|
 |
|||||||
We want to know how to serve you better. Through the development of additional eBooks, Video Downloads, and Direct Interactive Problem Solving Methods, we hope to do just that. But to do this we need information from you! Fill out our survey or send us an email so we can get directly involved in what you need & how best to provide answers:
THE JUST-IN-TIME (JIT) APPROACH
SUMMARY
JIT presentations often employ the analogy of a stream when describing proper inventory management. Well-managed systems achieve a flow of inventory from raw material to the customer like a smooth river, unimpeded by shoals of scrap or machine breakdown or other problems. This concept did not originate with the Japanese. Henry Ford's River Rouge plant regularly converted iron ore into a Model T in four days. However, in recent years, especially the 1970s, American business has not improved its manufacturing capability quickly enough to maintain a competitive position in cost or quality or market responsiveness or flexibility.
In 1983 APICS began a zero inventory crusade---strongly advocating JIT. Firms such as CM, Ford, Chrysler, Bendix, Harley-Davidson, IBM, Hewlett-Packard, AT&T, and others began the journey even earlier. There has been much progress on regaining competitiveness in recent years. This progress has been achieved by emphasizing continuous improvement, reduced inventories, expanded roles for hourly workers, fewer levels of management, longer term relationships with customers and suppliers, and an emphasis on providing value to the customer, Many American firms are once again at or near world class status. We should remember, however, that complacency is the principle barrier to maintaining world class status. We must adopt the philosophy of Kaizen, continuous improvement. The Japanese underscore the urgency of maintaining competitiveness with a phrase taught to every schoolchild, "Export or Die!" In yesterday's world, export or die was a truism for any island economy. In today's global village, export or die is a truism for all economies.
What is Just-In-Time? / Why is JIT Important?
To prosper---and often even to survive---manufacturing companies must provide value at least equal to that of competitors. Today, manufacturing competition includes plants located in many different parts of the world. For example, some refrigerators sold at major department stores in Canada are assembled in Wroclaw, Poland, using a condenser manufactured in Sao Paulo, Brazil. Much of the world is one big market, with goods crossing many different types of boundaries.
Although international trade always has existed, it has exploded in the last few decades. Improved communication and transportation have been contributing factors, but the primary cause has been dramatically improved manufacturing productivity---with emphasis on both quality and cost. Although Japan has been in the forefront of this advance, South Korea, Taiwan, Malaysia, Singapore, and Thailand have made remarkable strides. Progress is also taking place in Mexico and Brazil, and most Western European countries have continued to improve their industrial capability. In addition, it is not unreasonable to expect that Eastern European countries will improve their competitive position as they revise their economic policies.
These developments and a benign neglect of manufacturing by top management in many North American firms caused foreign trade balance deficits and a lower productivity growth in the United States during the 1960s and 1970s than in many other countries. Some U.S. companies lost market share and others lost markets. The MIT commission on industrial productivity reported a large and increasing balance of trade deficit in automobiles, consumer electronics, machine tools, semiconductors, and textiles (Dertouzos et al. 1989). Continuation of such a pattern can have dire consequences for the quality of life in any country. Foreign debt, currency devaluation, and loss of markets and profits eventually not only affect the ability of consumers to purchase material goods such as toasters and automobiles but also limit a nation’s ability to support health care, the arts, education, and recreation activities. In brief, the standard of living can decrease dramatically. For example, Argentina was a relatively prosperous country at the turn of the century, but today its economy is in shambles. Nearly all citizens suffer when such a change occurs.
The MIT commission observed:
A large continental economy like the United States will not be able to function primarily as a producer of services in the foreseeable future. One reason is that it would have to rely on exports of services to pay for its imports, and this does not seem realistic. In 1987 gross U.S. exports of services, excluding income from overseas investments and overseas sales of government services, were worth $57 billion, whereas the total value of goods and services imported into the United States was about $55 billion. - - , The United States thus has no choice but to continue competing in the world market for manufactures. The ultimate scale of American manufacturing is not known, but it will not be trivial. The important question is not whether the United States will have a manufacturing industry but whether it will compete as a low-wage manufacturer or as a high-productivity manufacturer. (Dertouzos et al. 1989, 39-40)
Clearly it is preferable to compete as a high-productivity manufacturer.
These considerations led many organizations in the United States, Canada, and other countries to examine successful manufacturing organizations in North America and throughout the world to identify the operating characteristics and practices of companies capable of competing in the present worldwide market. The essential characteristics of a such a company are that it produces high quality products at low cost and that it responds quickly to customer requests for delivery, changes in design, and changes in volume, When a company has achieved these goals, it can compete with anybody, anywhere. It is important to understand that both high quality and low cost are relative terms; continuous improvement is needed to maintain high relative quality and low relative cost. Referring to the degree of change needed to achieve world class status in Thriving on Chaos (1987), Tom Peters notes:
Radical changes in organizational structure and procedures are called for, Layers of management must be reduced in most big firms by 75 percent. Product development time and order lead time must be slashed by 90 percent. Electronic/telecommunication linkups to customers and suppliers must be developed posthaste. Just listening to customers and dealers needs to become the norm-and as yet it is not.
Different terms are used to identify the process of improving manufacturing productivity with emphasis on high quality and low cost: the Just-in-Time (JIT) approach, zero inventory, total quality management, world class manufacturing, and the search for excellence. We are using JIT because it seems to have been the first, and all of the essential concepts are inherent to it. The title of the process is not important; adopting the philosophy and pursuing its operating objectives are. This section includes concepts, approaches, and practices that may have originated under the aegis of programs with each of the different labels and titles given to various productivity improvement programs.
THE PHILOSOPHY OF JIT
JIT is a philosophy embodying various concepts that result in a different way of doing business for most organizations. The basic tenets of this philosophy include:
A. All waste, anything that does not add value to the product or service, should be eliminated Value is anything that increases the usefulness of the product or service to the customer or reduces the cost to the customer.
B. JIT is a never ending journey, but with rewarding steps and milestones.
C. Inventory is a waste. It covers up problems that should be solved rather than concealed. Waste can gradually be eliminated by removing small amounts of inventory from the system, correcting the problems that ensue, and then removing more inventory.
D. The customers' definitions of quality, their criteria for evaluating the product, should drive product design and the manufacturing system. This implies a trend toward increasingly customized products.
E. Manufacturing flexibility, including quick response to delivery requests, design changes, and quantity changes, is essential to maintain high quality and low cost with an increasingly differentiated product line.
F. Mutual respect and support based on openness and trust should exist among an organization, its employees, its suppliers, and its customers.
G. A team effort is required to achieve world class manufacturing capability. Management, staff, and labor must participate. This implies increasing the flexibility, responsibility, and authority provided to the hourly worker.
H. The employee who performs a task often is the best source of suggested improvements in the operation. It is important to employ the workers' brains, not merely their hands.
JIT is a very eclectic approach. It includes many old ideas and some new ones and relies on basic concepts from many disciplines, including statistics, industrial engineering, production management, and the behavioral sciences. But first and foremost, it is pragmatic and, thus, empirical. Discovering "what works" and why it works requires that plant operations be studied thoroughly. This requires the collection and analysis of relevant data concerning the plant's operation and its performance. This pragmatism causes the manufacturing process and its environment to be viewed as a research laboratory, similar to a university hospital, in that the primary task may be to produce quality output but another important goal is to learn how to do it better the next time.
Traditionally, inventory has been viewed as an asset, one that can be converted to cash. The Just-in-Time view is that inventory does not add value but instead incurs costs, and thus is a waste. Holding inventory is analogous to not receiving any interest for a deposit in a bank and, furthermore, paying to keep it there. Traditionally, holding inventory was seen as being less costly than correcting the production and distribution efficiencies that inventory overcame. For example, large lot sizes spread the cost of expensive setups across many parts. JIT takes a different view.
JIT views inventory as a symptom of inadequate management, a method of hiding inefficiencies and problems, see Figure 1. Inefficiencies that cause inventory include: long and costly setups, scrap, lengthy and widely varying manufacturing lead times, long queues at work centers, inadequate capacity, machine failure, lack of worker and equipment flexibility, variations in employee output rate, long supplier lead times, and erratic supplier quality. JIT emphasizes that solving each of these problems will reduce the need for inventory and improve productivity. It strives to have the right material, at the right time, at the right place, and in the exact amount. Thus, the name "Just-in-Time" is used by many to designate an organized and continuing program to improve operations productivity.
In Kaizen (1986), Masaaki Imai argues that the most important aspect of JIT is a philosophy of continuous improvement. He explains that although Westerners and Japanese both ascribe to improvement, he has discovered that the two cultures have 'different concepts of what this term means. Westerners think of improvement as a step function---a change represents a marked increase in performance. That level of performance is held until the next performance leap is introduced. The Japanese view continuous improvement as an upward sloping line-driven by numerous incremental improvements, Each improvement is in itself imperceptible, but collectively the changes made in a few months wilt represent a great deal of progress.
This difference in culture can be seen in the management of suggestions, Western companies offer large rewards for suggestions that substantially reduce company costs. In a typical year a few hundred suggestions may be received, a small percentage of which actually are implemented. Toyota, by contrast, offers a small, fixed fee---less than $1---per suggestion. They receive hundreds of thousands of suggestions each year and implement more than 90 percent of them. Imai contends that the total improvement achieved by emphasizing incremental improvement is greater than that achieved by emphasizing dramatic improvement. Certainly the performance of Toyota in recent years represents a strong ease for Imai's point of view.
The Japanese do not neglect dramatic improvement, either. The books of Shigeo Shingo (1985, 1986,1988) explore from an engineer's perspective the process of analyzing operations for opportunities for dramatic improvement, One story illustrates his approach. A client of Shingo's, a manufacturer of engraved brass plates, was seeking a way to efficiently remove the lubricating fluid used to cool the engraving pen, because cleaning the engraved plate represented the largest single process cost. Shingo, on reflecting on the purpose of the fluid-cooling plus debris removal, suggested the use of a focused stream of compressed air rather than fluid. The company's management felt that air would not properly protect the pen and would shorten the pen's life. On Shingo's insistence, the company tried it and found that the method not only eliminated the need to clean the plate after engraving but also actually extended the pen life by 30 percent. This story illustrates two aspects of continuous improvement---careful analysis combined with a willingness to try new approaches, even when they seem unpromising.
The JIT approach includes the following:
A. Reduction of setup times to achieve smaller production lot sizes
B. Increased use of sequential flow processes such as dedicated assembly lines and group technology cells
C. Increased use of multifunction workers
D. Increased flexibility of equipment and capacity
E. Increased use of preventive maintenance
F. Increased stability and consistency in the schedule
G. Longer term relationships with suppliers
H. More frequent deliveries from suppliers
I. Improved technical support of suppliers
J. Employee involvement programs such as quality circles
K. Statistical process control (SPC)
L. The stop production prerogative
M. Cause and effect analysis
Although this new philosophy affects all areas of a business, major changes take place in manufacturing management, purchasing, human resources management, and quality management.
Just-In-Time MANAGEMENT: A GENERAL OVERVIEW
Large work-in-process inventories can be the result of a number of difficulties, including: lengthy setups, long queues at work centers, material waiting to be moved to the next operation, long distances between work centers, uneven loads from one period to the next, equipment and workers with limited flexibility, unexpected equipment failure, and large safety stocks to cover possible scrap. Let's consider how the JIT approach can solve some of these difficulties.
Setups
Reduced setup times have two important benefits. First, they allow lot sizes to be reduced without any setup cost penalty. Second, because of the increased ability to switch production between different items, manufacturing can respond quickly to different customer demands. There arc a host of principles and guidelines for reducing setup times. Many of them apply to specific types of equipment; however, a few are universally applicable. The most prominent of these guidelines is that setup operations should be divided into internal elements (those that must be done when the machine is not operating) and external elements (those that can be performed when the equipment is operating). Performing as many of the setup activities as possible while the machine is operating can often reduce setup time substantially. Shingo (1985) estimates that setup time can be reduced 30 to 50 percent merely by separating internal and external procedures. Other universal guidelines include:
A. Modify equipment by the use of standard die height, locator pins, etc. to eliminate the need for adjustments.
B. Be sure all needed material and tools are available when the setup is scheduled to begin.
C. Videotape the setup operation for analysis and training.
D. Study the setup process in order to design a standard setup process. Prepare setup process sheets that list the elemental tasks in each setup. List elements according to their priority for future improvements using Pareto's (ABC) analysis.
E. Color code all connections: air, hydraulic, water, electrical, etc. Use quick disconnects.
F. Involve tool and die designers in setup reduction programs so that all new designs incorporate quick changeover concepts.
G. For large dies, improve transportation time by using a standard table to roll the die into position. Standardize machine bed heights so the die can easily slide from the table to the machine bed.
H. For small tools (e.g., hand power drills), use duplicate tooling to avoid setup.
I. Involve the people that do the work and know the equipment.
J. Make improvements as they develop---including the small ones.
See Shingo (1985), Hall (1983), and Schonburger (1987) for a further discussion of setups. The above recommendations point out the importance of "Housekeeping."
Housekeeping
It would seem obvious that an organized and uncluttered work place is conducive to efficient and effective manufacturing. Visiting some plants reveals that not everyone is sufficiently convinced. Poor housekeeping includes random location of tools, dirty equipment, poorly lighted areas, and cutting oil, material remnants, and chips on the floor. These lead to lost and damaged tools, accidents, and slower than necessary setups, and reflect an attitude that accuracy, appearance, and quality are not important. In general, poor house-keeping sends the message that: "What takes place here is not important." The effect of this message on effective operations may be greater than just the physical obstacles of poor work place organization. Few would argue that good housekeeping is a prerequisite to JIT. A few basic rules for housekeeping include:
1. Keep tools clean, lubricated, calibrated, sharp, and in their designated location.
2. Clean, inspect, and repair tools during or immediately after teardown (end of setup).
3. Classify tools, jigs, attachments, and supplies on the basis of the frequency of their use and store accordingly.
4. Store materials and supplies in designated locations Mark each location so that what belongs there is clearly visible, both when the spot is empty and when it is occupied. Visibility is crucial for identification of parts, tools, and supplies.
5. Avoid the "Call Housekeeping" syndrome. Let each employee or small group be responsible for the order and cleanliness in an area. This motivates everyone to keep things clean and in order and also raises the importance of housekeeping.
Although housekeeping is not the most exciting topic, the improved quality and productivity that often follow improved housekeeping can be very exciting.
Total Preventive Maintenance
Total preventive maintenance (TPM) includes both preventive maintenance (PM) and continual analysis of and improvements to equipment, tooting, and the work place organization. TPM increases flexibility, reduces material handling, and improves flow.
When equipment fails, work in process increases at upstream work centers as queues increase, and downstream work centers are idle due to the lack of incoming parts. In addition, deliveries are delayed, and scrap is often produced before the operation is stopped. One advantage of preventive maintenance over repair maintenance is that PM can be scheduled when the machine is not needed. Repair cannot. Furthermore, if PM is performed properly, there is less total maintenance effort required than when each component failure is repaired in isolation. Much lip service has been paid over the years to the value of PM. More often than not PM was practiced in the breach rather than on a regular basis.
Preventive maintenance is essential in many types of equipment, from an automobile to a commercial airliner. Both the failure of the braking system in the former and the failure of a hydraulic system in the latter can produce disastrous results. JIT stresses FM using statistical analysis and the knowledge of the operator. Statistical analysis has revealed when fan belts and hoses on an automobile should be replaced to avoid a breakdown. In the same way, it can reveal that specific adjustments and component replacements must be made on many types of equipment after so many hours of operation (or number of items processed) to avoid failure during processing.
Preventive maintenance begins with simple housekeeping procedures. For example, if equipment is kept clean, signs of trouble, such as oil leaks can be spotted before seals fail completely. And, just as the experienced driver can foresee potential failures by sensing unusual noises and vibrations or spotting a fluid leak, the experienced operator often can sense potential equipment failures by closely observing the behavior of equipment. When a worker is responsible for maintaining a machine, he or she tends to correct problems when they occur. When a maintenance person is responsible for repairing the machine, the worker tends to let the machine continue to run until it fails catastrophically. The combination of the statistical analysis of the reliability and life expectancy of components, the alertness of experienced operators, and scheduled preventive maintenance can substantially reduce unplanned downtime. This reduces the need for some inventory and improves delivery performance.
Manufacturing Flexibility
The ability to rapidly shift production from one item to another is a function of many factors, 'including setup times, work rules, worker flexibility, and equipment flexibility. Setups were discussed previously. Recently, labor unions have been more receptive to changes in work rules that allow workers to perform operations previously restricted to a specific worker classification. These changes combined with training workers to perform various operations are essential to increased flexibility. Such training is well received by most workers when iris introduced by an education program that demonstrates its relationship to the survival and prosperity of the firm and, thus, their future employment. When production quantities do not justify dedicated equipment, equipment flexibility can often be increased substantially with minor equipment modification.
APPENDIX “A” : PROCESS FLOW AND LAYOUT
Applications of JIT
Job shops have traditionally been organized with equipment performing similar functions located in the same department. For example, drill presses, lathes, milling machines, and welding equipment would each be in a different location (department). Thus, each order must be moved from department to department as required by the manufacturing process. This brings about material handling, material waiting to be moved, occasional damage to material during movement, and personnel and workers invested in moving material. None of these activities add value, JIT aims at eliminating these activities by changing from a job shop process to a flow process.
Because major benefits of JIT usually are achieved as Improvements in the process flow are made, we describe these changes and their benefits in detail using the following examples:
I. Deterministic Simulation, One Product, Four Operations
A. Job shop
B. Operation overlapping
C. Sequential flow
II. Stochastic Simulation, Five Products, Varied Operation Sequences
A. Job shop
B. Dedicated flow lines: two products
C. Job shop: reduced setups and improved quality
Our first example reveals the benefits of operation overlapping. Our second example illustrates the advantages of moving to a sequential flow process and of reducing batch sizes where non-sequential flow remains.
Example 1: Deterministic Simulation, One Product, Four Operations
The first step in the movement from a job shop to a sequential flow process is often operation overlapping followed by the application of group technology concepts and the development of manufacturing cells. This example portrays such a movement and its results. The process has four operations with the following operation, setup, queue, wait, and movement times. The lot size is 200 units.
A traditional measure of manufacturing efficiency (ME) is calculated by dividing the total of setup time and operation time by the total manufacturing lead time (MLT). Thus, for this example:
Total Setup Time + Total Operation Time = 315 + (200 x 21) = 4,515 minutes
MLT = Sum of Q’s + Sum of S’s + (N x Sum of O’s) + Sum of W’s + Sum of M’s
MLT = 1,920 + 315 + (200 x 21) + 960 + 45 = 7,440 minutes
ME = 7,440 / 4,515 = 0.607 or 60.7 percent
This seems to be a rather good performance and is not surprising because the queue and wait times are relatively low for a traditional job shop. However, the 60.7 percent is misleading because 199 parts are in queue or waiting while each part is being processed, and all of this time is counted as processing time. A better measure is required.
Value-Added Efficiency: A more accurate measure of manufacturing efficiency is obtained by dividing the processing time, the only time when value is added, by the total manufacturing lead time of the part. We have chosen to designate this measure as the value added efficiency (VAE). In the example,
VAE = 0 / MLT = 21 minutes / 7,440 minutes = 0.0028 or 0.28 percent
This result is not very impressive, but it is an accurate measure of the percentage of time each part is being processed, that is, having its value increased.
Operation Overlapping: Operation overlapping can increase the VAE substantially. For example, reducing the queue times to 30 minutes each and the waiting times to 15 minutes each and performing the setups prior to the operation as shown in Figure 2, gives the following transfer lot sizes for Operation 1.
=
O1 + O2
=
4.5 + 5.0
=The transfer lot sizes for the other operations are calculated in the same manner.
The analysis shown in Figure 2 reveals that the manufacturing lead time has been reduced to 2,680.5 minutes by processing the second transfer lot (Q2) for each of the first three operations while the first lot (Q1) is being processed in the next operation. This results in the following improved VAE.
2,680.5 minutesThis is nearly three times the former value of 0.28 percent, a marked improvement but still not very good. Proceeding further with operation overlapping and dividing the lot into 20 transfer batches of 10 units each provides the data for the analysis shown in Figure 3. Note that each transfer batch arrives at the next operation prior to the completion of the previous batch.
The VAE is calculated as:
1,395 minutesThis is approximately double the previous performance, a noteworthy improvement. But there is still ample room, 98.5 percent, for further improvement.
Manufacturing Cells: Group technology and manufacturing cells can improve the VAE by identifying parts with similar processes and locating personnel and equipment dedicated to manufacturing these items in one location. Manufacturing cells reduce setup times and lot sizes by using common setups. The proximity of equipment reduces material handling, and parts completed in one operation are immediately available in many cases for processing in the next.
Now we examine a manufacturing cell with a transfer batch of I unit, material arriving every hour in standard containers holding 10 units, no movement between operations, and 10 finished units being moved every hour from the final operation to finished goods. In addition, the process has been redesigned so that the smallest operation time is now 5.0 minutes, the largest is 5.5 minutes, and 6.0 minutes are allowed for each operation. The total processing time is still 21.0 minutes, Figure 4 describes this situation.
Figure 4, based on the average queue and wait times, gives a manufacturing lead time of 84 minutes. This results in a VAE of 21 / 84 = 0.25 or 25 percent. This performance is roughly one hundred times better than the performance of the job shop with which we began and roughly seventeen times better than the best operation overlapping performance. Improvement in actual situations depends on the present efficiency, the nature of the product and process, and the skills and creativity of those making the improvements.
Example 2--Stochastic Simulation, Five Products, Varied Operation Sequences
Example 1 purposely oversimplifies the situation for the sake of clarity. Now we use stochastic simulation to examine a more realistic example that has more than one product, different routings, scrap, and equipment failures. This example illustrates the advantages of moving to a sequential flow process and of reducing batch sizes where non-sequential flow remains.
Imagine a shop that has four departments, each with four machines. The shop produces five part types. The part routing, run time, and demand data are shown below:
Part Daily Demand Routing Run Time 1 400 A-B-C-D 4.5, 5, 5.5, 6 2 200 B-D-A-C 4.5, 5, 5.5, 6 3 80 C-A-D-B 4.5, 5, 5.5, 6 4 80 D-C-B-A 4.5, 5, 5.5, 6 5 40 C-A-B-D 4.5, 5, 5.5, 6 Thus, Part 1 begins at Station A, where it requires 4.5 minutes per part to produce; it then proceeds to Station B, where it requires 5 minutes per part; then on to Station C for 5.5 minutes per part; and finally to Station D for 6 minutes per part. Part 2 is routed first to Station B then to Station D, and so on. In addition, assume Station A requires 90 minutes to change from one part to any other part, Station B requires 60 minutes, Station C requires 75 minutes, and Station D requires 90 minutes. Average daily demand for each part is shown above. Each part is made in batches of 100 each. On the average, a batch of 100 is released to the shop every 3 hours. The process of releasing individual batches to the shop can be modeled as a Poisson arrival process (i.e., time between release of batches follows a negative exponential distribution with a mean inter-arrival time of 3 hours).
Each of the five parts is subject to scrap, There is an inspection station immediately after the final operation for each part. There is a 4 percent chance that any given lot is scrapped. When a lot is scrapped, a replacement lot is started at the first operation for that part. Each of the 16 machines is subject to random breakdowns; each machine is available 95 percent of the time. The mean time to repair a machine ranges from 50 minutes for Work Center D to 400 minutes for Work Center A. In all cases, the mean time to failure is 19 times as large as the mean time to repair. Figures 5 and Figure 6, show printouts of the simulation of the shop.
As discussed for Example 1, a traditional measure of manufacturing efficiency (ME) is calculated by dividing the total of setup time and operation time by the total manufacturing lead time (MLT). The total of setup time and operation time for a batch of 100 is:
Setup Time + Operation Time 315 + (100 x 21) = 2,415 minutes
The MLT for a batch of 100 was found by simulating the shop using GEMS. The average manufacturing lead time was found to be approximately 3,233 minutes, (In Figure 5, under Time and Associated Network Cost Statistics, find the row labeled "Stock" and the value labeled "Mean." This value represents the average time a batch spent from release to stock---the manufacturing lead time. The standard deviation of manufacturing lead time and the minimum and the maximum times also are shown.)
Thus,
ME = 2,415 minutes / 3,233 minutes = 0.747 or 74.7 percent
This seems to be a rather good performance and not surprising because the queue and wait times are relatively low for a traditional job shop. In fact, a scrutiny of Figure 5 reveals that only 557 minutes are spent waiting in queue. (In the section labeled "Queue Box Statistics," results are reported for five queues. Four of the queues are labeled "Work Center A," "Work Center B," etc. Within the statistics reported for each queue is a row marked "Waiting time." In the column labeled “mean” is the average waiting time for a batch of 100 parts. The average waiting time at Work Center A is approximately 121 minutes; at Work Center B, it is approximately 114 minutes; at Work Center C, it is approximately 150 minutes; at Work Center D, it is approximately 172 minutes. Adding these, one obtains an average time in queue of 557 minutes.) Each batch also spent 192 minutes being moved and waiting to be moved, (The queue labeled "Material Handling" reports a mean waiting time of 18 minutes. Actual move time averaged 30 minutes. The total of 48 minutes per move times 4 moves yields 192 minutes of move time.) There is a maxim that a job processed by a traditional job shop spends 90 percent of its time waiting in queue. For this job to spend 74.7 percent of its time actuality being worked on is exceptional.
However, the 74.7 percent is misleading because 99 parts are in queue or waiting while each part is being processed, and all of this time is counted as processing time. As noted earlier, a more accurate measure of manufacturing efficiency is the value added efficiency (VAE), obtained by dividing the processing time, the only time when value is added, by the total manufacturing lead time of the part. In the example,
VAE = 0 / MLT = 21 minutes / 3,233 minutes = 0.0065 or 0.65 percent
This result is not very impressive, but it is an accurate measure of the percentage of time each part is being processed, that is, having its value in-creased. A second problem is revealed if one studies the standard deviation of manufacturing lead time together with the minimum and maximum lead times. Note that there is a very high variation to lead time, including a maximum of more than 12,000 minutes. (These data were collected by measuring time from release until the batch, or its replacement, was completed. During 30 simulations of 90 days each, more than 19,000 batches were released. Of these, roughly 4 percent or 760 batches were rejected. Of the 760 replacement batches, roughly 4 percent of them also were rejected. Thus, about 30 batches were rejected twice. Of these 30, 4 percent were rejected, yielding an expected value of 1 batch rejected three times. The maximum of 12,000 minutes, which from Figure 6 occurred for only one batch, is consistent with an average lead time of 3,200 minutes times 4 batches required to finally have a batch accepted.) Although the number of batches having exceptionally Tong lead times is quite small relative to the total, to the individual customer whose batch is delayed inordinately, the delay may be a major problem. Each batch that is delayed a substantial length of time represents a potential lost customer.
Thus, the traditional job shop approach suffers from the fact that lead times are both long and highly variant. Notice especially the COUNT column on the right of Figure 6. The distribution of lead times is right skewed, that is, has a number of observations to the right of 'the mean that are quite far from the mean while no comparable events occur to the left of the mean. The existence of right skewness in the manufacturing lead time distribution suggests that a few jobs will consistently be very late. Furthermore, for most of the lead time, more than 99 percent in this case, the individual piece is merely sitting, waiting to be worked on.
Flow (Sequential) Process with Dedicated Lines: Now let's consider a JIT approach to this situation. Recall from the problem description that there arc four machines in each of the four work centers and that Part 1 is responsible for 50 percent of the demand and Part 2 is responsible for 25 percent of the demand. A JIT solution would be to move two machines out of each work center and set up two parallel lines dedicated to the production of Part 1. JIT would continue by moving one more machine from each work center to a third dedicated line, this one making only Part 2. The remaining four machines would produce the remaining three parts in a job shop fashion. The JIT approach immediately solves the problem of large batch size for Parts 1 and 2. Since both parts are manufactured continuously, no setup occurs. Manufacturing lead time is minimized because each part is passed to the next station as soon as it is completed. Figure 7 reports results from the three dedicated lines.
Note that lead time for Lines 1 and 2 averaged 62 minutes and that lead time for Line 3 averaged 107 minutes. VAE for Lines 1 and 2 is thus
VAE1 = 21 minutes / 62 minutes = 0.339 or 33.9 percent
VAE for Line 3 is
VAE3 = 21 minutes / 107 minutes = 0.196 or 19.6 percent
The VAE for Lines 1 and 2 is more than 50 times the former value of 0.65 percent. The VAE for Line 3 is about 30 times the old value. (The observant student may be wondering why Line 3 does not have the same manufacturing lead time as Lines 1 and 2 since the operation times are the same. The answer is that machine type A has the longest mean time to repair and, thus, is going to delay the part more, on average, than the other machines. For Part I, machine A has the shortest processing time of the four machines. For Part 2, machine A has the next to longest time. Thus, parts are delayed considerably longer at machine type A in Line 3 than in Lines 1 and 2.)
An interesting question is: Why does a dedicated line not achieve a VAE of 1? To achieve a VAE of 1, a part would have to be worked on 100 percent of the time, that is, move time must be zero and there must be no delay in a queue. In this model, arrivals of jobs to the line are constant, one every 7.2 minutes. If machine times were constant and if machines never broke down and if parts were never scrapped, a job would never wait. But the machines are down 5 percent of the time. This causes upstream machines to be blocked, since any queue wilt accept only five jobs. Also, because of scrap, a replacement part and a new part may arrive at the first station in line almost simultaneously, causing one to wait. Finally, machine times are random variables, so although average machine time may be only 5.5 minutes or 6 minutes for a given station, any single machine time may exceed the 7.2 minute cycle time.
Non dedicated Cell with Reduced Setup: Let's now consider what JIT would do with the remaining four machines and three parts. We should first note that setup times may be reduced for some machines by the mere fact that there are only three parts to be produced rather than five. For example, some machines have a tool magazine that will hold a fixed number of tools. Suppose the magazine holds 10 tools and that each of the five parts requires 3 distinct toots. Then for the original situation, the arrival of a new part might require that one or more tools be added to the tool magazine, requiring several minutes of setup time. However, with only three parts, the 9 tools may stay in the magazine all the time, and the setup is avoided.
The experience of many American firms that have adopted JIT is that about 75 percent of the setup time can be eliminated, without spending money to modify the equipment, by taking two steps. The first step is to separate tasks into internal task time and external task time, performing all external setup tasks which the machine is producing. The second step is to perform a methods improvement analysis on the setup and develop a standard setup methodology. (By traditional standards, setup time does not represent enough total time to warrant methods improvement analysis, so most setups have never been studied when a firm begins to implement JIT.) We assume each of the four setups is quickly reduced by 90 percent and, as a consequence, the batch size for the three jobs that continue to be produced in batches now is 10. The result of this simulation is presented in Figure 8.
Note that average manufacturing lead time now is approximately 1,183 minutes, and
VAE = 21 minutes / 1,183 minutes = 0.0178 or 1.78 percent
While this result is almost three times as large as the original value, it is somewhat disappointing, especially when one considers that setup time was reduced by 90 percent and batch size was reduced by 90 percent. This result emphasizes the value of getting the batch size down to I (at least the transfer batch, the number of parts required to be built at one station before parts are transferred to the next). One might also note from Figure 8 that the maximum manufacturing lead time is almost five times as large as the average time, emphasizing that the maximum lead time is strongly influenced by the scrap rate. (Recall that the maximum lead time typically occurs on an order that has a batch rejected, the replacement batch rejected, etc.)
Non dedicated Cell with Improved Quality and Reduced Setup: Figure 8 understates the effect on the manufacturing lead time of going from five parts to three parts using a job shop or non sequential flow method. The figure assumes that going to JIT has no effect on product quality. Let's consider the effect JIT has on how often a part is made. Parts 3 and 4 have a demand of 80 units each per day. Part 5 has a demand of 40 units. When these parts are made in batches of 100, the part is released to the shop about once each day or two. When there arc four machines in each work center, any single machine operator sees a part once every four to eight days on the average. When a worker processes a part only once a week, a learning process is required to regain top form in producing the part. This learning process almost certainly influences the scrap rate. When the batch size is reduced to 10, all three parts are built every day. Further, when only one machine is present in each work center, every machinist builds every part every day. No learning is needed to recall how to do something that is done every day for an extended period. Thus, from the learning effect alone, scrap should be reduced.
The use of standard die heights and locator pins to eliminate adjustments also improves quality. When die heights vary, there is a certain amount of guesswork concerning how a press should be set up to deliver maximum pressure to the desired point. In the absence of locator pins, there is also guesswork concerning the ideal positioning of the die on the press bed. The usual procedure is to use trial and error, expecting several defective parts to be made in the adjustment process. Once a good part is made, manufacture begins in earnest. Thus, the advantage of the JIT approach is that in addition to saving setup time, considerable scrap is eliminated almost immediately.
Let's examine the effect of reducing scrap from 4 percent to 2 percent, as shown in Figure 9 Note that both the maximum lead time and the mean lead time are reduced by about 25 percent. Maximum lead time is reduced because the likelihood of a job being rejected two or three times in a row is reduced substantially. (The chance of being rejected twice is reduced from 0.16 percent for 4 percent rejects to 0.04 percent for 2 percent rejects. The chance of being rejected three times is reduced from 0.00064 percent for a 4 percent rate to 0.00008 percent for a 2 percent rate.) The average is influenced quite a lot by extremely large values, Reducing the number 6f jobs that pass through the shop two or three times reduces the mean considerably. Another factor that reduces the mean is the fact that the smaller rejection rate also leads to slightly smaller queues and, hence, slightly smaller waiting time's throughout the entire facility.
The net effect of changing to a one-machine-per-station job shop handling three jobs with a scrap rate of 2 percent is the following:
VAE = 21 minutes / 775 minutes = 0.027 or 2.7 percent
This VAE is approximately four times the original 0.65 percent. Computing a weighted average for Lines 1 to 3 and the job shop yields
VAE = (0.5)(33.9%) + (O.25)(19.6%) + (0.25)(2.7%) = 22.5 %
Effect on Lead Time: Another way to express the benefit gained from moving to JIT is the reduction in lead time. In the original model, manufacturing lead time was 3,323 minutes or 55 hours. In the revised. model, lead time for the most popular part is 1 hour; for the next most popular, it is 2 hours; and for the last three parts, it is about 12.9 hours. There can be a large competitive advantage to promising delivery of a part in one or two hours, or at most one day, rather than promising delivery in three days.
The value of work in process inventory is greatly reduced in the Just-in Time example. Consider Figure 5, the original model. Examine the row marked "Queue Length." For Work Center A, this has a value of 0.615. This value is a time weighted value over the entire simulation. The simplest way that this value might arise is that 61.5 percent of the time there is a batch waiting to be processed, and 38.5 percent of the time there is not. There are also more complicated situations that achieve the same result. At any rate, since each batch represents 100 parts, there are 61.5 parts waiting to be processed, on the average, at Work Station A. Furthermore, there are an average of 2.348 operators busy at Work Station A. Each busy operator represents another 100 pieces. For each of the four stations, there is a positive value for queue length and busy servers. To determine total work in process, one must add queue length and busy servers for each of the four work centers and for material handling. These ten numbers add to 15.76 batches or 1,576 parts waiting to be processed or moved at any given time. By comparison, the JIT model results in an average work in process of 134 pieces, a 91.5 percent reduction in work in process.
Is our JIT example realistic? The Hewlett-Packard plant in Cupertino, California, actually did achieve a 94 percent reduction in work-in-process inventory when it implemented JIT (as did several others). Although an actual shop processes hundreds to thousands of parts rather than five, our example is not unrealistic. Typically, 5 percent of the parts account for some 60 percent of the product volume. Thus, most plants can move a few high volume parts into dedicated sequential flow lines. For the remaining parts, very often a number of medium volume parts have similar material content and part geometries (i.e., all are cylinders or all arc spheres or all are plates that require holes to be drilled, etc.). By moving all parts having material and geometry similarities into an area dedicated to only that family, much of the efficiency of a dedicated line can be achieved, There are several intermediate shop forms between the purely non sequential flow and the purely sequential flow. For example, a family of parts may share three or four operations in the same sequence. An area may be created to process sequentially a portion of the operations, while the remaining operations are performed non sequentially in another area of the plant. This arrangement provides the benefit of sequential flow for part of the routing, greatly reducing the lead time for the part and, hence, reducing the average WIP.
Taken all in all then, this example is quite realistic both in terms of the types of actions JIT would cause to be taken in moving from a pure non sequential process shop to a JIT shop and in terms of the magnitude of WIP and lead time reduction.
Uniform (Level) Flow
An objective of JIT is to have a smooth, relatively constant flow of work and material with a synchronized movement of small lots through the plant. However, a level load, day-in and day-out, is often difficult to achieve because of changing demand patterns, the mixed requirements that different level (smoothed) final assembly schedules place on upstream departments, and the necessity for freezing the final assembly and master production schedules for a month or so. Even when the load on the final assembly department is level, the corresponding load on the subassembly and parts fabrication departments may not be level, The following example from a major supplier of the automobile industry describes the objectives, benefits, and challenges of attaining uniform flow.
A major supplier of automobile sub frames receives a schedule, usually frozen for at least a month, for three subassemblies, The subassemblies arc manufactured (primarily welded) on a synchronized line that includes automatic welders, automatic movement, and a few manual operations. The representative data in Table 1 is used to describe the present operation, typical of many similar situations. At present the firm produces one lot of each of the three items each month, (They are working aggressively at implementing JIT concepts and have achieved substantial results in many parts of the plant.)
Because the assembly plant uses each of these parts at virtually a uniform rate, the subassembly manufacturer is carrying a little less than one-half month's inventory of each. The level loading policy shown in Table 2 eliminates nearly all this inventory and the resulting expense.
Changeover (setup) time usually is the impediment to producing the same amount of each item each day-even when only three items arc involved. When it takes five hours to change production from one item to another, manufacturing each item each day would require three shifts to produce the output they now accomplish in one. Clearly the challenge is to improve the design of the part, the line, and the changeover process to reduce setup requirements. The first objective is to make the improvements necessary to halve the lot sizes, especially for F-11 and F-12. The beauty of JIT is that it points out the inefficiency of the present changeover requirements.
Pull Production Control
Push systems are the traditional method of controlling orders and material in a plant. When an order is completed at one work center it is sent (pushed) to the work center where the next operation will be performed. A push system assumes the next work center will be ready to process the order. In a pull system, the parts are not forwarded to the next operation until they are requested.
A pull system works well in a sequential flow process environment because the source of incoming material is always the same work center. As variations in the process flow increase, implementation of a pull system becomes more difficult. For example, if the material at a work center arrives from a limited number of work centers (say two or three) in a constant pattern, a pull system may work effectively. However, if the master production schedule and the resulting schedules for individual items call for the arrival of parts from many different work centers in an irregular pattern, a JIT pull system usually is not feasible.
Many different methods may be used in a pull system to authorize a supplying work center to send parts. Returning empty containers is a common method; other methods include cards and tokens of various types. Another method requires the supplying department to observe the status of inventory in the receiving department and to forward material whenever the incoming material in the receiving department is reduced to a specified level. Using this method, the receiving department has specific shelves or locations on the shop floor that are designated for incoming materials and are clearly visible to the supplying department. Thus, empty shelves or empty floor spaces authorize the movement of material.
The objectives of a pull system are to:
A. Synchronize the movement of material throughout the manufacturing and distribution system at the rate of withdrawal of material from the system.
B. Limit the total inventory in the system.
C. Facilitate analysis, process improvements, and further reductions in inventory.
Because the Kanban pull system of Toyota is the best known pull system, pull systems are sometimes called Kanban systems.
Let's consider an example of a simple pull system that exists between a department manufacturing components and the assembly line it feeds. Standard containers, each holding 10 items, are used to move components to the final assembly line. As the last of 10 components in a container is used, the container is returned to the parts fabrication department where its arrival authorizes the fabrication of another 10 components. On completion of the 10 components, the container is sent immediately to the assembly department. The advantages of this type of system are apparent. The upstream (supplying) department cannot flood the assembly department with unneeded parts. Furthermore, if the need at the assembly department increases, it becomes apparent immediately to the supplying department because the containers are returned sooner and at a faster rate.
The number of containers required for the system to function properly between two departments depends on the demand rate (the production rate of the assembly line in this case), the movement time between the two departments, the time the container waits to be moved, and the processing time. This relationship is represented by the following model.
N
>
=D (M+P)(l.0+S) / Q where
N = an integer, the number of movement cards (containers) required
D = demand per hour (the rate at which the user department requires the parts)
M = average wait time (wait time includes the processing time at the user department) and move time required. Thus, M is the total round trip time from the source (parts producing work center) to the user (assembly line) and back.
P = average setup, run, and inspection time required to manufacture the parts in a container
S = the safety factor, expressed as a percentage to compensate for varying rates of production and the efficiency of the producing department
Q = quantity of parts held by each container
In this example the demand rate of the assembly department is 20 parts per hour; the move time is 15 minutes (0.25 hours); the assembly time is 30 minutes (0.50 hours); the total processing time for a container of parts in the parts fabrication department is 24 minutes (0.40 hours); the safety factor is 0.05; and there are 10 parts in each container. The number of cards is then calculated as follows:
N = 20(0.75 + 0.40)(l.0 +0.05) / 10 = 2.415 or 3 containers
A Gantt chart analysis of the movement of containers in this situation will reveal that three containers is the minimum that can be used to keep the assembly department operating at all times. This is based on the containers not being returned until all the incoming parts have been used in the assembly process. Since the safety factor used to estimate the number of Kanbans cannot be determined with precision, in reality the number of Kanbans arc determined by trial and error.
HUMAN RESOURCE MANAGEMENT
If there is a single key to attaining JIT objectives, it is a genuine respect for fellow human beings, for their aspirations, capabilities, and integrity. This requirement is the foundation of the recommendations concerning the treatment of customers, employees, and suppliers.
It is not surprising that changes in this area present the greatest challenges to Western companies. This is especially true for companies in which the work force and management have a long-standing adversarial relationship. Adopting JIT requires that all personnel--management, staff, and labor-must perceive changes as enhancing their personal goals as well as the goals of the organization. Employees must be confidant that improvements will not jeopardize their employment and that they will share in the resulting benefits.
The following are critical in winning the trust, participation, and wholehearted support of all employees:
A. Employees must be convinced that the improvements they suggest will not result in their unemployment.
B. Orientation, education, and training programs must exist so that employees understand the objectives and policies of the company and the rationale of related programs. Furthermore, they must be given the opportunity to increase their skills and to participate more fully in the improvement activities.
C. Employees must be given more responsibility as decision making is driven downward in the organizational structure,
D. Formal procedures must be developed for tapping the experience and knowledge of all employees through improvement suggestions. A system must exist for evaluating suggestions quickly and rewarding them fairly.
E. Employees must be united, team spirit must be developed, and performance evaluations and rewards must be based on the performance of functional groups and the whole organization. In many cases this requires the development of a new organizational culture. Developing a new and pervasive culture takes commitment, leadership, patience, and time.
Companies with seasonal demand and a tradition of seasonal production and layoffs face a substantial challenge. Developing a flexible work force capable of performing equipment and plant maintenance during normally dormant demand periods can reduce the need for layoffs. Dormant periods can also be profitably used for employee education and training. The Sunnen Corporation of St. Louis has followed this approach successfully for many decades.
Quality circle programs, productivity improvement programs, and other similar programs have been used successfully by many companies. Such programs have a twofold effect. First, they generate cost savings and profit improvements. Equally important, they go a long way in convincing employees that management does value employee ideas and competence. These programs, combined with opportunities for skill development, are an indication of an organization's true respect for its employees. Most employees respond positively when they perceive themselves as being treated as important in a company's operation. A few isolated and fragmented projects have little value; a continuing and ever-present management attitude supporting teamwork and employee participation can work wonders. Case studies of JIT implementations are contained in Crawford, Cox, Blackstone (1988) and Sepehri (1986).
JIT: PURCHASING & SUPPLIERS
Developing Long-Term Relationships
The relationship between purchaser and supplier often has been one of mutual suspicion. The phrase caveat emptor (buyer beware) was the watchword of the purchaser. On the other hand, suppliers were often treated unfairly. For example, a supplier might provide excellent products on time and at a reasonable cost only to lose the next order due to a 10 cent lower bid by a competitor.
Under JIT, emphasis is on developing long-term relationships with suppliers. The relationship is based on mutual trust with quality the main objective. The supplier, and the supplier's supplier, are viewed as links in the industrial chain meeting the needs of the customer. If any link fails to perform satisfactorily, the final product is unsatisfactory and the. entire system fails. The objective is to reduce costs and increase quality and productivity by:
1. Involving the supplier in the product design effort in order to take advantage of the supplier's specific competencies
2. Reducing the number of suppliers and the continual bidding process
3. Increasing the technical support provided to suppliers
4. Providing the schedule of order releases in a time frame that encourages the supplier to commit resources to improving quality, delivery, and cost
5. Lowering costs through the increased learning curve effects that result from the long-term relationship
6. Increasing communications through electronic transmission of engineering changes and delivery schedules
7. Locating suppliers nearby to reduce average delivery times and their variance, to increase frequency of deliveries (daily if possible), and to reduce lot sizes
8. Aiding suppliers in establishing statistical process control to improve quality
9. Reducing inspection requirements as quality levels increase
10. Improving product design through the supplier's innovations that result from increased experience and commitment
11. Increasing detection and correction of defects through the supplier's frequent deliveries
12. Using standard containers and simplifying the count of incoming parts
13. Obtaining dollar volume discounts through larger purchase commitments
These improvements do not occur overnight or automatically; they result from consistent and painstaking analyses with a sprinkling of failures and false starts.
TOTAL QUALITY MANAGEMENT
The conventional wisdom concerning quality has been that as the quality of a product was increased, the cost of manufacturing it increased exponentially. This contention was debunked by Phil Crosby (1979) who is credited with the statement, "Quality is free." Dave Garwood (1988), among others, illustrates this graphically by what is known as the “Mount Fuji Effect." (See Figure 10.) According to this concept, the cost of quality increases up to a point, the top of the mountain, and then the net costs decrease dramatically-they go down the other side of the mountain-as the benefits of good quality exceed the costs. The point is that conventional wisdom overlooked both the costs of poor quality and the value of good quality, as given in Table 3.
Total quality management (TQM) differs from traditional quality control in a number of other important ways. They are:
A. Quality begins at the source, namely the product design, the design of the manufacturing process, and the supplier in the case of purchase parts.
B. Operators arc responsible for the quality of their output. Inspection is not left until the batch is completed or the product completes the final operation.
C. Statistical process control (SPC) is used to spot trends in output that presage tool wear or equipment requiring adjustment before unacceptable output is produced.
D. A supplier's processes and its statistical process control system is evaluated before the supplier is certified as an approved supplier.
E. Preventive maintenance, based on statistical analysis of past performance and output characteristics, is emphasized.
Some of these approaches have existed in many companies for some time. It is the synergistic effects of all of them combined with a corporate culture giving top priority to quality that makes the difference. Certain methods including SPC, stop production authority, mistake-proofing, preventive maintenance, and fishbone analysis, are used to achieve the goals of TQM.
Statistical process control (SPC) is based on a prior analysis that deter mines those process characteristics that are critical in producing a quality product. Measurement and analysis of these process characteristics in relationship to their acceptable averages and variances alerts management to the likelihood of out-of-control processing conditions before they occur.
Stop production authority (jidoka) permits an employee to stop the production process, an entire line if appropriate, when a quality defect occurs. The problem is analyzed immediately, with the assistance of manufacturing and equipment engineers if necessary. The objective is to eliminate quality problems before they multiply and before additional work is performed at downstream work centers on parts that are already scrap. It is consistent with the attitude that quality is more important than output quantity and that bad parts arc unacceptable.
Mistake-proofing (poka-yoke) is targeted at the process design stage in an attempt to eradicate mistakes. There may be limit switches that automatically prevent a machine from moving too far in a given direction or other devices that physically prevent the mistake from occurring.
Cause and effect analysis is an excellent method for analyzing defects and for educating employees concerning the importance of different product and process design characteristics. This approach uses a schematic diagram resembling a fishbone, and thus, it is often called fishbone analysis. In Figure 11, the bone structure represents the hierarchical relationship of contributing causes to the specific quality characteristics of a release product. A release product is any label or sticker which is sold fixed to a carrier, but which is pulled from the carrier and stuck to another object. Preprinted price stickers or mailing labels are examples of release products. Quality control for the release product is critical because the label must have a weak enough adhesive (not very sticky) to pull from the carrier (backing paper) without tearing the label, yet be strong enough to adhere to the recipient object surface. This measure of stickiness is called the release value.
The fishbone diagram shows four principle contributors to defective release values: strength of the label paper, strength of the release adhesive, coating of the carrier paper, and surface of the recipient object (i.e., envelope or paper on which the label is placed). Once the quality problem has been isolated to one of the four principle contributors to defective release values, possible secondary reasons are pursued by the analysis. For example, if the strength of the release adhesive were found faulty, secondary reasons would be evaluated, including: humidity, temperature, speed of application rollers, chemical formulation of the adhesive, electron dryer mechanism, and consistency of the mixture. The fishbone diagram is very useful in quality diagnosis of complex processes because it encourages a logical analysis to sequentially isolate quality problems.
JIT AND COSTS
JIT can affect the bottom line in a variety of ways. Improvement in quality and delivery times can increase demand and, thus, revenue. Costs are also affected; the JIT philosophy contends that inventory reduction and increased quality reduce costs. Traditional cost accounting Systems often make it difficult to measure the effects of changes except in very aggregate terms. One of the tenets of JIT is to account for these effects more accurately.
Cost Accounting Systems
Costs are a major factor in PIM decisions. Unfortunately, traditional cost accounting Systems often do not tell the decision maker how much a specific decision wilt affect actual expenditures. This is due to overhead costs being hidden by the allocation methods. For example, overhead costs usually are allocated to departments (cost centers) rather than to activities, such as setup, and inspection and maintenance operations. In addition, allocation based on the material or direct labor required to manufacture an item ignores the fact
that different items are in different stages of their life cycles. Thus, different items may have different manufacturing, engineering, and tooting costs, may have quite different quality and inspection requirements, and may require different marketing and distribution expenditures. When these costs are aggregated and allocated on the basis of the average direct labor cost of a part-as is the case with most traditional cost accounting systems-some products are allocated costs considerably below the actual expenditures required for their manufacture and distribution and others are allocated more than their true cost. Thus, decisions often are based on inaccurate information.
In order to manage costs and base decisions on accurate information, the causes (source) of the expenditures must be identified. Various expenditure causes; such as setup times, shop and purchase order processing, receiving, and material handling deserve more discussion. These basic causes of indirect costs are called cost drivers. The cost accounting system must report the cost of these activities to accurately determine the costs of individual products. Such reporting enables manufacturing management to treat setup, inspection, receiving, and transaction costs as direct costs, to base decisions on accurate information, and to focus on reducing high cost elements. An ABC analysis, can be used to select the activities that are appropriate for cost reduction studies.
The Correlation Between Cost and Quality
The JIT philosophy contends that there is an inverse relationship between quality and cost: as quality increases, cost declines. While some find this belief enigmatic, a growing number agree with this notion. Garvin (1988) notes that quality has eight dimensions. He posits that some aspects of quality are inversely related to cost while others are directly related to cost. People disagree in their definition of quality rather than in the relationship between cost and quality.
Garvin lists the eight dimensions of quality as:
1. Performance
2. Features
3. Reliability
4. Conformance (to specifications)
5. Durability
6. Serviceability
7. Aesthetics
8. Perceived quality
Garvin found empirical evidence that reliability and conformance are inversely related to cost. He found no empirical evidence supporting either an inverse or a direct relationship between cost and any of the other six quality dimensions. Garvin argues that the Japanese (JIT) viewpoint arises from a strong emphasis on reliability and conformance in defining quality.
Reliability---the mean time to first failure (and for repairable items, the mean time between failures)---heavily determines warranty cost and product liability cost. Conformance is the inverse of scrap and rework percent, a measure of ''doing it right the first time.'' It is not surprising that Garvin found evidence supporting inverse relationships between cost and reliability and between cost and conformance.
However, Garvin's thesis that the inverse relationship between cost and quality extends only to reliability and conformance ignores the JIT emphasis on designing for ease of manufacture. Designing for ease of manufacture is intended to produce a high quality, low cost product by (1) reducing options and (2) avoiding requirements beyond the capability of available equipment.
Reducing the number of options reduces cost by reducing inventory. For this strategy to work, the model choices must each please a large number of consumers. Pleasing consumers requires listening carefully-a Japanese design engineer spends much more time with customers than does an American designer. Once the most popular options have been determined, they can be made standard features at a much lower cost than the cost of providing for each as options. The Toyota Cressida at one time sold with a standard luxury package---leather seats, impressive sound system, maximum use of electronics, etc.,---for thousands of dollars less than a comparably equipped American or German luxury car.
A very strong argument can be made that by designing for ease of manufacture, there can be an inverse relation between cost and features, aesthetics, and perceived quality. Because design influences durability, there may also be an inverse relationship between cost and durability (largely offset by a higher material cost to build in durability). Because only a few firms today truly design for ease of manufacture, it is not surprising that Garvin was unable to find empirical support for this relationship.
High quality, low cost items will be essential if a company is to compete in future world markets. These items require a strong emphasis on product design and manufacturing process design and a high degree of cooperation between product design and process design teams. The product design must emphasize performance, features, aesthetics, serviceability, durability, and perceived quality, while enabling inexpensive manufacture. The process design must emphasize conformance and, hence, reliability, durability, and perceived quality.
Low manufacturing cost also requires the realization of economies of scale-volume production. To achieve world class production volume requires worldwide markets. Unfortunately, many American firms do not emphasize export and lack the skills needed to customize products for foreign markets. Manufacturing cannot propel a company to world class status by itself-dramatic change is also required in marketing. But that's another topic.
PERFORMANCE MEASUREMENT
In a survey of JIT implementation to identify problems, Crawford, Blackstone, and Cox (1988) found that the largest single problem was failure to change to an appropriate performance measurement system before JIT was introduced to the shop floor. In a follow-up study, Crawford, Cox, and Blackstone (1988) identified appropriate JIT performance criteria as follows:
Raw Materials: Inventory dollar days, raw material stock outs, raw material reduction, vendor delivery, vendor quality
Equipment: Machine breakdowns, preventive maintenance, setup reduction
Facility: Space requirements
Employee: Morale, education and training acquired, labor effectiveness
End item: Cost of goods sold, customer service, schedule flexibility, inventory dollar days, inventory reduction, lead time, output per employee, scrap, rework
Transformation: Cycle efficiency, process improvement, lot-size reduction, material stock outs, WIP reduction
Most companies reported the use of relative measures rather than absolute measures, i.e., what mattered was the improvement and the trend, not the level that existed when the improvement process began. Most companies displayed results graphically in the work center area so that workers could take pride in their accomplishments.
Crawford, Cox, and Blackstone report the following performance measure system principles:
1. The performance measurement system should have multiple criteria.
2. The primary purpose of the performance measurement system should not be to reward or to punish.
3. Performance-to-schedule measures must use group, not individual, results.
4. Specific goals must be established for performance-to-schedule criteria and must be revised when met.
5. Specific goals are not necessary for inventory and quality criteria; improving trends are needed.
6. Performance measures must be understood by those who are being evaluated.
7. Performance data should be collected by the person being evaluated.
8. Graphs should be the primary reporting method.
9. Performance data should be available for constant review.
10. Schedule performance should be reported daily.
11. Inventory and quality performance should be reported monthly.
12. The performance system must include frequent performance review sessions.
13. Suppliers should be evaluated on quality and delivery.
IMPLEMENTATION of JIT
Implementation of JIT involves six phases: organization, education, evaluation, planning, execution, and review. A prerequisite to success is top management's long-term commitment. Employees quickly sense when management is half-hearted or not fully convinced of the ultimate benefits and will quickly relegate JIT to the burial grounds that hold many previous “panaceas of the month."
Organization
A broadly based steering committee should be formed with representation from purchasing, design and manufacturing engineering, manufacturing management, production control, industrial engineering, quality control, maintenance, and operations. The leader should be the champion of change and have an understanding of the requirements for, and preferably some experience in, implementing change. The members should possess a certain discontent with the present yet be able to express this discontent and support change in a constructive manner. In addition, including a facilitator, often someone from outside the firm, aids in providing a broader frame of reference and overcoming those blind spots that naturally develop in most firms due to long-accepted ways of operating.
Education
The development of knowledge, understanding, confidence, and trust throughout the organization begins in the organization phase and is solidified in the education phase. Although education is continual in a JIT mode of operation, it is most intense and crucial in the beginning. It should begin with top management and cover virtually everyone in the organization. JIT must be understood and appreciated throughout the organization to achieve its full benefits. Nothing will work on the plant floor unless the workers are convinced of its benefits; and it will not reach the plant floor if staff and middle level managers do not support it. First, the education should cover the basic objectives and philosophy of JIT and its importance to all employees (their livelihood and development). It should stress that JIT is not a "microwave" program; benefits do not come overnight. Patience is required. It is an evolution not a revolution, and not all changes will be success stories. Education should also encompass basic concepts concerning such areas as the importance of the customer, quality, the cost of inventory, lead time, and productivity.
After the initial education program, specific and focused education training programs concerning topics such as setup reduction, working with suppliers, statistical process control, and group technology are appropriate.
Evaluation and Assessment
Because organizations have different environments and are at different stages in developing their manufacturing activities, each should make a thorough assessment of its environment, decide on its strategic objectives relative to JIT, and evaluate its present status relative to the major operating objectives of JIT. Assessment of the present status is a prerequisite for deciding the priority of proposed improvement activities. Ken McGuire (1984) recommends that three teams perform independent assessments. The reconciliation of these assessments then provides a consensus final assessment that serves as a basis for selecting initial JIT activities.
Figure 12 is an adaptation of the chart McGuire proposes for rating and ranking areas for improvement. It assesses each area on the basis of its importance relative to the success of the firm, the current operating effectiveness of the area, the resources required to improve it substantially, and the time required to complete the improvement. This method of assessment is similar to that often used in establishing the priority of information system improvement projects. Clearly, an area that is critical to the success of the firm, whose present performance is inadequate, and which can be improved with little investment and in a relatively short time would likely be attacked first. Although it would seem that there are few such opportunities, there are usually at least a few substantial and visible improvements that can be made. For example, Miller Fluid Power of Bensonville, Illinois, reorganized the assembly area for one of its product lines by dedicating specific lines to assemblies that require many common parts and that use the same assembly fixtures and by stocking the appropriate parts next to each line. These changes in layout, allocation of lines, and stocking locations reduced lead time from approximately two weeks to one day, reduced inventory levels, and released space for other uses. On the other hand, A. 0. Smith in Granite City has heavy presses that cannot be relocated easily. In this case, provisions were made for rapid movement of smaller transfer batches to the next manufacturing operation. Manufacturing cells with cross training of workers were established in other areas. The point is that each firm must work within the constraints and opportunities of its environment.
Outside assistance is often very helpful in providing perspective and objectivity and in gaining a consensus concerning the present status and the priority of different potential actions. The objective is to begin with low risk, high yield tasks-those that have a high probability of success.
The Plan
The initial plan begins by obtaining the commitment of top management and is followed closely by the introductory education programs for the entire work force. Different programs may be appropriate for. different groups. Evaluation and assessment provide the basis for developing the plan for the initial improvement activities.
A strong case can be made for stressing quality improvement early in a JIT program (Hall 1983). An analysis of quality requires a study of the customers' requirements. Improved quality inherently reduces inventory requirements, reduces scheduling problems, and improves personnel and equipment capability. Good housekeeping should be considered a prerequisite, and improvements should be initiated post haste, if required.
(Click Here to view Figure 12, JIT Assessment Chart)
Most people respond to performance measures: Students want to know what the exam will cover, and production personnel act to achieve good scores on their performance measures. Early replacement of inappropriate performance measures is essential. For example, if a manager's performance is measured by output volume alone rather than by completion of the right products (orders) at the right time, the manager will be hard pressed to take JIT seriously. Department heads and workers, who for years have been encouraged to keep equipment utilization high and output volume high and have been measured on that basis, will continue to produce unneeded parts at non bottleneck work centers unless their performance is measured in terms of improved processes, reduced setup times, reduced work in process, reduced lead times, improved quality, and an improving percentage of deliveries right on schedule.
RESULTS
The following results have been reported by American companies applying JIT to American plants:
The Apple Macintosh factory, is months into JIT, reported that rejects were reduced from 28 percent to 1 percent, inventory turns were twice the industry average, space requirements were reduced 35 percent, labor productivity was increased 60 percent (Sepehri 1986).
Omark Industries, in the first year, reduced inventory 25 percent ($20 million), increased productivity 30 percent, reduced lot sizes, shortened lead times, and improved quality. Later into the program, raw material was reduced 95 percent and WIP 96 percent. In the case of WIP, the reduction was from 100,000 pieces on the floor at any given time to 4,000 pieces, with an eventual goal of 1,000. The consequences of this reduction to material scheduling and control are that material can be much more tightly controlled. (Sepehri 1986)
Harley-Davidson reports a 50 percent inventory reduction, a 50 percent reduction in scrap and rework, a 32 percent productivity increase, an increase in inventory turns from 5 to 17, and a decrease in warranty claims despite a longer warranty period (Sepehri 1986).
IBM's plant in Raleigh, North Carolina, which makes terminals for mainframe computers, while not reporting numeric results did report that manufacturing costs were greatly reduced, inventory turns increased, mean time between failures was reduced, and cycle time from product inception to customer availability was reduced (Sepehri 1986).
JIT's applicability is not limited to discrete parts manufacturers or to large companies. ChemLink, a small petroleum processor, reports that inventory was reduced by 21 percent, sales grew by 9 percent, obsolete inventory was reduced by 30 percent, and transportation cost was reduced 8 percent (Crane 1989).
Hay (1988) estimates the range of improvement possible for a western JIT implementation to be about 83 to 92 percent reduction of lead time, 5 to 5O per-cent reduction of direct labor, 21 to 60 percent reduction of indirect labor, 26 to 63 percent reduction in the cost of poor quality, 6 to 45 percent decrease in purchased material costs, 35 to 73 percent reduction in purchased materials, 70 to 89 percent reduction in work in process, 0 to 90 percent reduction in finished goods inventory, 75 to 94 percent reduction in setup time, and 39 to 80 percent reduction in space requirements.
JIT presentations often employ the analogy of a stream when describing proper inventory management. Well managed systems achieve a flow of inventory from raw material to the customer like a smooth river, unimpeded by shoals of scrap or machine breakdown or other problems. This concept did not originate with the Japanese; Henry Ford's River Rouge plant regularly converted iron ore into a Model T in 4 days. However, in recent years, especially the 1970's American Business has not improved its manufacturing capability quickly enough to maintain a competitive position in cost or quality or market responsiveness or flexibility.
Back...
Keeping Promises...
Overcoming Resistance...
Turning Communication Into a System...
Identifying the Heart of What Really Matters...ERP-MRP Evolution…
ERP & Hoshin Kanri…
ERP Implementations…
Profit-Ability Improvement... (¬Click here to see definitions)
Profit-Ability Management Principles... (¬Click here to see definitions)
People, Empowerment & Profit-Ability… (¬Click here to access articles)
Hoshin Kanri & Deming's Plan-Do-Check-Act... (PDCA) Cycle…
