Ultimate flexibility means having viable alternatives in any situation. Standardization of parts, processes, and production systems, so that these elements are interchangeable, creates options for using them where there is a shortfall. In the event of a disruption, firms can substitute alternative parts (or part suppliers), swap out damaged components, use alternative processes, or reroute the flow of business activities.
Intel has plants in places ranging from Ireland to Israel to China to the United States. With such a widely dispersed organization, disruptions that could affect Intel include rolling blackouts in California, earthquakes in Oregon, typhoons in the Philippines, terrorism in Israel, or freight hijackers in Malaysia.
When SARS struck in Asia in 2003, chip industry analysts worried about the virus’s effect on computer chip making. Intel, Hewlett-Packard, and other multinationals reported potential SARS cases among Hong Kong-based sales and marketing staffers; almost all trade shows in Asia were canceled.1 Possible disruption of Asia’s entire electronics manufacturing base loomed large. For example, a Motorola plant in Singapore shut down when one of the plant’s 532 workers became infected. Some feared that other plants, such as Intel’s Shanghai facility, might also close as the infection swept across Asian cities and curtailed normal public activities. Intel, however, was confident that if it had to, it could withstand the closure of the plant with minimal consequences.
In a strategy called Copy Exact! Intel builds each of its semiconductor fabrication plants to the same exact specifications, creating interchangeable processes and interchangeable fabs throughout the world. Copy Exact! began in the mid-1980s as a means of coping with the inscrutable complexity of semiconductor manufacturing. The smallest variation in temperature, pressure, chemistry, or handling can mean the difference between a wafer full of hundreds of expensive chips, worth hundreds of dollars each, and a useless silicon disk. Once Intel has a new semiconductor manufacturing process debugged at one facility, it copies that process—down to the lengths of the hoses on the vacuum pumps—to other Intel facilities.2 This strategy also provides flexibility; Intel can transfer capacity and work-in-process back and forth between facilities to eliminate manufacturing bottlenecks and overcome disruptions in any given facility.3
When SARS erupted in Asia, Intel spokesman Chuck Mulloy said, “If there’s a problem, we can move the capacity around—we have designed the system so we don’t have chokepoints.”4 Copy Exact! transforms Intel’s global portfolio of facilities into a large virtual fab. Wafers can be partially completed in one fab and flown to another for finishing, without affecting yield.
The Copy Exact! mentality extends beyond semiconductor fabrication to the assembly and test factories and to the contractors who support building PC motherboards. According to Intel’s Steve Lund, “If something happens to a facility, we roll over to another subcontractor who can pick up the same assembly test and make sure that we get the same product at the same amounts for our shipping plans.”5 Copy Exact! even extends to Intel’s IT infrastructure. Identical software and hardware architecture support a range of activities, such as ordering and production planning at 18 manufacturing, testing, and assembly sites across three continents.
Other companies also replicate factories and standardize manufacturing processes. General Motors (GM), for example, operates near-identical plants in Argentina, Poland, China, Thailand, and Brazil for enhanced flexibility. The plants were built to a common template that uses the same design, processes, and technology.6 At GM, the standardized processes include material systems, body shop, stamping, fabrication, painting, and general assembly. GM even standardized the human resources strategies as well as the financial and quality control systems. GM’s interchangeability, however, is more subtle than Intel’s replication of plant and equipment. “It is not necessary to make a cookie-cutter for all our operations,” said James Wiemels, vice president, manufacturing, GM Europe.7 GM’s new manufacturing process starts with designing common primary locating points and assembling vehicles in basically the same sequence, even if a plant’s process is slightly different.
Given the difficulties of predicting demand for a particular vehicle years in advance of building a manufacturing plant, the resulting flexibility helps GM cope with uncertainty in demand. For example, GM’s Thailand plant came on line in 1997, in the aftermath of the collapse of several Asian currencies. The plant was initially intended to produce cars for the local market, but the slumping Thai economy caused GM to switch the plant to export production, making engines and cars that were sold in Japan, Europe, and Chile.8
In addition, to be able to adjust plants to changing requirements, none of GM’s five flexible plants has fixed conveyer lines that carry car bodies in a pre-designed assembly pattern. Instead, each plant can be reconfigured over a weekend.
Shifting production from one disrupted facility to an alternative one requires more than standardization of the technical capabilities of the factories (including equipment, personnel, and processes); it also requires the ability to shift the flow of inbound material into the alternative facility and the ability to service the customers from that facility. Such capabilities require flexible supply contracts and flexible distribution capabilities.
To provide strategic flexibility across product lines, beyond the ability to move production among plants, GM, like most other large automobile manufacturers, uses a joint platform strategy. Several car models share similar chassis, engines, transmissions, and many other components that are opaque to the customer. The factory, then, can use a single set of fixtures and jigs to handle multiple models of cars. It also allows the company to produce components (such as engines and transmissions) before committing to the production of specific models, thereby delaying this commitment until demand can be forecast with greater accuracy. A similar strategy is used by most automakers. For example, the Volkswagen “B Platform” is the basis for a wide range of models, from the budget-priced Skoda Octavia to the midpriced VW Golf, Beetle, and GTI to the luxurious Audi A3 and TT. All share about 70 percent of their content.9
Standardizing manufacturing processes is not the only way to create flexibility through interchangeability. When Lucent’s Switching System group in Spain faced a looming deadline for a major project in Saudi Arabia, the company’s Tres Cantos factory could not handle the project’s needs for circuit packs and cabinets. Worse, differences across Lucent’s product line meant that Lucent could not use its factories in Oklahoma City, Poland, and Taiwan to help out. In response, Lucent redesigned the products to use standardized parts and adjusted its processes so that several factories could contribute to the project. Standardizing the designs and processes let Lucent use multiple capacities, complete the project on schedule, and avoid disrupting an important customer.10
The Saudi project made Lucent realize the high, ongoing cost of the company’s proliferation of independent designs, components, and suppliers. To make this point internally, Lucent Supply Chain Networks (SCN) president Jose Mejia brought into one room all of the different types of product enclosures that Lucent used for their network systems and invited Lucent’s senior supply chain executives to come take a look. The visual effect of seeing 47 different enclosures with similar functionality got the point across. As Mejia commented, “these products looked like they were made by different companies.”11
In fact, Lucent’s SCN organization was created by Mejia in 2001, in part, to solve this problem.12 The organization was given broad responsibility for all supply chain costs, from researchand-development, to engineering, to manufacturing, to sales, and to distribution and installation.13 SCN drove standardization throughout Lucent’s operations with impressive results. The number of enclosures was reduced from 47 to seven; the number of shelves from several hundreds to three, the number of fan controller configurations from 92 to four platforms, and the number of specified filters from 466 to 15. The increased flexibility that resulted from this action helped return the company to profitability in the third quarter of 2003.
Lucent applied the same concepts of standardization to its supplier relationships. It went from spending 40 percent of its procurement expenditure with its 1,000 top suppliers to concentrating 80 percent of its supply expenditures with 60 top suppliers. Having narrowed the number of suppliers, Lucent then reached out to develop deeper relationships with them, putting emphasis on risksharing relationships to help seek new business for both Lucent and its suppliers. This included collaboration on proposal development, in which all parties shared the costs of investing in proposals and prototyping for potential customers (without bidding with other partners at the same time), thus winning or losing together.
In addition, the suppliers shared with Lucent the liability for failing to meet any contractual obligations to the customers. Inevitably, Lucent and its suppliers became intimately familiar with each other. The resulting trust and mutual awareness provided Lucent with even more flexibility, and together with its suppliers, Lucent could sense and respond more quickly to rapidly changing market conditions or to special requests. In particular, Lucent’s average supply lead times to its customers were almost halved between 2002 and 2003.
When Toyota examined its use of P-valves following the Aisin fire (see chapter 13), it realized that it had too many different types of the small auto brake component. One hundred varieties of the simple part had complicated and delayed the recovery efforts after the fire. Toyota launched a redesign effort across its product line and reduced the number of different P-valve types to a dozen. This action was taken with two goals in mind:
Increase corporate resilience by simplifying requirements from other manufacturers, in case the need ever rose again to bring new suppliers on line quickly.
Simplify the forecasting requirements by using the same P-valve across several product lines, allowing demand to be aggregated over all the models that shared a given P-valve.
In some cases, existing standards are so commonplace that their contribution to flexibility is taken for granted. For example, when Ford recalled 6.5 million tires used on Ford Explorer SUVs, the standardization of automobile tire-mounting throughout the industry allowed Ford to quickly procure tires from other tire makers. Had tire sizes been nonstandard, the time required to provide a safe alternative for customers would probably have been months.
In general, the use of standard parts and commodity components presents a trade-off between the exact fit to a specific purpose (as presented by specially engineered parts) and the faster time-to-market and higher availability of standard parts. Using standard designs broadens the natural base of suppliers, letting the company procure from other suppliers in case one supplier is disrupted. It requires, however, higher levels of internal coordination while engineering new products.
The familiar brown uniforms of the UPS driver only hint at a deeper kind of uniformity at the global package carrier. UPS standardizes more that just the appearance of its 357,000 employees; it standardizes work processes to the minutest detail, down to telling package delivery truck drivers to snag the keys of the truck with the third finger of the right hand as they open the truck door with the left hand. Such uniformity creates flexibility for the giant package delivery company, which in 2003 handled shipments representing 7 percent of the U.S. GDP. That year it delivered an average of 13.6 million packages a day to eight million customers in 200 countries with its 88,000 brown vehicles and 575 brown planes.
UPS Worldport in Louisville is the central sorting facility for UPS air packages. Late each night, hundreds of planes come in and unload their packages. The packages are then sorted and loaded onto outbound planes, lifting off in the early hours of the morning to make deliveries all across the United States, when firms open for business. In 2002, the facility had the capacity to sort through more than 300,000 packages an hour, supported by computers twice as powerful as those running the New York Stock Exchange and with software capable of more than a million database transactions per minute.14
Sunday, January 16, 1994, began as a normal winter’s day for UPS’s bustling Louisville hub. Weather forecasts predicted a light dusting of snow as a cold front wound its way across Kentucky in the evening hours. But the forecast was wrong. Instead of one or two inches, the storm blasted Louisville with 16 inches of sleet and snow. With the blizzard came record cold temperatures—as low as 22 degrees Fahrenheit below zero (-30°C)—which froze the heavy snow to the roads. These arctic conditions paralyzed all transportation in and around Louisville.
What aggravated the conditions was that the city of Louisville was unprepared. Local snow plows had rubber-edged blades, which were designed to reduce wear and tear on the roads but such blades made them incapable of removing the thick ice laid down by the storm. Many of the plows broke down as they attempted to clear the heavy snow. The ultra-low temperatures made road salt ineffective. The city had no choice but to close all roads and institute a travel ban. Even ambulances became mired in the snow, forcing the city to use fire trucks instead. And even the fire trucks had trouble.15
The storm began at 8 P.M., just as some 100 UPS planes were preparing to take off all around the country loaded with packages bound for Louisville for sorting and delivery. As the weather took a turn for the worse, UPS realized that this would not be a normal night. The first order of business was to divert the airborne planes to alternative airports as Louisville’s airport shut down. Next, the company had to attend to the hundreds of thousands of packages already sitting in its Louisville hub, each representing a customer eager for delivery.
Clearing the snow-clogged runways at the Louisville airport took less than a day, but the roads were a different matter. Local and state government declared a state of emergency and forbade nonessential travel for five days as the heavy snow bogged down vehicles and disabled two-thirds of the snow plows in the region. “We actually were up and ready for the next-day-air sort on Tuesday night, but of course, we had thousands of employees who couldn’t get to work,” spokeswoman Patti Hobbs said.16
But the company did not give up. “Our employees couldn’t drive to work . . . but they could fly,” Hobbs said.17 With the airport open, UPS flew in employees from other UPS locations to help process the stranded packages and load the planes, getting out the packages stuck in the Louisville hub. The company negotiated with local authorities for a limited waiver of the travel ban so that the jet-commuting employees could be driven to nearby hotels for the night.
UPS could use workers from other facilities because of its uniform practices. UPS’s sorting machines were interchangeable, the processes were interchangeable, and the people who were familiar with one location could operate any of them.
Standard processes also help UPS respond to demand fluctuations. For example, when the package volume booms every December, supervisors and managers help work the hubs—unloading, sorting, and loading packages. During a Teamsters strike against UPS in 1997, these seasoned managers kept the business going. The standardized processes mean that anyone who knows any UPS operation can contribute, and that means that UPS is flexible, whether it faces an unexpected blizzard or a busy Christmas season.
For UPS, widespread standardization not only provides flexibility but also underpins performance improvement processes. Because every UPS outpost uses the same processes, the same machinery, and the same training, the company can cross-compare performance at different divisions, different regions, and different delivery routes. UPS then works to improve low performers (whether they are individual employees, a facility, or a region) and lauds and copies the high performers.
Helix Technology, a maker of vacuum instruments and equipment, uses Demand Flow Technology, an operational strategy descendent of lean manufacturing. The strategy is designed to increase manufacturing flexibility in a lean manufacturing environment. As part of the implementation of this strategy, Helix divided its manufacturing processes into small, well-documented pieces (sub-processes). Workers in “cells” are trained to perform the set of standard tasks constituting each sub-process. The result is a much-simplified set of work assignments requiring less training for each worker to achieve proficiency.18
In concert with other process changes, this strategy helped Helix improve asset utilization and accelerate material flows throughout its manufacturing processes. It had, however, a powerful disruption-related benefit as well. By atomizing the tasks, workers could perform a greater variety of sub-processes, each consisting of a set of simple subtasks. This created a flexible worker environment. The inherent cross-training means that many workers can substitute for one another. It also means that new or substitute workers can be quickly trained in the event of disruption. If one of its facilities is interrupted, Helix can even transfer tasks to suppliers, because the work has been so scripted that it can be easily transferred. More than 3,200 companies, including GE, Boeing, and Flextronics, use Demand Flow Technology to improve performance and flexibility.
Other companies also use various types of standardization to create interchangeability, which in turns creates flexibility. Southwest Airlines flies nearly 400 airplanes—all of them Boeing 737s. The company operates a tight schedule with each plane spending only 20 to 30 minutes on the ground between flights and being expected to make seven flights each day, giving the company little time for recovery from disruptions. The use of a single aircraft type means that Southwest can swap airplanes and swap crews.
Southwest’s penchant for standardization extends to the last detail; the company standardizes the cockpits, even as cockpit technology has evolved over the years. In more recent models of 737s, Boeing designed a “glass cockpit” with computer screens replacing old-fashioned analog dials. But in order to maintain interchangeability, Southwest asked Boeing to program the new displays to look like the old “steam gauge” dials and indicators that are so familiar to Southwest pilots.19 Interchangeable cockpits mean interchangeable pilots and reduced training costs. If a plane has a maintenance issue or a pilot gets delayed, Southwest need not disrupt its time-sensitive flight schedule; any Southwest pilot can fly any Southwest plane on any Southwest route.
Other low-cost airlines use the same strategy. Ryanair, the European discount airline, uses only Boeing 737s, while JetBlue, the U.S.-based discounter, uses Airbus A320s on all of its long-haul flights and Embraer 190s for all short-haul flights. When EasyJet (which is Ryanair’s main competitor in Europe) decided to buy Airbus A319 aircraft to add to its 48 Boeing 737s on October 13, 2002, its stock fell by 5 percent. According to the BBC, the market had “worries that the cost of mixing Airbus planes into Easyjet’s Boeing fleet would prove expensive for the airline.”20
The benefits of interchangeability come from risk pooling. Assume that a computer manufacturer makes four types of PCs, each with a different case. Because demand for each of the computers is uncertain, the manufacturer needs to keep a certain level of safety stock of these parts to ensure that its production line does not run out of the parts too often. (That is usually expressed in terms of “fill rate,” which is the percentage of occurrences when the needed part is available in inventory.) If the manufacturer, instead, decided to use the same case for all four computers, it would be able to cut its safety stock in half but still keep the same fill rate.21 Alternatively, the manufacturer could keep the same level of safety stock, resulting in higher fill rates. The increased fill rate is rooted in the flexibility offered by the interchangeability of the cases. If the demand for one computer type increases, the demand for another may go down, and vice versa.
Standardization creates interchangeability, which creates flexibility to respond to disruptions. Standard factories (Intel, GM), standard equipment (Southwest, Ryan Air), standard components (Lucent, Dell), and standard processes (Helix, UPS) allow companies to respond to disruptions effectively. Companies can move personnel around, move work across their plants (and even their suppliers’ plants), or change suppliers of parts and services.
In many cases, interchangeability implies “sameness” of parts or products so that they can be moved from surplus to deficit locations. But growing customer expectations, increased rate of product introductions, and strong competition mean that companies have to introduce more varieties of many of their products, thereby increasing the need for extra stock and diminishing the flexibility offered by interchangeability. The next chapter looks at a strategy of combining large production batches with the ability to offer many product variants, thus retaining flexibility while satisfying customer preferences.