Create Initial Process Stability

First Get to Basic Stability

If you have not already been using lean methods and improving processes, in all likelihood your individual processes are unstable. Stability is defined as the capability to produce consistent results over time. Instability is the result of variability in your process. It could be that equipment is not well maintained and so breaks down regularly. It could be that for any number of reasons defects are regularly produced. Or perhaps there is no standard work, and the amount of time it takes to perform a given process varies tremendously from person to person, across shifts, or over time.

The first step in creating lean processes is to achieve a basic level of process stability. The primary objective in developing stable processes is to reach a consistent level of capability. Based on the spiral model of continuous improvement presented in the last chapter, there are increasing degrees of stability. The initial level of stability is generally defined as the capability to produce consistent results some minimum percentage of the time. This is measured based on the outcome and is related to producing the same quantity of products, with the same amount of resource time (people and equipment), with a high degree of reliability (the exact level may vary according to the process and conditions, but a reasonable rule of thumb is 80 percent or better). A simpler indicator would be the ability to meet the customer requirements with quality products the first time through on time (again, 80 percent or better). In many cases the “customer requirement” is not clearly defined and becomes one of the first tasks of the stability phase.

Indicators of Instability

There is a wide belief that stability is indicated mainly by equipment performance. As a result, the pursuit of certain lean tools—like “quick changeover”—and attacking equipment failures through preventative maintenance become primary activities. Developing process stability is not an end onto itself. In fact, it’s more about creating a foundation for further aspects of a lean process. Through direct observation, an unstable process is indicated by the following conditions:

• A high degree of variation in performance measures—either pieces produced or pieces per labor hour.
• Changing the “plan” often when a problem occurs. This includes relocating labor or leaving a position vacant when an absence occurs, moving product to another machine when a breakdown occurs (and thus not producing the planned product), and stopping work in the middle of an order to change to another order.
• It is not possible to observe a consistent pattern or method to the work.
• Batches or piles of work in process (WIP) that are random—sometimes more, sometimes less.
• Sequential operations that operate independently (island processes).
• Inconsistent or nonexistent flow (also indicated by random WIP piles).
• Frequent use of the words usually, basically, normally, typically, generally, most of the time, when describing the operation, followed by except when, as in: “Normally we do this . . . except when . . . happens, then we do this. ”
  (By its very nature, an unstable operation does not often experience “normal” in terms of consistent method. In fact, the abnormal becomes the normal.)
• Statements such as, “We trust the operators to make decisions about how the work is done” (part of a misguided application of employee empowerment).

It’s important to realize that no operation will ever achieve a perfect level of stability, and thus to some degree these conditions will always exist. In fact, stability is not only a requirement for flow, but developing flow helps motivate disciplined approaches to stability—they go hand in hand. The main consideration is how unstable the process is, and how stable it needs to become in order to move into the next phase of achieving some degree of flow. Based on the spiral model of continuous improvement, during the incremental leveling phase the operation will be “squeezed” and a higher level of stability will be necessary to meet the tighter requirements. This, in turn, will force a refinement in the methods, beginning a new turn around the spiral in ever tightening cycles.

Clearing the Clouds

The Japanese are prone to using metaphors to describe situations. Toyota Production System (TPS) masters often refer to “clearing the clouds” when discussing the initial creation of a lean process. This was often compared to a photograph that was cloudy or unclear. Many issues often cloud processes that have not achieved a basic level of stability. They may or may not be truly related to the process; however, the cloudiness makes it difficult to determine this. Most important, these “clouds” obscure our view, and our ability to see and understand the true underlying image. On more than one occasion a TPS trainer was seen circling his hand, palm down, around his head and muttering, “Very confused,” indicating the effect of many obfuscating issues. Upon initially observing an operation, it’s easy to confuse the activity seen with beneficial or necessary (value adding) work. People are busy, they’re moving quickly, they’re “doing” things, and it can be challenging to ascertain what the underlying true image should be.

Processes fraught with randomness and chaos tend to lead us to incorrect conclusions about what is real, what is possible and what’s not. The ability to adapt to surrounding conditions is a human characteristic necessary for survival, which is fortunate, and yet it makes creating lean processes especially difficult.

By our nature, we adapt to our surrounding conditions and within a short time come to accept them as “normal” and no longer give them consideration. In many cases we even come to consider these conditions part of what we “have to do.” Fortunately, we can be shaken from this paradigm, and when the situation is considered from a different angle, understanding is developed. Utilization of the lean philosophies and tools will force us to take a fresh look from a different perspective, and if we allow our minds to accept the new information, real transformation can occur. And then—human nature again— once the transformation occurs and we become accustomed to the new condition, it may never occur to us to reevaluate again and to seek another level. This is the challenge of continuous improvement. Diligent application of the lean transformation spiral model will force continual evaluation and removal of another layer of cloudiness in pursuit of the underlying crystal clear image.

Objectives of Stability

The primary objective of the stability phase is to create a basis for consistency so the “reality” can be seen and random activities removed, thus establishing a foundation for true improvement. This includes reducing the variability of the demand rate (prior to the establishment of takt time, rate of customer demand) and the creation of basic daily volume leveling. Additionally, each phase in the continuous improvement spiral provides necessary preparation for the development of succeeding phases. Thus, the stability phase is crucial for the preparation of the flow phase. Major impediments to flow must be targeted and removed. If connected flow is attempted prior to achieving stability, the impediments may be too large and the creation of smooth, consistent flow will be impossible. A stable process will also have a higher degree of flexibility and capability of meeting varied customer requirements.

TRAP

The Fallacy of Perfect Stability
We were involved as consultants in the early day of the implementation of the Ford Production System, modeled after the Toyota Production System. There was general agreement on the importance of process stability before moving to the higher levels of lean. There was also a strong belief that all plants around the world (over 130) had to move forward in roughly the same time frame. So the first year was spent on process stability issues in one model area that each plant selected, including 5S (see page 64), preventive maintenance, and standardized work. The first year extended into year two. It became clear that these seemingly simple tools required a great deal of discipline and understanding and the plants had little incentive other than “corporate wants us to do it and is going to check up on it.” In later years this moved at Ford to a more integrated approach where flow, pull, and stability were better integrated in model areas. Process stability should have a reason—to support value-added flow. Reducing waste and creating flow will make stability a necessity instead of a necessary evil to please the corporate lean group.

On the other hand, it’s possible to spend years trying to achieve perfect stability without moving to higher levels of flow and pull. Experience suggests that this will lead to cycles of stability: dropping back to instability, reattaining stability, and on and on. The reason is that there is no motivation to sustain the higher levels of capability because the system is not tightened to require the improved level. In a large batch operation without flow, a high level of stability is actually not needed and thus the only motivation to continue using disciplined process is to keep “lean managers” off your back.

Strategies to Create Stability

Table 4-1 shows the strategies utilized during the stability phase, as well as the primary and secondary lean tools often utilized. Any particular tool may or may not be used, depending on the circumstances of the operation. The objectives and strategies, however, always apply.

Table 4-1. Strategies and Tools Used in the Stability Phase

As we noted earlier, it is not our intent for this to be a “how to implement lean tools” book. There are already numerous books filled with excellent descriptions of each of these tools. Our objective is to focus on the philosophy and an understanding of the process.

Identify and Eliminate Large Waste

As mentioned previously, the identification and elimination of waste is a primary philosophy of lean. If this is a virgin site for lean, there’s a lot of low hanging fruit. For example, simply using 5S to label where inventory should be held and setting visible maximum and minimum levels can have a large impact. Standardized work and 5S can significantly improve manual operations. Improvements in equipment uptime and reductions in lost time by reducing changeover times will add capacity and improve process throughput.

Removing the first, large layer of waste generally yields significant improvements in overall performance. At this point most of the improvements are at the individual process level, not at the level of flow-connecting processes. Subsequent cycles through the continuous improvement spiral will connect processes and can have even larger impacts, and reinforce motivation to maintain the stability of individual processes.

Standing in the Circle Exercise

Learning to identify the seven types of waste begins immediately during the stability phase and is reinforced by “standing in the circle,” the exercise used by Taiichi Ohno to train new members. This is part of the philosophy of genchi genbutsu, which emphasizes going to the actual place to observe and understand. During this exercise, the member is directed to stand and observe an operation carefully, and to identify the waste within the operation and the conditions that cause the waste to exist. Members are often left standing for 8 hours or more before the sensei is satisfied that they have seriously seen the waste. Ironically, this is even harder to do when you’re already familiar with the operation. Because you understand the “reason” that the waste exists, you will be inclined to rationalize its existence (why it is that way) and to conclude that nothing can be done to improve it. During the circle exercise it is best to simply acknowledge that the waste exists, without the need to explain it or to try to figure out how to “fix” it.

If the exercise is taken seriously, the amount of waste observed can be over-whelming. A common reaction is to immediately seek out solutions to remedy the situation. In Part IV, which deals with problem solving, we explain that the first step is to develop a thorough understanding of the situation prior to beginning corrective action. Standing in the circle for many hours will allow a thorough understanding, which is necessary before any true countermeasures can be identified.

The circle exercise may be likened to a distance race, such as the marathon. (Though we have never run anything close, we all know people who have.) About 20 miles into the 26.2 mile race, runners describe a sensation known as “hitting the wall.” Some have described the physical aspect as a sort of “transcendence” of the body. The circle exercise is similar in nature. During the first few minutes to an hour, the mind is observing the larger issues and capturing the “big picture” and might conclude that everything has been seen and there is no need to continue. Stand some more! The real learning is just beginning. Depending on the individual, it may take four to eight hours before “hitting the wall” and transcending to a higher level of awareness. This is an extremely powerful exercise. Do not view it as merely “standing around.” Rather, use it as a method of “tuning” your awareness skill. Once this skill is mastered, a shorter observation will provide a clear understanding of the details of an operation. Fortunately, it will not require eight hours every time!

Standardized Work as a Tool to Identify and Eliminate Waste

After you have mastered the ability to observe and identify waste, it is possible to document the situation using the standardized work tools. Often standardized work is thought to be mainly a set of instructions for the operator. In reality one of the most powerful uses of standardized work is for analyzing and understanding waste in the operation. The documented work procedure will be a visual representation of the waste (opportunity for improvement) that exists. It is part of the analysis that helps to remove the “clouds” and see the underlying image. It will also provide beneficial information for establishing balanced work flow during the creation of continuous flow.

In Chapter 6 we will provide greater detail on standardized work and how it’s used to establish and document the standard method, but at this phase suffice it to say that the tools are simply used to aid in identifying waste. There are three critical elements in analyzing the work and identifying waste during the stability phase:

1. Identify the basic work
2. Record the time for each
3. Draw a picture of the work area and the operator’s flow within the

Bear in mind that the intent is to identify waste, and it is important to start with the “big” waste first. As an analysis tool, standardized work will primarily aid in the identification of motion (walking, reaching) and waiting (when the work cycle is below takt rate). It’s best to first analyze from a higher level and then work down to a detailed level. If the work requires the operator to walk out of the work area, we begin by identifying this major component. If the operator walks within the work area, we begin with the walking pattern. If the operator is stationary (in a chair, or does not walk), we begin by observing his or her hand motions.

There are no hard rules about how to document the work at this stage. The objective is to record what is happening in such a way that the big waste can be seen and understood by everyone. The level of detail for describing the work steps is relatively basic. It is not intended to prescribe how the work is to be performed; rather, it is a description of what is happening.

Since we’re looking for big waste, the general rule of thumb is to record each time the operator takes a step from place to place for walking jobs or moves his or her hands for stationary jobs. We are looking for the waste, and not necessarily the details of what is being done at each step.

After the steps have been identified, the amount of time for each step is captured and recorded. Separate the time into two basic categories: work time, and walking (or reaching) time. Finally, a bird’s-eye drawing of the work area is made, the location of the work steps is added, and the steps are connected with a line. This drawing is very important, and visually powerful. Make it large enough to get a clear picture. Do not worry if the drawing looks “messy” with too many lines and circles. That’s the point! When the picture is completed, look

at it and ask, “What does it look like?” Perhaps the answers will be, “Messy, lots of moving around, lots of crisscrossing, backtracking, etc.” Visually, people will see that the work flow is not good. If you are fortunate to have good work flow at this stage and do not see a messy picture, you’re ready to go a level deeper and analyze for smaller hand motions.

Figure 4-1 shows a completed waste analysis, including the work steps, the work and walking time (in seconds), and the pictorial view of the operator work flow. As you can see, the walking time is two-thirds as long as the work time, and the picture shows a nonlinear work pattern with significant distances, backtracking, and crisscrossing paths.

Figure 4-1. Completed waste analysis

Remember, the first step is to thoroughly understand the current situation. Only then should you start to identify an optimal condition (reduced walking time) and then work toward how to create it. There are many options and techniques, but the basic idea is to have the work “flow” in a continuous fashion with no moving back over tracks already made. (See Chapter 6 for more on the use of standardized work, including examples of the documents used.) During the stability phase the process is used primarily to identify the waste rather than to establish “standardized work,” which is not possible until a certain degree of stability is established.

5S and Workplace Organization

We group 5S (Figure 4-2) and workplace organization together, and some would argue they are in fact one and the same, being primary methods for clearing the first layer of “clouds” by physically removing the clutter in a work area. Many people mistakenly believe that 5S is merely a clean-up initiative, perhaps because a clean work area is one outcome. The primary purpose of the first S in 5S is to clear the clouds, which involves eliminating the waste of motion from moving things and the waste of looking for tools and materials. However, other components of the 5S process—Straighten or Set in order, and Standardize—develop disciplined work habits that are crucial in later phases of lean implementation.

Figure 4-2. The 5S process

Consolidate Waste Activities to Capture Benefits

This strategy is often overlooked because of misguided beliefs. One such belief is that individual efficiencies can be maximized if each person works independently. In this way the problems encountered by one operation do not negatively influence others. As we will see in the next chapter, this philosophy will allow problems to be minimized, which makes the urgency to correct them also minimal. In addition this thinking allows for isolated waste activities that are absorbed by each person. Each operation then carries a high waste burden, and in many cases the waste is identical to wasteful procedures required at other operations.

Case Example: Consolidate and Conquer the Waste

In this example several operations were working independently to assemble various models of a product. Each operator had non-value-added activities in common with all other operations, such as retrieving material from the storage area, preparing the material for assembly, completing the shipping documents, and transporting the completed orders to the shipping area. Operators did this for themselves. Standing in the circle and carefully observing all operations revealed that about 20 percent of each operator’s total time was consumed by these activities (see Figure 4-3).

Multiply this across all operators and the waste was enormous. This did not include other non-value-adding activities within the work process.

Utilizing standardized work charts to analyze the work indicated that these waste activities could be consolidated to one “line support” operator, who would be able to minimize the waste by performing those activities collectively, thereby reducing the waste of conveyance. This meant that one operator was removed from the line to perform this “consolidated waste,” which was at first resisted by management (see Figure 4-4). By streamlining these consolidated activities, the time required for these tasks was reduced. The line support function then had free time available to perform other duties, such as data collection and reporting, and problem solving.

Figure 4-3. Each operator wastes effort getting materials

Line Support

Figure 4-4. Non-value-added activities consolidated

In addition to establishing an available resource, the consolidation of non-value-added tasks creates a cyclical, repeated process for the pickup and delivery of material. This activity should be performed on a timed cycle, or pitch. This pitch is defined based on the needs of the operations and other factors and is the foundation for standardizing the movement of material.

Standardizing this activity includes what will be done, who will do it, and when it will be done. It is important that these tasks be cyclical and repeatable so the foundation is established for standardization. Once it is established, additional improvements can be added, such as specific containers and delivery carts, and racks for presenting material to the operator. Many companies get the cart before the horse and attempt to create the devices (carts, racks, containers) before they establish the process—a standardized repeatable method. Once the process is standardized we can then look for opportunities to reduce the labor required by rebalancing the work among other operators. Ultimately we usually find labor savings both in direct labor and in material handling.

Another misguided belief is that it’s preferable to perform certain activities less frequently so that waste is minimized. This is most commonly applied to the movement and delivery of materials in a facility. There are additional factors that contribute to this belief, namely the distinction between “direct” and “indirect” labor. Within Toyota, all manufacturing employees have the same classification. They are all called “production team member,” and there is no distinction regarding the type of work performed. All employees are viewed as assets, and the cost of the asset is the same regardless of the type of work being performed. Waste is waste, and the cost effect is the same regardless of job function.

In contrast, managers of other companies are more often evaluated based on their ability to control indirect labor costs, which means fewer material handlers. If there are fewer material handlers, the obvious solution is to deliver larger quantities to the line less frequently. In many ways, this method increases the overall waste, and the net result is a higher “total cost.” (Most cost accounting systems are focused on the individual cost of labor, or piece production costs rather than the overall cost for the entire system.)

The case example below compares the two thought processes. The Toyota Way is to always focus on optimizing value-adding activity, and any system established begins with consideration for the operator and minimizing waste. We use the expression, “Treat the value-adding operator as a surgeon.” A surgeon needs to focus exclusively on the patient, and when reaching out to request a scalpel, an assistant places the right tool directly in his or her hand. This philosophy leads to increased quality and generally lower overall waste.

At an assembly plant of a major automotive manufacturer, the manager directed a continuous improvement team to focus on reducing indirect labor costs by minimizing the number of times parts were moved from a warehouse area to the assembly line. The plant manager was intent upon bringing the material from the truck to the line directly, with a minimum number of trips. It was hard to understand why he focused on this. It’s likely this was driven by years of being beaten up by upper management to reduce labor cost as the primary directive. This narrow objective can often lead to eliminating one waste but creating other more serious wastes. In this case the plant manager was convinced that producing in larger batches and putting product in large containers would save material handling costs. But what are the other consequences of this “big box” mentality?

Reviewing Figure 4-5 below, and beginning with the value-adding operator, we see that the overall length of this one sample work zone

on the assembly line is 30 feet long. The reason for the length is a combination of the variety of parts the operator is responsible for installing and the size of each container. The containers are approximately four feet wide by four feet deep by four feet high. Because of this size and weight, only one box of each type part can be placed on a heavy-duty roller conveyor. Keeping one box for each type of part creates a huge amount of inventory line side.

Figure 4-5. Long work area dictated by large containers

Because of the long work space (three times the length of the vehicle), the operator has an excessive amount of walking. When there are about 20 pieces remaining in the box, the operator contacts the material handler, who is driving back and forth waiting for an empty container. The material handler brings the new container to the line, sets it down, and removes the empty container. Because the timing is not exact, the operator must remove any remaining parts from the container and set them on top of one of the neighboring containers. (Aside from the fact that this is wasteful activity, there’s a good chance of product damage as well as mixing up similar products, leading to the installation of incorrect items.) The old container is removed and set aside, the new container picked up and set on the line, and the operator replaces the parts previously removed from the old container (again wasteful, and potential for damage ).

A closer look reveals another problem. The indirect labor, which is the focus of attention, is actually not very effective. In this case the material handler is only able to service one customer at a time (the operator), and only one item (part) because the material is delivered in large heavy containers requiring a forklift. There is substantial waste in the material handler’s job going back and forth throughout the day.

Also, the work method in this situation is impossible to standardize. Since the line will only hold one box of each type of part, the timing of the exchange is driven by the consumption of each part (which is driven by the model mix), and that timing will therefore never be consistent. It is impossible to dictate a specific time to deliver a certain part to the line.

Anytime it’s not possible to standardize a job task, the result will be a less efficient operation. It is impossible to define cyclical work and to ensure that the methods are refined. Consolidating this waste allows the creation of a standardized process for material handling that also allows the delivery of a large variety of parts to many operators.

The Toyota Way is to begin with a focus on the value-adding operation. This view would conclude that for the operator to be most effective, there would have to be a minimum of walking and the operator should be able to install a greater variety of parts. This leads to the understanding that a greater variety of parts need to be delivered to the work area, in a smaller space, and that the replenishment of parts should not require the operator to remove parts from containers before they are needed on the vehicle.

A lightweight “flow rack” can be constructed that will accommodate a large variety of parts in a much smaller space. Since the containers are smaller in height, the rack is designed to handle multiple levels, as shown in a front view in Figure 4-6, and provides for the return of empty containers that are collected by the material handler. The rack is also deep enough to hold several containers of each part type, and the exchange and replenishment of material is handled without interrupting the operator.

10 Feet

Figure 4-6. Front view of flow rack

Since each operator is not required to walk long distances, they can install additional parts. This consolidation will reduce the number of operators on the line by about 20 percent.

If these non-value-adding activities of many operators are consolidated and “pushed” out of the work area, the resulting waste becomes the burden of the material handler, who is now required to service many customers simultaneously and must create an efficient work pattern that will meet their needs. The material handler can drive a small electric cart that pulls a chain of dollies carrying a large variety of “right sized” containers for a large number of operators. Because this method requires smaller containers with lower quantities per container, the frequency of replenishment will be increased, which will increase inventory turnover, a desirable characteristic; however, it will not increase the labor needed. In fact, it is likely to reduce the overall labor requirement for material handling.

Improve Operational Availability

Often, we find processes that struggle to meet the requirements of the customer. The root causes are generally attributable to production opportunities that are lost due to the unavailability of equipment. The causes for lost time are numerous; however, they fall into two main categories:

1. In-cycle losses. These are losses that occur during the work cycle (as the equipment is operating). They may include excessive motion and equipment travel distance. One such case involved a spot welder who had a six-inch stroke when only three inches was necessary to clear the work This extra distance traveled required two seconds every cycle. Cycle losses are generally considered first because they may be easily corrected, the improvement is immediate, and it is gained each and every cycle. Multiplying the small amount of time by the frequency of occurrence (every cycle), these small changes can amount to significant gains in operational availability.
2. Out-of-cycle losses. These generally occur when the equipment is not operational. The losses per occurrence tend to be significant, but the frequency of occurrence is less. One of the significant losses is equipment setup or a tooling The principles pioneered by Shigeo Shingo, known as SMED or Single Minute Exchange of Dies, can be used to dramatically reduce this time. Also known as “quick changeover,” or “rapid changeover,” this method can be applied any time equipment is “changed” from one physical state to another. This may include tool changes, material changes, or changing to a different product or configuration. Additional causes for out-of-cycle losses are easily identified using a simple comparison of value-adding and non-value-adding activities as shown in the following case example.

Case Example: Improving Operational Availability at the Cedar Works

The Cedar Works produces wood birdhouses. The first step of the operation involves slicing the raw wood stock into thin slabs using a band saw. As a result of a sharp increase in demand, this operation was running seven days a week, 24 hours a day, in an attempt to maintain production levels. After four hours of standing in the circle, it was estimated, however, that only about 30 percent of the saw capacity was being utlilized. The department manager, was incredulous. “That’s crazy!” he said. “We’re working 24/7! How can we get more out of this operation?” Having not had the opportunity to stand in the circle, he’d fallen into the trap of confusing “work” and “activity” with value-adding time, confusing the activities of the person and the machine.

To improve his understanding, we first reviewed the concept of the seven forms of waste (non-value-adding) and value-adding activities. Beginning with the easier side of the comparison, we identified the valueadding activity and agreed that the saw added value when the blade was cutting wood. We also agreed that there are other “necessary” activities performed, though they do not help achieve the end goal of cutting more wood. We then agreed that only when the blade is cutting wood is value truly added by the saw. Now the comparison was simple: On the value-added side we had “blade cuts wood,” and on the nonvalue-added side, “everything else.”

By standing in the circle and observing, we saw many situations when the blade was not cutting wood. This list was shared with the operators, who were asked to add any additional items that were not observed. We suggest standing in the circle at various times of the day and on multiple days to get a fairly complete understanding of the situation.

Figure 4-7 shows a side-by-side comparison of value-added and nonvalue-added activities. It shows a typical situation for any operation. There will generally be few items on the value-added side and many on the non-value-added side. This provides a large selection to capture lost time opportunity by shifting from the non-value-added to the value-added side.

Value-Added Non-Value-Added

• Blade cuts wood

Non-Value-Added

• Handling wood
• Clean up
• Quality checks
• Banding bundles
• Moving bundles of wood
• Changing blades
• Breaks/Lunch
• Meetings
• Breakdown
• Adjusting saw
• Paperwork
• Waiting for wood
• Waiting for helper to return wood for additional cut

Increase Value-Added

Reduce

Figure 4-7. Comparison of value-added and non-value-added activities

From the non-value-added list we first focused on in-cycle losses— those occurring during the operation of the saw. The operators realized that simply changing the wood-handling method would increase the value-adding time nearly 25 percent. In addition, shifting activities that were currently performed “internally” (while the saw was stopped) to “external” (performed while the saw continues to add value) was borrowed from the quick changeover technique. These changes were easy to implement, and the cost was minimal.

Out-of-cycle losses were the secondary focus: primarily, reductions in time for blade change (quick changeover) and in cleaning time. The blade change time was reduced from 10 minutes per change (average two times per shift) to 2 minutes, and cleanup time was reduced from 30 minutes to 15 minutes per shift.

Reduce Variability by Isolating It

Reduction of variability is the key to achieving stability. Variability comes in two forms:

1. Self-inflicted variability—that which you
2. External variability, which is primarily related to the customers, but also to suppliers and to the variation that is inherent to the product itself (different sizes, shapes, and complexities). External variability may not be within your ability to change. However, you can build systems to compensate for the effects of the variability, mitigating the

A common example of self-inflicted variability is the way many companies apply resources—people and equipment. Many companies that operate with the “island” method—with each operation working independently of others— have no way to fill a position if an employee is absent. This includes vacation, which is a planned occurrence. In most companies, planned and unplanned absences amount to between 10 and 20 percent of workdays. During these occurrences, planned work is not completed, workers are shifted around to the “hot” jobs, and other work—much of it already in process, thus wasting the time and effort already expended—is left to wait. Once this first domino falls, a chain reaction begins of chasing the hottest jobs and shifting resources (now people and machines), all of which magnifies the variation.

The problem with variation is that once it gets started and an “adjustment” is made, it sets off additional waves of variation, making it difficult to return to “normal.” We should note here that many people incorrectly believe that a lean process is rigid and inflexible because the ability to make “adjustments” randomly is removed. We will explore this in Chapter 6, but for now we can say that the idea is that a standard condition will manage planned occurrences such as absences, and that response plans will handle unplanned events such as equipment failure—without negative consequence to the customer, and with a quick return to the standard method.

A common example of external variability is product demand, or model mix. The reality in today’s manufacturing environment is a shift from high volume, low variety (HVLV) to low volume, high variety (LVHV). This creates many challenges since the different products require varying amounts of time and/or processing steps to manufacture. Balancing the resources (people, machines, and material) with this inherent product variation is virtually impossible without employing the concept of isolating variability. If you are unable to control the variability, the next best option is to isolate it, which reduces the impact on the whole. In the last chapter, where we discussed value stream mapping, we brought up the concept of a product family. In fact, separating products into “similar families” that belong to a common value stream is an example of isolating variation.

In considering methods for isolating variation, it’s important to think about future steps. These will include the creation of flow and pull, as well as standardizing. The value stream mapping process is a useful tool for developing an understanding of the relationship of the different processing steps and times, and the effect on creating balanced flow later.

The 80/20 Rule

The 80/20 rule is useful when considering divisions in products that will isolate variation. The time required to complete the product at each operation is the critical element for the creation of future flow, so look at the products to determine where the variation occurs relative to time. To reduce variability in processing time we consolidate similar products based on the required pro cessing time. Time is also the important factor in establishing the alignment of resources.

In fact, some operations are not affected by product variability. (We call these operations “flow through” processes, because all products flow through without any change in time required.) For example, a washing or cleaning operation is not affected by the variation-in-part complexity, or model type, and thus requires the same amount of time regardless of what is produced. We are looking for the operations that are most affected by the product variation, especially if they create a bottleneck.

The tricky thing about variation is that 20 percent (the minority) of the product often provides 80 percent of the total variation. This may be difficult to see, because the ripples of prior variation create new ripples. A great deal of variation can seemingly be “removed” from the overall results of an operation by simply isolating this minority—“seemingly” because the variation is in fact not removed at all, but the magnifying effects are reduced, providing greater consistency.

Case Example: Isolating Variability in a Low Volume Aerospace Supplier

This company produced welded tube sections for the aircraft industry, an operation with a spectrum near the extreme end of low volume (average order, about five pieces) and high variety (in the thousands). Long lead times are the norm in this industry, and the desired outcome was to reduce the throughput time through the bending and welding portion of the operation. Figure 4-8 shows the average throughput time by month. It indicates an unstable process, and the range in time is between 14.5 and 21 days, with the average near 17.5 days.

Using a value stream map, it was determined that the welding area was the controlling point in the flow. Observation and discussion revealed that the complexity of the tubes caused significant variance in the weld ing time per tube. This contributed to the fact that daily output in pieces varied greatly as well. Also, in reviewing the entire value stream, it was determined that the welding operation is the most critical, time-consuming, and difficult process, and is most affected by the variation in product complexity. These characteristics made the welding area a good choice as the initial work area for achieving stability, since performance in the other processes in the value stream were more capable and stable.

Evaluation of the products showed that while each tube is unique (high variety), they fell into common groupings in regard to the time required to weld. On the low end, tubes required less than 10 minutes per piece; the middle range was between 10 and 30 minutes;

Figure 4-8. Throughput time prior to isolating variation of welding time

and on the high end, tubes required 30 minutes to several hours to complete (some could take days). From this perspective, the low-end tubes had a narrow range of variation and the high end had a large range. On a volume basis, 80 percent of all tubes were in the low to middle range, allowing the variation in time, relative to the total range, to be isolated within a narrow range.

This narrow range provided an opportunity to effectively align resources to the workload. The narrow range on the low and middle time tubes allowed us to establish a takt rate and then to determine the number of welders needed to meet the rate.

Since the mix of product varies as a result of customer requirements, it’s necessary to be able to determine the alignment of resources with the workload on an ongoing basis. The “standard” was determined based on average volume history, which is a fair indicator, but current reality rarely matches the average. In this case, current “real-time” indicators were needed so that everyone could see the actual product mix at any time and adjust accordingly, in order to maintain flow.

During the creation of lean processes, it’s often necessary to bring forward concepts from subsequent phases and to introduce them “early.” Utilizing basic concepts of connected flow and pull (to be described in the next chapter), the team created visual awareness of actual demand by setting up defined locations and quantities of work in process (WIP) for each category (low, mid, high). Minimum and maximum quantities were defined for each location, providing a standardized indicator—brought forward from the standardize phase—for the team so they could make decisions about allocating resources. These visual indicators were added throughout the value stream so that each operator worked to maintain consistent flow.

Defining and controlling the WIP at each operation reduced the range of throughput time, and decreasing the quantities further will drive down the overall time. Figure 4-9 plots the results. Clearly, the process is more consistently achieving a throughput time of 15 days, and the graph indicates that the process is “stable” in terms of this performance. Having achieved a basic level of stability, this value stream is ready to move forward in the continuous improvement cycle.

Level the Workload to Create a Foundation for Flow and Standardization

As we saw in the previous case example, the establishment of product groupings in order to isolate variation is a crucial step in the development of stability and a foundation for creating flow and establishing standardization. In essence, this isolation of variation is a basic application of heijunka, or leveling. By grouping similar products, we were able to level the workload for the majority of the process. The highly variable work is still difficult to standardize, but in this case 80 percent of the total is possible. This is an important aspect of creating stability. Some basic applications of leveling can be done in the stability phase, and there are advanced applications of heijunka as well, that will incrementally tighten the timing and pressure on the system in later phases. (We will discuss this in detail in Chapter 7.)

Figure 4-9. Process stability after variation of welding time is isolated

One common mistake is to attempt to establish flow or standardization too soon. As we will go into in the next chapter, creating flow between operations is designed to surface any issues quickly and to make them critical in nature (ignoring them would be disastrous). If this step is taken before eliminating major obstacles, the result will be too many problems and a consequent retreat to the “old way.” Likewise, an attempt to standardize a chaotic process with a high level of variability will most certainly cause frustration, since it is not possible to standardize variation.

If we liken the creation of lean processes to building a house, we understand that in order to support the roof, we will need walls and trusses. Foundations and subfloors, in turn, support the walls. This is easy to see and understand because a house is a real, visible, tangible object with common elements (they all have roofs of some type). A lean system, on the other hand, is not so clear. If you focus your effort on developing an understanding of the intent of each phase, rather than the application of lean tools, this process will be more successful. Understand the what before trying to apply the how. The lean tools are applied to address specific needs, and should not be applied simply because they are in the toolbox.

Reflect and Learn from the Process

1. Develop a current state map of your The primary purpose is not to complete a map, but to see what is actually happening in your organization.
   1. List at least 50 examples of waste that you observed while developing the map. At this time do not be concerned with “fixing” the problems you Simply look and notice the opportunities.
   2. If you cannot identify at least 50 examples, walk through the process again, taking more time to stop and observe (repeat as necessary).
2. Identify one specific operation from your current state map where you believe the greatest need for improvement
   1. Complete the “stand in the circle” activity at this operation for at least two hours or more (longer is better).
   2. List at least 50 examples of waste within this single operThis should be a simple task. If you have trouble identifying 50 items, you’re overlooking many examples of waste. Take time away from the process; then return with a fresh mind. Begin with the most obvious examples (big waste), and then become more focused on smaller and smaller examples of waste. If 50 examples is a simple task, keep adding to the list until you are challenged to find additional examples. This is when you will develop your powers of observation.

Identify indicators of instability in this one operation (chaos, variation, firefighting, inconsistent performance). Do not think about why these conditions exist or how to correct them. The purpose is simply to observe the current condition.

1. Make a list of the indicators of instability that you observed.
2. Separate the list into two categories based on whether the instability is caused by external issues (customer demand and product variation) or by internal issues (changes made that are within your control).
3. Review the suggestions in this chapter and determine the strategies and lean tools needed to address the