than body panels to the final assembly line. As it turned out, materials technology advanced in the meantime, and most of the control surface parts were changed from aluminum structure to graphite composites. That helped to make them one of the more complex jobs on the airplane.
JADC, on the other hand, was responsible for the several large body sections. The Japanese participants had been interested in working with Boeing for years and had done progressively more important work on other aircraft. Now, their workmanship was considered exacting enough to meet Boeing standards for the production of major sections of structure.
Boeing worked closely with all of its subcontractors, from initial planning to final delivery. Cerf observed:
Generally, at Boeing we do not contract with suppliers and then walk away. We feel responsible for them and have to make it work. This was especially true of the 767 program participants. Because the content of their work was so significant, a failure would have precluded our ability to salvage an industrial operation of this size.
To begin, the Italian and Japanese participants were asked to work together with Boeing engineers. Engineering management helped to select the Italian and Japanese engineers who would participate in the 767 program, and rated them according to their skill levels. The Italian and Japanese engineers then worked alongside Boeing engineers in Seattle. At the 25 percent structures release point (a critical milestone, at which point stress analyses had been completed), they returned to their home companies, accompanied by their Boeing engineering counterparts, who were then integrated into the Italian and Japanese engineering organizations. At the same time, in mid-1978, Boeing established residence teams in Italy and Japan, consisting of some of Boeing's best operations people. The operations teams evaluated and helped to establish participants' facilities, training, and manufacturing processes, and also certified their quality assurance processes. If problems arose, rapid communication with Seattle was often necessary; this was assured by a private telephone network connecting Boeing to each participant.
An Example of Supplier Management: The Japanese Transportation Plan
Initially, JADC had argued that transporting body sections from its factories in Japan to Boeing's assembly plant near Seattle would present few problems. Boeing, to be absolutely certain, had insisted that scale models of all sections be built and carried along the proposed route. The parts proved to be too large for Japan's narrow, rural roads; as a result, an old steel factory, located closer to shipping facilities, was converted by one Japanese company to assemble major sections. Another company constructed a final assembly plant located directly on the water. As insurance, Boeing also requested that the body sections be air transportable, and their designs were sized accordingly.
Boeing then put one of its transportation specialists to work with his Japanese counterparts to develop a transportation plan. This effort took several months, as Cerf recalled:
We went through a major exercise to prove that all of the Japanese companies could support our assembly schedule in Seattle. We brought their representatives to see the complete plan, which covered the walls of a huge meeting room, and worked with them carefully to plan what would be on their shipping docks, what would be on the high seas, and what would be in our plants at any one time.
The level of detail was quite astounding. We kept asking them representative questions, such as ''Do you have the right permits and who will get them? What does the transportation container look like and has it been stressed properly for transport by sea?" Surprisingly, the Japanese didn't object to this process at all. They weren't just cooperative; they were used to working at this level of detail and wanted to learn all we knew.
All of this was a good thing because there was no backup once the decision was made to build the major body sections in Japan. We were committed because our plants at Boeing were working at capacity.
Part fabrication began in July 1979, minor (subsection) assembly in April 1980, and major assembly in July 1980. Such long lead times were necessary to meet the planned rollout of the first 767 in August 1981. Flight tests began immediately after rollout, and FAA certification was expected in July 1982.
All 767s were assembled in Everett, Washington, in the same facility used for 747s. Half of the building was devoted to assembly of major subsections; the other half to final assembly. In the final stages of assembly, a line flow process was used, with seven major work stations (see Exhibit 10-8 for a rough sequence of manufacturing operations). Every four days, partially completed planes were moved, using large overhead cranes, from one work station to the next. At each work station, teams of skilled employees positioned a single plane in massive tools and fixtures, and then riveted, wired, and connected parts and pieces.
During the assembly stage, managers faced two critical tasks: maintaining schedule, and ensuring that learning curve goals were met. Both were complicated by a key difference between airframe manufacturing and other industries: the difficulty of managing a large number of engineering change orders. Haas observed:
An airplane is not something you design, turn over to manufacturing, and then forget. The configuration is constantly changing. So you commit to a schedule, and then incorporate changes and improvements as they come.
This task was especially critical because cost estimates assumed that assembly labor hours would decline predictably over time, following a preset learning curve. Managers therefore had to ensure that learning goals were met at the same time that they were accommodating unanticipated changes.
Requests for changes came from internal and external sources. Some, such as the color of carpeting or seating arrangements, were negotiated by airline customers; others, such as parts or wiring changes, were proposed by engineers. In total, the two sources generated 12,000 changes on the first 767.
Managers tracked these changes carefully. Even before the plane's basic design was frozen, all major changes had to be filed using the same formal procedure. This was done to ensure that specifications remained accurate. Once assembly began, a Production Change Board, chaired by the operations department, reviewed all engineering change requests and assessed their likely impact on schedule and cost. If the changes were approved, an implementation plan was then developed. Three general approaches were used: incorporating changes into the normal flow of production; installing old parts as originally planned and then retrofitting new parts off-line, outside the normal flow of production; and expediting changes by assigning additional workers, a process known as "blue streak."
Exhibit 1Q-S. ?ó?-2QQ Manufacturing Sequence
Exhibit 1Q-S. ?ó?-2QQ Manufacturing Sequence
ukc km fwt ij. nun tuaimML-
TM WCt tHHI KlfDtl W
ukc km fwt ill Ml i,ti nun ij. nun tuaimML-
UJtlCWIMt m cuto toot ira nan Q—tA
TM WCt tHHI KlfDtl W
In all cases, a primary concern was maintaining schedule. Boeing faced substantial penalties if a plane was delivered even one day late, because airlines planned their schedules around promised delivery dates and expected a new plane to be flying immediately. According to Haas:
For a long time, we have stressed the importance of schedule performance. The airplane will move [from one work station to the next] on the day that it is supposed to move. Management will get in a lot more trouble for not moving an airplane, assembly, or part on schedule than for a budget overrun. Over the years, budgets have gained significantly in importance, but not at the expense of schedules.
To ensure that schedules were maintained, Boeing employed a management visbility system. Schedules were prominently posted, and marathon status meetings, which were attended by representatives of all affected departments, were held weekly to review slippages and highlight potential problems. Every manager discussed what he or she was doing and what he or she was owed by others. The emphasis was on early notification, as Dennis Wilson observed:
If I'm at a status meeting and I find that someone has missed a critical milestone, the first question I ask is, "Why didn't you tell me about the problem last week?," not, "Why did you miss the milestone?"
In June 1981, as assembly of the first 767 moved into its final stages, a First Flight Committee was established. The committee reported directly to Dean Thornton and met daily during the six weeks before the plane's first test flight. At that point, the test pilot had final say in setting priorities and selecting the tasks to be completed.
Learning curves were also used to manage the assembly process. Based on historical experience, Boeing had developed learning curves for every major work center. Machining, assembly, and sheet metal fabrication had curves of their own, each with a different slope. However, curves were used in the same way at all centers.
To begin, an optimum crew size was defined for the operation, based on available work space, engineering guidelines, and tooling to be employed. For example, the optimal size for forward body section assembly was eight people. A parametric estimate was then made of the number of labor hours needed to assemble that section of the very first 767. The total (in this case, 6,000 hours) was then divided by the number of labor hours available each day (in this case, 128 hours, equal to eight people working eight hours per shift, two shifts per day) to give the number of days to complete the very first assembly (forty-seven days).
At this point, a learning curve was invoked. The next assembly would be scheduled not for forty-seven days but for a lesser number, to reflect the historical rate of learning on that operation. The same number of people would be employed, but they would work faster and more efficiently. (When precise calculations were impossible, Boeing varied staffing levels within minimum and maximum values, rather than sticking to a single, optimal crew size.)
Learning curves were also applied to change management. Work centers were initially staffed to reflect a large number of changes. For example, of the eight people assigned to forward body section assembly, three might initially be responsible for in corporating changes. But because the number of changes fell sharply as more planes were produced—the first 767 had 12,000 changes, while the seventieth 767 had only 500—fewer people would be needed for the activity as time passed, and staffing would be reduced over time.
Such improvements did not come automatically. Three tools were used to ensure that targets were met: specific work station goals; stand-up meetings with first-line supervisors; and the management visibility system discussed earlier. Hourly goals were set for every employee and displayed prominently on bar charts by their work stations. The game, as one manager put it, then became "worker versus bar chart." Stand-up meetings were held only if targets were not met. First-line supervisors had to stand up at these meetings and identify what was impeding their ability to meet learning curve goals. Managers were then responsible for solving the problems.
In the late 1970s, airframe manufacturers, led by Boeing, proposed a switch from three- to two-person cockpits. Advanced technology, they argued, had made a three-person crew unnecessary. The Air Line Pilots Association (ALPA) objected strongly to these arguments, claiming that safety levels were certain to fall if the number of crew members was reduced. To resolve the debate, a presidential task force was convened; both parties agreed to accept its findings. In July 1981, the task force concluded that two-person cockpits presented no unusual safety problems, and that manufacturers could offer them on all planes.
Airlines, including those that had already ordered 767s, soon expressed an interest in having their planes delivered with two-person cockpits. Boeing had anticipated such a response and, years earlier, had conducted preliminary studies to determine how best to convert the 767 from its original, three-person cockpit design to a two-person model (see Exhibit 10-9 for a comparison of the two cockpits). Further studies were immediately begun; their goal was to identify the number of planes then in process that would require rework or modification to become two-crew models, and the likely impact of these changes on cost and schedule. Engineers concluded that the thirty-first 767 was still far enough from completion that it, and all subsequent planes, could be built with two-person cockpits without modification. Thirty planes, however, were in relatively advanced stages of production. Some were nearly ready to be rolled out and flown; others had complete cockpits but were not yet tested; others had bare cockpits without any electronics installed. But since all thirty were being built according to the plane's original, three-person cockpit design, all would require some modification.
Customers were notified of the additional cost and delivery delay they could expect on these thirty planes. The impact was not large: a small percentage increase in costs and on average delay of one month from promised delivery dates. All but one airline chose to have their planes built with two-person cockpits.
In August 1981, a special task force, reporting directly to Thornton, was formed to determine the best way of modifying these planes. It soon narrowed the choice to two alternatives: (1) building the thirty airplanes as they had originally been designed, with three-person cockpits, and then converting them to two-person cockpits after they had left the production floor (but before delivery to customers), and (2) modifying the production plans for the thirty airplanes so that conversion would take place during production and no parts would be installed only to be removed later (which
Exhibit 10-9. Three-Crew and Two-Crew Cockpit Designs
Exhibit 10-9. Three-Crew and Two-Crew Cockpit Designs
meant leaving some cockpits temporarily unfinished while drawings and parts for two-person cockpits were being developed).
Completion of Production and Subsequent Modification
In this approach, production would continue as planned, without delay. Neither learning curves nor schedules would be disrupted by attempts to modify airplanes during the assembly process. The modification program would be managed as a separate, tightly controlled activity, apart from the normal flow of production, and special teams of "modification experts," skilled at parts removal, modification, and repair, would be assigned to it. Approximately one million additional labor hours were thought to be required if this method were used.
The primary advantage of this approach was that flaps, ailerons, landing gear, hydraulics, and other airplane systems would be functionally tested during the final assembly process, as originally planned. Problems would be identified and corrected on the spot, rather than hidden or disguised by subsequent assembly activities. And because the airplane that rolled out of production would be fully tested and functional, any problems identified after installation of the two-person cockpit could be isolated, with some assurance, to the cockpit area.
The risk of this approach was the potential "loss of configuration" (i.e., when the plane was actually built, the integrity of the overall design might be compromised). Parts required for three-person cockpits would be installed firmly in place, only to be removed and replaced later by modification experts. (Because these parts had been ordered months before and were already on hand and paid for, this option did not impose greater scrap costs than the other option.) If the modification was not done carefully, many of the plane's operating systems might be disrupted. Boeing experts, however, believed that the management controls used for modification would prevent this from occurring. To minimize the risk, additional functional testing would be required after modification.
Space was also a problem. There was not enough room within the factory to modify all thirty planes. Work would therefore have to be done outside, but even then space was limited. A special parking plan would have to be developed, and the planes being modified would have to be parked extremely close together. The required arrangement would violate fire regulations, so special fire control plans and waivers would be necessary.
Several managers had reservations about this approach, for they objected to its underlying philosophy. The end result would be an airplane that had been modified, after the fact, to accommodate a two-person cockpit. As Standal put it: "It goes against our grain and better judgment to roll out an aircraft and then tear the guts out."
In this approach, all modification of the thirty planes would be done during production, rather than after the fact. No parts would be installed only to be removed later. Instead, all panels, instruments, and switches that were associated with three-person cockpits would be identified and their installation halted. Meanwhile, production would continue on other sections of the plane. Once plans and parts were available for two-person cockpits, they would be incorporated within the flow of production.
This was the traditional method of making engineering and design changes. It was used routinely for the thousands of configuration changes on every new airplane. The primary advantage of this approach was that all parts were installed only once.
Because there would be no installation and subsequent removal, the configuration was more likely to remain secure. Moreover, because modification would occur during production, all activities would be controlled by normal management procedures, rather than by a separate program.
The primary disadvantage of this approach was that the original production plan would be disrupted. Separate plans would have to be developed for the first thirty airplanes, which required modification, and all subsequent planes. Learning curves would be disrupted as well, because a large number of additional workers would have to be added temporarily, at selected work stations, to complete the modification of the first thirty planes. If this method were used, modification was expected to require approximately two million additional labor hours.
Because all cockpit work would be deferred until engineering drawings and parts were available for two-crew models, test procedures would also have to change. Traditionally, functional testing was done sequentially, with each system (flaps, ailerons, etc.) tested as it became operational. That approach would be impossible here because all cockpit work would be deferred until complete plans and drawings were available. Functional testing would therefore have to be done after the two-person cockpit was fully installed. Problems might not be detected and corrected immediately and might well be hidden by systems that were installed later, making problem diagnosis much more difficult.
Thornton knew that it was time to make a choice between the two approaches so that production could continue. The risks, however, were great; as his staff kept telling him, the decision was a potential "show-stopper." He wondered: "Should I authorize after-the-fact conversion of planes or modification during production? And for what reasons?"
The Boeing 767:
From Concept to Production (B)10
Thornton elected to retrofit the thirty 767s with two-person cockpits, rather than installing them inline. The project was managed as a separate production program with its own schedule and learning curves, and the storage problem was solved by special parking arrangements in the large stalls normally used for 747s. Managers were greatly pleased with the results. In August 1981, the first 767 was rolled out as planned, and only a few deliveries were delayed by as much as a month.
Six years later, the world of airframe manufacturing had changed. By August 1987, Boeing had received orders for 263 767s; of these, 181 had been delivered. But the monthly production rate was down from the planned level of eight to two, and forecasts had not been fully met. One reason was that most U.S. airlines had developed different needs because of deregulation. Routes that had previously been limited to a small number of carriers had been opened up. The resulting competition had depressed ticket prices and profits and led ultimately to both bankruptcies and mergers.
10 Copyright © 1988 by the President and Fellows of Harvard College. Harvard Business School case 688 -041. This case was prepared by Lee Field, Janet Simpson, and David A. Garvin as the basis for class discussion rather than to illustrate either effective or ineffective handling of an administrative situation. Reprinted by permission of the Harvard Business School.
At the same time, airlines had altered their route structures, moving from the traditional pattern of direct connections between city pairs to hub and spoke systems that encouraged shorter flights.
Airbus had also proved to be a formidable competitor. By 1987, it had captured 20 percent of the world market and 30 percent of the market outside the United States. Airbus continued to receive support from the French, British, Spanish, and German governments, and was actively seeking new members for its consortium.
In this environment, Boeing was preparing to launch its newest plane program, the 7J7, an advanced technology airplane targeted for the 150-seat market. The plane was based on a radically new engine, the prop-fan, that promised fuel savings of up to 50 percent over turbofan engines, as well as a new airframe and new optical fiber and hydraulics technologies. McDonnell Douglas, by contrast, was also planning a new plane, but it had decided to combine prop-fan engines with an existing airframe. Both companies were hoping that their planes, scheduled to appear in 1992, would take sales from Airbus' A-320, a 150-seat model, due in 1988, that was without the new engine technology.
To Boeing the challenge was clear. To be successful, the 7J7 had to offer an operating cost advantage over the A-320 while providing the airlines with enhanced revenue-generating capability. The question was, how? Technology was one route. But should the 7J7 program also be managed differently than its predecessor, the 767?
Was this article helpful?
What you need to know about… Project Management Made Easy! Project management consists of more than just a large building project and can encompass small projects as well. No matter what the size of your project, you need to have some sort of project management. How you manage your project has everything to do with its outcome.