What is the current state of manufacturing at Whistler? Highlight some symptoms from the case that...

Transcribed Image Text:

What is the current state of manufacturing at Whistler? Highlight some symptoms from the case that may indicate problems with their manufacturing performance. 2. What are the underlying causes of Whistler's problems? Identify both external and internal causes. 3. Based on your analysis above, which one of the three optionsshould Whistler consider to solve its manufacturing crisis? What are the risks and benefits of that one manufacturing option for Whistler? 4. Can you explain some of the potential ethical challenges associated with either one of the three options? Can you come up with a fourth option that can possibly address some of the ethical issues you raised while not sacrificing the benefits of the firm, employees as well as the end consumer? Whistler Corporation (A)¹ In the summer of 1987, Charles Stott, the recently appointed President of the Whistler Corporation, realized that the company had reached a critical juncture-the once profitable maker of radar detectors was losing $500,000 per month. Stott had been brought in by Whistler's corporate parent, the Dynatech Corporation, to return the company to profitability. Within the next few weeks, Stott was to present a decision to Dynatech regarding the future of Whistler's manufacturing operations. Stott was considering radically restructuring Whistler's domestic manufacturing operations in an attempt to become cost competitive with off-shore manufacturers. A pilot "just-in-time" (JIT) synchronized production line that had been in operation for less than three months would be the model for the new manufacturing system. If the synchronized production system could be successfully scaled up. Whistler could expect significant cost savings from reduced work-in-process inventory, better quality, higher labor productivity, and more efficient utilization of floor space. Major savings in fixed costs would also come from being able to close the company's plant in Fitchburg, Massachusetts. A second option being contemplated was instead to expand the company's long standing and successful relationship with a Korean consumer electronics company. This company was already supplying complete "low-end"2 radar detectors to Whistler at very attractive prices. Turning over additional low-end products to this Korean supplier would allow Whistler to be more cost competitive in this segment of the market. It would also alleviate some of the capacity problems the company was experiencing in its two domestic plants (in Westford and Fitchburg, Massachusetts). A third, and more extreme option was also being given serious consideration: Move all production off- shore and shut down the company's two domestic plants. This route had been taken by all but one of Whistler's competitors in the radar-detection market. Background Radar detectors, known in the vernacular as "Fuzz-busters"TM3, are small electronic devices which alert drivers to the presence of police band radar used to track vehicle speeds (Exhibit 1). The device contains three basic functional (internal) parts. The microwave assembly is an antenna which picks up microwave signals (emitted by police "radar guns") and converts them into lower frequency radio signals. These signals are then processed and interpreted by the radio frequency assembly. If police band radar signals are detected, the control assembly alerts the driver through flashing lights, a beep, or some combination of visual and audio cues. The first radar detectors were introduced in 1972. Truck drivers were then the overwhelmingly dominant users of radar detectors. In the late 1970's, the use of radar detectors began to spread to automobile "enthusiasts", salespeople, and others who frequently drove long distances on highways. It was still a small, specialized niche market when the Whistler Corporation decided to enter the business in 1978. Whistler was founded by Dodge Morgan in the early 1970's. In its earliest days, the company designed and manufactured electronic specialty products (voice scramblers, marine radars, and gas leak detectors) in the founder's garage. As revenues began to grow, the company relocated first to an old mill and then to a larger, more modern facility in Westford, Massachusetts. Because of the company's strong design and engineering capabilities, it quickly became a dominant and profitable player in the small, but growing market for radar detectors. Its first radar detector, designed in 1978, became a leading seller. In 1982, Whistler introduced two more models. In 1983, Whistler was one of six companies competing in what was still a rather specialized, but profitable, niche market. In this same year, Whistler was acquired by Dynatech, a Burlington, Massachusetts company whose strategy was to purchase small to medium sized companies with dominant positions in niche markets. Explosive Growth: 1983-1987 Shortly after Dynatech acquired Whistler, a number of interrelated changes occurred in the market for radar detection devices. As indicated in Exhibit 2, unit aggregate demand exploded. Between 1982 and 1987, the total number of radar detectors sold in the United States increased by more than 450%. Annual rates of market growth during this period averaged 35.6%. This growth was associated with a fundamental change in the composition of demand. "Mass" consumers became the dominant buyer segment. The market became segmented according to price, quality, performance, and purchasing convenience. Distribution expanded from specialty auto and truck shops to a variety of general retail outlets (electronics shops, mass merchandisers, and mail-order catalogues). To serve as many segments of the rapidly growing consumer market as possible, Whistler introduced nine new models between 1982-1987. As could be expected, demand growth attracted new competition. By 1987, Whistler had 19 competitors. While many of these were American companies, virtually all of them sold radar detectors which were manufactured under subcontracting arrangements with Asian suppliers. The majority sourced exclusively from Asia. In the wake of intense competition and access to low-cost off-shore producers, the average price of radar detectors declined steadily (Exhibit 2). Manufacturing Operations Manufacturing at Whistler was divided into two sets of operations: subassembly production and final assembly. Before 1985, both subassembly and final assembly operations were done in Whistler's 40,000 square foot plant in Westford. In 1985, Whistler moved final assembly operations to a new 20,000 square foot plant in Fitchburg, Massachusetts (approximately 25 miles from its Westford plant). Subassembly As indicated earlier, radar detectors consist of three internal major subassemblies: 1) a microwave subassembly, 2) a radio frequency (RF) subassembly, and 3) a control subassembly. Whistler designed and built all three of these subassemblies in-house. Both the RF and control subassemblies consisted of printed circuit boards containing through-hole soldered, as well as surface mounted, electronic components. The RF and control boards were assembled from electronic components and bare printed circuit boards purchased from outside vendors. The microwave subassembly was manufactured using zinc die castings supplied by an outside vendor. Normally, these subassemblies were produced in batch sizes large enough to meet the requirements of one month of final assembly. Circuit board production required a number of steps. First, all of the components required to produce a batch of a particular circuit board were brought from the stockroom to the circuit board production area. Then, electronic components were automatically mounted onto the surface of the bare panels. After this "surface mounting", the batch was inspected to ensure that all of the appropriate components were correctly placed. The entire batch then moved to the "hand stuffing" work area where components that could not be automatically mounted were manually inserted into the board. These "hand-stuffed" components were then soldered into place on an automated wave solderer. The entire batch of boards was then tested; boards which failed the test were marked. All panels were then sent to the "waterknife" which used a high-pressure stream of water to cut the panel into individual circuit boards, a process known as "depalletization". After depalletization, any defective boards were sent to the printed circuit board re-work area. The boards which had passed the first step were sent to a another work station for RF tuning and functional testing. Boards which failed this second test were also sent to the re-work area. Finally, the entire batch went through a final quality control audit to ensure that only good boards would be sent to the board inventory in the stockroom¹ and ultimately on to final assembly. Microwave assemblies were assembled in a separate part of the factory. Like the RF and control boards, the microwave assemblies were produced in batches large enough to meet one month's final assembly requirements. Final Assembly Final assembly consisted of six steps. First, the three major internal subassemblies were wired together. Second, a "quick test" of the integrated electronic system was done. Defective units were sent to a final assembly rework area. Third, the now-integrated electronic systems were attached to the bottom half of the unit's plastic molded shell. Fourth, the top half of the exterior shell was fastened onto the bottom half (now containing the electronic system). Fifth, the entire unit was tested (defective units were sent to the re-work area). Finally, the unit was packaged with instructions and sent to the finished goods storage area. Production Control The flow of materials was controlled by what is commonly referred to as a "batch-and-kit" method. In a batch-and-kit operation, all subassemblies and final assemblies are produced in batches-at Whistler, normally monthly batches. For example, if 5000 Spectrum 1 radar detectors were scheduled for production in a month, all 5000 would be assembled in one batch. Before the final assembly of a batch could begin, all requisite subassemblies and parts had to be ready. Several weeks before final assembly of a batch of Spectrum Is was scheduled to begin, the subassembly area produced 5000 of the appropriate RF boards, 5000 of the appropriate control boards, and 5000 of the appropriate microwave assemblies. The finished batches were then sent to the stockroom. Just prior to the time at which final assembly was scheduled, all the required subassemblies were picked from the stockroom and organized into a "kit"-all those parts required for the final assembly of one batch of finished product of that model. The kit was then sent to final assembly. Most of the small electronic components used by Whistler were supplied by vendors located in Asia. For these components, the lead time from order to delivery was about ten weeks. Due to these long lead times, and because of the relatively low cost of these components, Whistler generally stocked enough raw materials to satisfy the next two months of scheduled production. Strains in Manufacturing Strains in Manufacturing As production volume began to increase rapidly after 1983, manufacturing operations experienced a number of problems. When Whistler had been producing four models in relatively low volumes, the number and size of production batches were manageable. However, increases in both total production volume and the number of models meant an increase in the number and size of batches. With a total of 13 models in production, the factory often found itself with several batches of in-process units lined up and waiting between work areas. With available stockroom space full, the floor of the Westford factory quickly became cluttered with work-in-process. Floor space was not the only constraint in the Westford plant. There were many electronics companies in the Westford-Lowell area. Because of defense spending and a general expansion in the economy, these companies were experiencing a boom in business. Demand for semiskilled assembly workers was extremely high. Whistler, competing for labor with such companies as Digital Equipment Corporation (DEC) and Wang, found it difficult to staff its production operations at the wages it could pay. A decision was made to lease a factory in Fitchburg, Massachusetts where both labor and space were more economically available. Management viewed the Fitchburg plant as a way to Whistler Corporation (A) 690-011 quickly add capacity and capture some of the business Whistler had been losing. In 1985, the Fitchburg plant was opened. All final assembly and test operations were moved to Fitchburg. Subassembly operations and packaging remained in Westford. Thus, the process flow was altered as follows: RF boards, control boards, and microwave subassemblies were produced and tested in Westford. These subassemblies and all the other parts required to make a batch of a particular model were "kitted" and then trucked to Fitchburg. In Fitchburg, the kits were assembled into final products and tested. The entire batch (or multiple batches) were then trucked back to Westford where finished product was packaged and stored. During this time period, Whistler began to experience quality problems in the subassembly of RF and control boards. Some of these problems were due to defects in components sourced from outside vendors. However, a major part of the problem was caused by the new surface mount technology (SMT). Components which are surface mounted take up less space on the circuit board than those that must be inserted and soldered into holes. As a result, surface mounting allows more components to be packed onto a smaller circuit board. The trend toward smaller radar detectors made this technology desirable. When automated, the surface mounting process reduced cycle time dramatically. Given the need to increase production volume to keep up with demand as well as the increased demand for smaller detectors, SMT was a natural technology to adopt, even though it was a far more complex process than traditional methods. The first SMT equipment was installed in 1986. The yields on the SMT process were quite low while operators, designers, and process engineers learned about and adjusted to the subtleties of the new process. By early 1987, the first-pass yield (the percentage of "good" output the first time through, before rework) had climbed to 75%. Through diligent inspection and extensive reworks, the quality of products reaching customers remained extremely high. However, ensuring that reliability in the field was quite costly. For example, at the end of 1986, 100 of the company's 250 production workers were deployed to fix defective boards. About 30% of the Westford plant's floor space was taken up with in-process rework. Rework accounted for approximately $600,000 of the company's $2 million of work-in-process inventories. Defective boards also hampered smooth material flow. A strict batch-oriented material flow discipline requires that an adequate number of subassemblies of each type be available to complete a kit. High rates of defective subassemblies created problems in matching subassemblies to create final assembly kits. In final assembly, incomplete kits could sit for weeks waiting for defective subassemblies to be reworked. The pressure to ship products was so great that common parts from other batches in process would sometimes be used as replacements. Unfortunately, this "borrowing" of components often went unrecorded. As a result, the missing parts were not available when the other batch was ready for final assembly. High defect rates became so normal that they were built into the production schedule. For example, as a matter of policy, the Production and Control Department ordered 20% more subassemblies than the final assembly production schedule called for. Work-in-process piled up because of frequent unexpected schedule changes. When Whistler had been producing only four models for a rather narrow market, the schedule could be set well in advance. The "mass" market, however, was far more volatile. As one production scheduler put it: Out of the blue we could get an order for 5000 Spectrum 2's from a big retailer running a weekend promotion. If we had the products to ship, we had the order. If not, we lost it. This was something we just weren't used to. Truckers don't go out and buy a radar detector just because it's George Washington's birthday. production controller who had worked on the production line described the scene: "The place was a total zoo. Boards and half-assembled units were piled up from floor to ceiling. We even had to rent three trailers to handle the overflow." Work-in-process had also become a materials-handling nightmare. The longer kits remained on the factory floor and the more they were reshuffled (to make room for other work-in-process), the greater the likelihood of damage to delicate electronic parts and further problems in final assembly. In 1987, by the time a unit had completed final assembly, it had spent an average of 23 days in process, although actual production time was only eight hours. Performance Until late 1986, manufacturing was of little concern to the Whistler management. Innovative design and good marketing were, as one executive put it, "the name of the game". Sales had been growing rapidly and manufacturing, despite its internal problems, had managed to keep up. In 1985, profits before taxes were 20% of sales; the company had a pretax return on assets of 40%. In that year, the company was the market share leader with 21% of the domestic radar detector market (units). The "build as many as you can, any way you can" strategy seemed to be working. In 1986, however, financial performance deteriorated rapidly. High manufacturing costs were making it difficult to compete with off-shore manufacturers. The manufacturing costs of Asian subcontractors were substantially lower than those of Whistler (Exhibit 3). A marketing analysis suggested that, due to their performance, quality, reputation, and brand image, Whistler's products could support a 10% price premium, but not much more. By the end of 1986, Whistler's market share had fallen to 12%. By the end of the summer of 1986, the company began to lose money for the first time in its history, at the rate of almost $500,000 per month. The problems in manufacturing began to draw notice. According to Jack Turner, Vice President of design and engineering, who headed manufacturing at the time: "We knew manufacturing was sick. We knew something had to be done. We just didn't know what it was." In September 1986, a consulting firm was hired to study Whistler's manufacturing process and to make suggestions for change. The RACE-ME Program The consulting firm suggested a comprehensive program to reform manufacturing operations. The program was dubbed RACE-ME (Restoring A Competitive Edge Through Manufacturing Excellence). Its goal was to make Whistler's manufacturing as efficient as its Far Eastern competitors within 24 months. The RACE-ME program included reforms in materials handling practices, operator training. process lay-out, inspection and quality control, tooling, and production flow. A model (pilot) production line (MPL) was set up in a corner of the Westford plant to implement, evaluate, and demonstrate various reforms. The line consisted of a series of connected work benches. For symbolism, the benches were a different style and different color than those found in the main manufacturing operations. The relatively high-volume Spectrum 1 was chosen to be produced on the The MPL was designed to change the flow of materials and work-in-process. Rather than a batch flow, the MPL would operate as a "repetitive" or a synchronous line-flow manufacturing process with only very small buffer stocks between adjacent stages of production. In the traditional system, a work station received large batches of components and subassemblies as scheduled, whether or not it was ready to start working on them. On the MPL, a work station would receive materials or work-in-process only when it requested them. The second major change was in the production schedule. The MPL was scheduled to produce the same quantity of detectors day after day. After some trial-and-error, the MPL process evolved as follows: RF boards, control boards, and microwave subassemblies were produced with the same equipment and through the same process as before. However, separate inspection points were eliminated. Each work station was made responsible for identifying and correcting its own quality problems. Batch sizes and production flow were dramatically different. Rather than produce boards in one-month batches, the SMT operation ran only enough boards to supply one day of final assembly. The batch sizes were only 2 hours in the board assembly operations after SMT (hand stuffing, wave soldering, and depalletization). The production flow was controlled by "kanban"6 racks between each work station. Workers at any particular work station were instructed to work on a particular batch of boards only when an empty tray appeared in the rack between it and the next work station. When a specific work station needed more of a certain type of board, a worker would place an empty tray (with a production control card indicating the type and number of boards required) in the rack. This would signal a worker in the preceding work station to begin making a batch of those boards. The worker receiving the "produce" signal first checked the inventory in the rack between it and the station preceding it. If the appropriate work-in-process or components were there, the worker could begin processing immediately. If not, he or she would order them from the preceding station by putting an empty tray with the appropriate production control card on the rack. Through this kanban chain, materials and work-in-process were pulled as needed by downstream stations. The flow of materials through subassembly was similarly controlled by the final assembly line. When required, final assembly would pull small lots of boards (a 15 to 30 minute supply) from the kanban rack located after the depalletization area. The small lot would then be tested. Boards that failed were sent to a rework area assigned especially for the MPL. A limit had been set on the maximum number of boards permitted in the board rework station. Once the limit was reached, the board subassembly operation would be shut down until the cause of the quality problem could be identified and corrected. It was hoped that this line shut-down procedure would help keep the process under control by drawing attention to defects and by forcing shop management to deal with their causes. The boards were then passed along to the unit integration station where the three major subassemblies were joined. The MPL operated as a worker-paced assembly line. After completing a particular assembly step, the operator would pass the work piece a few feet down the bench to the next station's incoming work tray. These trays could hold a maximum of six work pieces. Once the tray was filled to capacity, the operator at the preceding station was instructed to stop working. He or she could resume assembly only when the tray contained fewer than six in-process work pieces. At the end of the MPL, an enclosed work station performed the final test. Any unit failing this final test was sent to another rework station. The entire production line would be shut down if one day of inventory accumulated in this rework area. MPL Results: April - June 1987 The MPL began producing Spectrum 1's on April 7, 1987. By late June, a study of the pilot line was completed. The data (summarized in Exhibit 4) suggested that it might be possible for Whistler to achieve substantial improvements in the productivity of people, equipment, and space if the manufacturing methods used on the MPL were used to produce all of Whistler's radar detectors. In addition, the MPL enabled Whistler to produce a Spectrum 1 from start to finish in 1.5 days, as compared to the 23 days required in the traditional system. One option was to implement MPL concepts throughout the Westford plant. The expected gains in productivity were expected to allow Whistler to close the Fitchburg plant and reduce its total labor force. However, several questions remained. Some managers in the company wondered whether the results of the pilot project could be replicated on a plant-wide basis. Which product lines might be best suited to the repetitive manufacturing process? Ed Johnson, the Director of Marketing, argued: This is extremely risky. You're talking about changing the entire manufacturing system and maybe shutting down one of our plants. Frankly, the possibilities scare me to death. Any kind of major disruption in output could really hurt us. I know things aren't very good now, but, at least we are getting product out the door. I think we should make changes in manufacturing, but let's phase them in more slowly. Larry Santos, the plant manager at the Fitchburg plant, was also concerned: I know this sounds like I am just looking after my own job, but I have to say I am not sure closing Fitchburg is the best decision. First, the plant has only been open for two years. The people there have been showing some real commitment. In fact, most of the quality and scheduling problems have originated in Westford. The Director of Operations Planning, Sharon Katz was also concerned about closing Fitchburg: I'd have to agree with Larry about closing Fitchburg, but for a different reason. If we plan to grow, we're eventually going to need the extra capacity Fitchburg gives us. It's going to very expensive to start up a new plant in a few years. The Executive Vice President of Finance, Margaret Curry, had a more extreme viewpoint: I am very impressed by the results of the MPL experiment, but I just don't think the cost savings will be enough. I just don't believe we can compete by manufacturing our products domestically. The offshore option buys us tremendous sourcing flexibility. We can shop around for the best suppliers in cost, quality, and delivery. If one isn't working out, What is the current state of manufacturing at Whistler? Highlight some symptoms from the case that may indicate problems with their manufacturing performance. 2. What are the underlying causes of Whistler's problems? Identify both external and internal causes. 3. Based on your analysis above, which one of the three optionsshould Whistler consider to solve its manufacturing crisis? What are the risks and benefits of that one manufacturing option for Whistler? 4. Can you explain some of the potential ethical challenges associated with either one of the three options? Can you come up with a fourth option that can possibly address some of the ethical issues you raised while not sacrificing the benefits of the firm, employees as well as the end consumer? Whistler Corporation (A)¹ In the summer of 1987, Charles Stott, the recently appointed President of the Whistler Corporation, realized that the company had reached a critical juncture-the once profitable maker of radar detectors was losing $500,000 per month. Stott had been brought in by Whistler's corporate parent, the Dynatech Corporation, to return the company to profitability. Within the next few weeks, Stott was to present a decision to Dynatech regarding the future of Whistler's manufacturing operations. Stott was considering radically restructuring Whistler's domestic manufacturing operations in an attempt to become cost competitive with off-shore manufacturers. A pilot "just-in-time" (JIT) synchronized production line that had been in operation for less than three months would be the model for the new manufacturing system. If the synchronized production system could be successfully scaled up. Whistler could expect significant cost savings from reduced work-in-process inventory, better quality, higher labor productivity, and more efficient utilization of floor space. Major savings in fixed costs would also come from being able to close the company's plant in Fitchburg, Massachusetts. A second option being contemplated was instead to expand the company's long standing and successful relationship with a Korean consumer electronics company. This company was already supplying complete "low-end"2 radar detectors to Whistler at very attractive prices. Turning over additional low-end products to this Korean supplier would allow Whistler to be more cost competitive in this segment of the market. It would also alleviate some of the capacity problems the company was experiencing in its two domestic plants (in Westford and Fitchburg, Massachusetts). A third, and more extreme option was also being given serious consideration: Move all production off- shore and shut down the company's two domestic plants. This route had been taken by all but one of Whistler's competitors in the radar-detection market. Background Radar detectors, known in the vernacular as "Fuzz-busters"TM3, are small electronic devices which alert drivers to the presence of police band radar used to track vehicle speeds (Exhibit 1). The device contains three basic functional (internal) parts. The microwave assembly is an antenna which picks up microwave signals (emitted by police "radar guns") and converts them into lower frequency radio signals. These signals are then processed and interpreted by the radio frequency assembly. If police band radar signals are detected, the control assembly alerts the driver through flashing lights, a beep, or some combination of visual and audio cues. The first radar detectors were introduced in 1972. Truck drivers were then the overwhelmingly dominant users of radar detectors. In the late 1970's, the use of radar detectors began to spread to automobile "enthusiasts", salespeople, and others who frequently drove long distances on highways. It was still a small, specialized niche market when the Whistler Corporation decided to enter the business in 1978. Whistler was founded by Dodge Morgan in the early 1970's. In its earliest days, the company designed and manufactured electronic specialty products (voice scramblers, marine radars, and gas leak detectors) in the founder's garage. As revenues began to grow, the company relocated first to an old mill and then to a larger, more modern facility in Westford, Massachusetts. Because of the company's strong design and engineering capabilities, it quickly became a dominant and profitable player in the small, but growing market for radar detectors. Its first radar detector, designed in 1978, became a leading seller. In 1982, Whistler introduced two more models. In 1983, Whistler was one of six companies competing in what was still a rather specialized, but profitable, niche market. In this same year, Whistler was acquired by Dynatech, a Burlington, Massachusetts company whose strategy was to purchase small to medium sized companies with dominant positions in niche markets. Explosive Growth: 1983-1987 Shortly after Dynatech acquired Whistler, a number of interrelated changes occurred in the market for radar detection devices. As indicated in Exhibit 2, unit aggregate demand exploded. Between 1982 and 1987, the total number of radar detectors sold in the United States increased by more than 450%. Annual rates of market growth during this period averaged 35.6%. This growth was associated with a fundamental change in the composition of demand. "Mass" consumers became the dominant buyer segment. The market became segmented according to price, quality, performance, and purchasing convenience. Distribution expanded from specialty auto and truck shops to a variety of general retail outlets (electronics shops, mass merchandisers, and mail-order catalogues). To serve as many segments of the rapidly growing consumer market as possible, Whistler introduced nine new models between 1982-1987. As could be expected, demand growth attracted new competition. By 1987, Whistler had 19 competitors. While many of these were American companies, virtually all of them sold radar detectors which were manufactured under subcontracting arrangements with Asian suppliers. The majority sourced exclusively from Asia. In the wake of intense competition and access to low-cost off-shore producers, the average price of radar detectors declined steadily (Exhibit 2). Manufacturing Operations Manufacturing at Whistler was divided into two sets of operations: subassembly production and final assembly. Before 1985, both subassembly and final assembly operations were done in Whistler's 40,000 square foot plant in Westford. In 1985, Whistler moved final assembly operations to a new 20,000 square foot plant in Fitchburg, Massachusetts (approximately 25 miles from its Westford plant). Subassembly As indicated earlier, radar detectors consist of three internal major subassemblies: 1) a microwave subassembly, 2) a radio frequency (RF) subassembly, and 3) a control subassembly. Whistler designed and built all three of these subassemblies in-house. Both the RF and control subassemblies consisted of printed circuit boards containing through-hole soldered, as well as surface mounted, electronic components. The RF and control boards were assembled from electronic components and bare printed circuit boards purchased from outside vendors. The microwave subassembly was manufactured using zinc die castings supplied by an outside vendor. Normally, these subassemblies were produced in batch sizes large enough to meet the requirements of one month of final assembly. Circuit board production required a number of steps. First, all of the components required to produce a batch of a particular circuit board were brought from the stockroom to the circuit board production area. Then, electronic components were automatically mounted onto the surface of the bare panels. After this "surface mounting", the batch was inspected to ensure that all of the appropriate components were correctly placed. The entire batch then moved to the "hand stuffing" work area where components that could not be automatically mounted were manually inserted into the board. These "hand-stuffed" components were then soldered into place on an automated wave solderer. The entire batch of boards was then tested; boards which failed the test were marked. All panels were then sent to the "waterknife" which used a high-pressure stream of water to cut the panel into individual circuit boards, a process known as "depalletization". After depalletization, any defective boards were sent to the printed circuit board re-work area. The boards which had passed the first step were sent to a another work station for RF tuning and functional testing. Boards which failed this second test were also sent to the re-work area. Finally, the entire batch went through a final quality control audit to ensure that only good boards would be sent to the board inventory in the stockroom¹ and ultimately on to final assembly. Microwave assemblies were assembled in a separate part of the factory. Like the RF and control boards, the microwave assemblies were produced in batches large enough to meet one month's final assembly requirements. Final Assembly Final assembly consisted of six steps. First, the three major internal subassemblies were wired together. Second, a "quick test" of the integrated electronic system was done. Defective units were sent to a final assembly rework area. Third, the now-integrated electronic systems were attached to the bottom half of the unit's plastic molded shell. Fourth, the top half of the exterior shell was fastened onto the bottom half (now containing the electronic system). Fifth, the entire unit was tested (defective units were sent to the re-work area). Finally, the unit was packaged with instructions and sent to the finished goods storage area. Production Control The flow of materials was controlled by what is commonly referred to as a "batch-and-kit" method. In a batch-and-kit operation, all subassemblies and final assemblies are produced in batches-at Whistler, normally monthly batches. For example, if 5000 Spectrum 1 radar detectors were scheduled for production in a month, all 5000 would be assembled in one batch. Before the final assembly of a batch could begin, all requisite subassemblies and parts had to be ready. Several weeks before final assembly of a batch of Spectrum Is was scheduled to begin, the subassembly area produced 5000 of the appropriate RF boards, 5000 of the appropriate control boards, and 5000 of the appropriate microwave assemblies. The finished batches were then sent to the stockroom. Just prior to the time at which final assembly was scheduled, all the required subassemblies were picked from the stockroom and organized into a "kit"-all those parts required for the final assembly of one batch of finished product of that model. The kit was then sent to final assembly. Most of the small electronic components used by Whistler were supplied by vendors located in Asia. For these components, the lead time from order to delivery was about ten weeks. Due to these long lead times, and because of the relatively low cost of these components, Whistler generally stocked enough raw materials to satisfy the next two months of scheduled production. Strains in Manufacturing Strains in Manufacturing As production volume began to increase rapidly after 1983, manufacturing operations experienced a number of problems. When Whistler had been producing four models in relatively low volumes, the number and size of production batches were manageable. However, increases in both total production volume and the number of models meant an increase in the number and size of batches. With a total of 13 models in production, the factory often found itself with several batches of in-process units lined up and waiting between work areas. With available stockroom space full, the floor of the Westford factory quickly became cluttered with work-in-process. Floor space was not the only constraint in the Westford plant. There were many electronics companies in the Westford-Lowell area. Because of defense spending and a general expansion in the economy, these companies were experiencing a boom in business. Demand for semiskilled assembly workers was extremely high. Whistler, competing for labor with such companies as Digital Equipment Corporation (DEC) and Wang, found it difficult to staff its production operations at the wages it could pay. A decision was made to lease a factory in Fitchburg, Massachusetts where both labor and space were more economically available. Management viewed the Fitchburg plant as a way to Whistler Corporation (A) 690-011 quickly add capacity and capture some of the business Whistler had been losing. In 1985, the Fitchburg plant was opened. All final assembly and test operations were moved to Fitchburg. Subassembly operations and packaging remained in Westford. Thus, the process flow was altered as follows: RF boards, control boards, and microwave subassemblies were produced and tested in Westford. These subassemblies and all the other parts required to make a batch of a particular model were "kitted" and then trucked to Fitchburg. In Fitchburg, the kits were assembled into final products and tested. The entire batch (or multiple batches) were then trucked back to Westford where finished product was packaged and stored. During this time period, Whistler began to experience quality problems in the subassembly of RF and control boards. Some of these problems were due to defects in components sourced from outside vendors. However, a major part of the problem was caused by the new surface mount technology (SMT). Components which are surface mounted take up less space on the circuit board than those that must be inserted and soldered into holes. As a result, surface mounting allows more components to be packed onto a smaller circuit board. The trend toward smaller radar detectors made this technology desirable. When automated, the surface mounting process reduced cycle time dramatically. Given the need to increase production volume to keep up with demand as well as the increased demand for smaller detectors, SMT was a natural technology to adopt, even though it was a far more complex process than traditional methods. The first SMT equipment was installed in 1986. The yields on the SMT process were quite low while operators, designers, and process engineers learned about and adjusted to the subtleties of the new process. By early 1987, the first-pass yield (the percentage of "good" output the first time through, before rework) had climbed to 75%. Through diligent inspection and extensive reworks, the quality of products reaching customers remained extremely high. However, ensuring that reliability in the field was quite costly. For example, at the end of 1986, 100 of the company's 250 production workers were deployed to fix defective boards. About 30% of the Westford plant's floor space was taken up with in-process rework. Rework accounted for approximately $600,000 of the company's $2 million of work-in-process inventories. Defective boards also hampered smooth material flow. A strict batch-oriented material flow discipline requires that an adequate number of subassemblies of each type be available to complete a kit. High rates of defective subassemblies created problems in matching subassemblies to create final assembly kits. In final assembly, incomplete kits could sit for weeks waiting for defective subassemblies to be reworked. The pressure to ship products was so great that common parts from other batches in process would sometimes be used as replacements. Unfortunately, this "borrowing" of components often went unrecorded. As a result, the missing parts were not available when the other batch was ready for final assembly. High defect rates became so normal that they were built into the production schedule. For example, as a matter of policy, the Production and Control Department ordered 20% more subassemblies than the final assembly production schedule called for. Work-in-process piled up because of frequent unexpected schedule changes. When Whistler had been producing only four models for a rather narrow market, the schedule could be set well in advance. The "mass" market, however, was far more volatile. As one production scheduler put it: Out of the blue we could get an order for 5000 Spectrum 2's from a big retailer running a weekend promotion. If we had the products to ship, we had the order. If not, we lost it. This was something we just weren't used to. Truckers don't go out and buy a radar detector just because it's George Washington's birthday. production controller who had worked on the production line described the scene: "The place was a total zoo. Boards and half-assembled units were piled up from floor to ceiling. We even had to rent three trailers to handle the overflow." Work-in-process had also become a materials-handling nightmare. The longer kits remained on the factory floor and the more they were reshuffled (to make room for other work-in-process), the greater the likelihood of damage to delicate electronic parts and further problems in final assembly. In 1987, by the time a unit had completed final assembly, it had spent an average of 23 days in process, although actual production time was only eight hours. Performance Until late 1986, manufacturing was of little concern to the Whistler management. Innovative design and good marketing were, as one executive put it, "the name of the game". Sales had been growing rapidly and manufacturing, despite its internal problems, had managed to keep up. In 1985, profits before taxes were 20% of sales; the company had a pretax return on assets of 40%. In that year, the company was the market share leader with 21% of the domestic radar detector market (units). The "build as many as you can, any way you can" strategy seemed to be working. In 1986, however, financial performance deteriorated rapidly. High manufacturing costs were making it difficult to compete with off-shore manufacturers. The manufacturing costs of Asian subcontractors were substantially lower than those of Whistler (Exhibit 3). A marketing analysis suggested that, due to their performance, quality, reputation, and brand image, Whistler's products could support a 10% price premium, but not much more. By the end of 1986, Whistler's market share had fallen to 12%. By the end of the summer of 1986, the company began to lose money for the first time in its history, at the rate of almost $500,000 per month. The problems in manufacturing began to draw notice. According to Jack Turner, Vice President of design and engineering, who headed manufacturing at the time: "We knew manufacturing was sick. We knew something had to be done. We just didn't know what it was." In September 1986, a consulting firm was hired to study Whistler's manufacturing process and to make suggestions for change. The RACE-ME Program The consulting firm suggested a comprehensive program to reform manufacturing operations. The program was dubbed RACE-ME (Restoring A Competitive Edge Through Manufacturing Excellence). Its goal was to make Whistler's manufacturing as efficient as its Far Eastern competitors within 24 months. The RACE-ME program included reforms in materials handling practices, operator training. process lay-out, inspection and quality control, tooling, and production flow. A model (pilot) production line (MPL) was set up in a corner of the Westford plant to implement, evaluate, and demonstrate various reforms. The line consisted of a series of connected work benches. For symbolism, the benches were a different style and different color than those found in the main manufacturing operations. The relatively high-volume Spectrum 1 was chosen to be produced on the The MPL was designed to change the flow of materials and work-in-process. Rather than a batch flow, the MPL would operate as a "repetitive" or a synchronous line-flow manufacturing process with only very small buffer stocks between adjacent stages of production. In the traditional system, a work station received large batches of components and subassemblies as scheduled, whether or not it was ready to start working on them. On the MPL, a work station would receive materials or work-in-process only when it requested them. The second major change was in the production schedule. The MPL was scheduled to produce the same quantity of detectors day after day. After some trial-and-error, the MPL process evolved as follows: RF boards, control boards, and microwave subassemblies were produced with the same equipment and through the same process as before. However, separate inspection points were eliminated. Each work station was made responsible for identifying and correcting its own quality problems. Batch sizes and production flow were dramatically different. Rather than produce boards in one-month batches, the SMT operation ran only enough boards to supply one day of final assembly. The batch sizes were only 2 hours in the board assembly operations after SMT (hand stuffing, wave soldering, and depalletization). The production flow was controlled by "kanban"6 racks between each work station. Workers at any particular work station were instructed to work on a particular batch of boards only when an empty tray appeared in the rack between it and the next work station. When a specific work station needed more of a certain type of board, a worker would place an empty tray (with a production control card indicating the type and number of boards required) in the rack. This would signal a worker in the preceding work station to begin making a batch of those boards. The worker receiving the "produce" signal first checked the inventory in the rack between it and the station preceding it. If the appropriate work-in-process or components were there, the worker could begin processing immediately. If not, he or she would order them from the preceding station by putting an empty tray with the appropriate production control card on the rack. Through this kanban chain, materials and work-in-process were pulled as needed by downstream stations. The flow of materials through subassembly was similarly controlled by the final assembly line. When required, final assembly would pull small lots of boards (a 15 to 30 minute supply) from the kanban rack located after the depalletization area. The small lot would then be tested. Boards that failed were sent to a rework area assigned especially for the MPL. A limit had been set on the maximum number of boards permitted in the board rework station. Once the limit was reached, the board subassembly operation would be shut down until the cause of the quality problem could be identified and corrected. It was hoped that this line shut-down procedure would help keep the process under control by drawing attention to defects and by forcing shop management to deal with their causes. The boards were then passed along to the unit integration station where the three major subassemblies were joined. The MPL operated as a worker-paced assembly line. After completing a particular assembly step, the operator would pass the work piece a few feet down the bench to the next station's incoming work tray. These trays could hold a maximum of six work pieces. Once the tray was filled to capacity, the operator at the preceding station was instructed to stop working. He or she could resume assembly only when the tray contained fewer than six in-process work pieces. At the end of the MPL, an enclosed work station performed the final test. Any unit failing this final test was sent to another rework station. The entire production line would be shut down if one day of inventory accumulated in this rework area. MPL Results: April - June 1987 The MPL began producing Spectrum 1's on April 7, 1987. By late June, a study of the pilot line was completed. The data (summarized in Exhibit 4) suggested that it might be possible for Whistler to achieve substantial improvements in the productivity of people, equipment, and space if the manufacturing methods used on the MPL were used to produce all of Whistler's radar detectors. In addition, the MPL enabled Whistler to produce a Spectrum 1 from start to finish in 1.5 days, as compared to the 23 days required in the traditional system. One option was to implement MPL concepts throughout the Westford plant. The expected gains in productivity were expected to allow Whistler to close the Fitchburg plant and reduce its total labor force. However, several questions remained. Some managers in the company wondered whether the results of the pilot project could be replicated on a plant-wide basis. Which product lines might be best suited to the repetitive manufacturing process? Ed Johnson, the Director of Marketing, argued: This is extremely risky. You're talking about changing the entire manufacturing system and maybe shutting down one of our plants. Frankly, the possibilities scare me to death. Any kind of major disruption in output could really hurt us. I know things aren't very good now, but, at least we are getting product out the door. I think we should make changes in manufacturing, but let's phase them in more slowly. Larry Santos, the plant manager at the Fitchburg plant, was also concerned: I know this sounds like I am just looking after my own job, but I have to say I am not sure closing Fitchburg is the best decision. First, the plant has only been open for two years. The people there have been showing some real commitment. In fact, most of the quality and scheduling problems have originated in Westford. The Director of Operations Planning, Sharon Katz was also concerned about closing Fitchburg: I'd have to agree with Larry about closing Fitchburg, but for a different reason. If we plan to grow, we're eventually going to need the extra capacity Fitchburg gives us. It's going to very expensive to start up a new plant in a few years. The Executive Vice President of Finance, Margaret Curry, had a more extreme viewpoint: I am very impressed by the results of the MPL experiment, but I just don't think the cost savings will be enough. I just don't believe we can compete by manufacturing our products domestically. The offshore option buys us tremendous sourcing flexibility. We can shop around for the best suppliers in cost, quality, and delivery. If one isn't working out,

Answer rating: 100% (QA)

The detailed ... View the full answer