Chapter 3 The First Step Toward Sustainability—Lean and Six-Sigma – Sustainable Operations and Closed Loop Supply Chains, Second Edition

CHAPTER 3

The First Step Toward Sustainability—Lean and Six-Sigma

3.1 The Lean Management Philosophy: Types of Waste

Lean and six-sigma are two mainstream process improvement methodologies. Several books have been written about these two methodologies; as a result, this chapter will only summarize some main points. Although some purists may not agree completely, “lean manufacturing,” “Toyota Production System (TPS),” and “Big Just-in-Time” are three different names referring to the same thing. In a nutshell, the objective of lean is to reduce waste (muda, in Japanese) in all aspects of a firm’s production activities: human relations, vendor relations, technology, and the management of materials and inventory. There are seven forms of waste:

1. Overproduction. This is waste from producing products and services when they are not immediately needed. Examples: production in large batches in manufacturing, and returning excess surgical supplies to the shelf in health care.

2. Transportation. This is waste from moving products or parts from one physical location to another. As an example, work-in-process (WIP) inventory is moved from one end of a manufacturing plant to the other end, as a result of the layout of the facility, that is, the location of manufacturing equipment within the facility.

3. Inventory. In lean, inventory is considered waste. Of course, a certain amount of inventory is necessary to protect against uncertainties inherent in processes, but lean views inventory as an impediment to the observation and correction of process problems. This is illustrated with the “rocks in the river” analogy in Figure 3.1. In this analogy, the water level represents WIP inventory, and the rocks represent problems in the process. By lowering the amount of inventory (through, for example, a kanban system, which we explain later), the firm is able to uncover problems in the production process. In Figure 3.1, for example, if inventory is lowered by a small amount, the firm will notice machine breakdowns (that is, inventory was high to protect against machine breakdowns). The firm then implements a preventive maintenance program. The firm continues to lower inventories, and then uncovers materials that are out of spec; the firm then works with suppliers to address that. And so forth.

4. Waiting Time. This is waste incurred when people (employees) wait for the arrival of some part, information, or equipment to accomplish a task. For example, when operators wait for parts to arrive from an upstream process step, or when doctors are waiting for lab results.

5. Motion Waste. This is similar to transportation waste, but here it is related to people. For example, when a machine operator walks 20 ft to another room to pick up a tool that is needed to setup a machine for production. As another example, when a nurse has to take an elevator and go down two floors to pick up supplies to care for a patient.

Figure 3.1 “Rocks in the River” analogy: Water level must be lowered.

6. Processing Waste. This is waste related to unnecessary process steps. For example, if there is too much paperwork, or when a patient is required to go through redundant or unnecessary steps because the doctor practices defensive medicine.

7. Defects. This is a natural definition of waste for most people—fixing defects. For example, rework (when a defect is caught in production before reaching the final customer), warranty claims, recalls, hospital infections.

The seven forms of waste are summarized in Figure 3.2.

Three Japanese words are frequently used to characterize the lean philosophy:

Heijunka. This means, level the load by reducing workload imbalances between different steps of the production process, reducing setup times (and costs), and reducing variability within the process. All of these result in reduced inventories. We discuss this in more detail later when we address pull production, one of the “tools” of lean.

Figure 3.2 The seven forms of waste.

Jidoka. This means, stop the process immediately (sacrificing output) to address an immediate production problem. At Toyota manufacturing plants, there are andon cords that run parallel to the assembly line, and a worker can pull an andon cord at any time (if a defect or problem is found) to stop the production line. In auto manufacturing, stopping a production line is an expensive proposition, since a typical line manufactures one car every 90 seconds or so. Thus, the use of the andon cord provides strong incentives to find and fix defects where they occur.

Kaizen. This means a relentless, long-term commitment to process and quality improvement.

3.2 Lean Toolkit

Pull Processes and Setup Time Reduction

The basic idea behind pull processes is to significantly reduce WIP in the system, and as a result, reduce throughput, or lead times (i.e., the total time a unit spends in the system, including waiting and processing). This is because inventories and throughput times are related through Little’s Law:

3.1

A consequence of Equation 3.1 is that there are only two ways to decrease throughput times:

Increase the production rate. This most likely necessitates investments in equipment or labor, because the production rate of a process is equal to the rate of the bottleneck resource.

Decrease WIP. This can be accomplished by using a “pull process,” which sets the amount of WIP in the system to a constant value, as discussed later.

In a push process, work is scheduled and pushed through each stage in the process in order to meet specified delivery dates for finished products and services. That is, work happens in anticipation of demand forecasts. Consider the simple two-stage process in Figure 3.3, where each circle with a number represents one unit of WIP. Stage A will deliver unit 6 to WIP Pile B as soon as it finishes its processing. In a push process, the upstream stage (stage A) authorizes the downstream stage (stage B) to work by delivering WIP.

Figure 3.3 Illustration of push and pull processes in a two-stage process. (Each circle with a number represents one unit of WIP.)

On the other hand, in a pull process, work at each state in the production process is pulled through the system by actual demand for final products and services. One way to implement a pull process is through a kanban system (kanban means card in Japanese), where the WIP for each stage in the process is set to a constant (i.e., the amount of WIP at that stage is equal to the number of kanbans designed for it). This can be seen in Figure 3.3, where the size of the kanban at stage B is four units (that is, the maximum WIP waiting to start processing at stage B—WIP Pile B—is four units). If stage A completes its work on unit 6, it will only deliver that unit to WIP Pile B when stage B completes work on unit 1. At that time, unit 6 is delivered to WIP Pile B, and stage A can start working on unit 7. In a pull process, the downstream stage (stage B) authorizes the upstream stage (stage A) to work in order for WIP to be constant.

The size of the kanban (i.e., the amount of WIP the system is allowed to maintain) is related to the amount of variability in the system, including quality. A highly variable system with lower amounts of WIP results in frequent process stoppages. To see this, consider again Figure 3.3: if the throughput rate of stage B becomes higher than that of stage A, then it is possible that stage B will consume all the WIP in WIP Pile B before A is able to “catch up,” which may result in an empty WIP Pile B; stage B will be starved and the process will stop. Lean is also concerned with removing variability and workload imbalances in the system, so that there is less idle time (or production stoppages) overall.

In addition, if stage B identifies quality issues with units 2–5 and these have to be re-sent to stage A for rework, then stage B again will be starved. That is why jidoka and kaizen are essential components of lean: without quality, it is not possible to implement a pull process effectively.

Setup time (cost) is a fixed time (cost) needed for an operation to take place, regardless of the quantity processed. As an example, consider the process of slicing cheeses at a deli, which requires positioning of the cheese at the machine and calibrating it for the required thickness before the slicing can begin, regardless of the number of slices processed. Setup times are not only common in manufacturing, but they also exist in services: consider the loss in productivity that result from a worker switching tasks. Setup times result in inventories, because operations need to be carried out in batches (so that the fixed setup time can “distributed” among more units), and as a result, units are processed in advance of when they are needed. Lean seeks to achieve small batches through the use of setup reduction techniques, such as SMED.1 SMED typically starts by filming the setup process, and identifying possible ways to reduce the time involved in it, for example, positioning needed tools closer to the worker, or designing and using fixtures that make calibration less time consuming.

Value Stream Mapping

Value stream mapping can be thought of as a process flow chart where activities are identified as belonging to one of three types:

Value Added activities (VA): activities in a process for which the customer is willing to pay. Examples include nursing care, surgery, or a worker assembling a wheel in a car production line.

Non Value Added activities (NVA): activities for which the customer is not willing to pay, and are not necessary for business. Examples include a nurse walking 50 ft to another room to search for supplies, or a worker retrieving inventory from a warehouse.

Business Non Value Added (BNVA): activities for which the customer is not willing to pay, but are necessary for accounting, legal, or regulatory purposes. Examples include preparing financial statements, or disclosing information to regulatory agencies.

In a value stream map, additional information is typically included in the chart displaying the activities. Examples include processing time per unit, wait time, estimated cost, and changeover cost. The objective of a value stream map is to reduce waste in the process by eliminating or reducing NVA activities. Figure 3.4 displays a value stream map for part of a process at an assembly plant. The only VA activity in Figure 3.4 is the assembly of the product by the technician, and that is shaded in gray. The other activities constitute waste. Moving parts around the plant constitute transportation waste. Inspection activities constitute processing waste (if there was an effective quality program in place, there would be no need for inspection). Storage constitutes inventory waste. For any value stream map, we define process efficiency as the percentage of time the process is spent in VA activities. For the process in Figure 3.4:

3.2

Efficiencies around 20 percent are common in batch manufacturing, where there is considerable waiting time. At some service organizations, efficiencies can be much higher. For example, at Virginia Mason Medical Center, a hospital in Seattle (WA), the process efficiency for breast cancer diagnosis and treatment was reported to be 70 percent.2 In fact, Figure 3.5 shows a value stream map for a nurse’s time at an actual hospital in the mid-Atlantic area of the United States. It can be seen that VA activities constitute 282 minutes (204 + 78) out of the 480 minutes of the nurse’s time, for an efficiency of 59 percent. BNVA activities constitute 120 minutes (85 + 35), or 25 percent. The 36 minutes of waste in this example comprise activities such as looking for equipment, looking for supplies, waiting delays, and others. Although 36 minutes of waste constitute only 7.5 percent of the total time of the nurse, improvements here can have an impact, considering one would have to multiply these numbers by the number of shifts and the number of nurses in the hospital.

Figure 3.4 Value stream map for assembly plant.

Figure 3.5 Value stream map for a nurse’s time during an 8-hour shift (480 minutes).

5S

5S is a system of procedures that are used to organize and arrange the workplace, in order to optimize performance, cleanliness, and safety. The meaning of each of the 5Ss is displayed in Table 3.1.

Table 3.1 The meaning of each of the 5Ss

Japanese S

English S

Seiri

Sorting

Seiton

Simplifying access (or, Set in order)

Seiso

Sweeping (or, Shine)

Seiketsu

Standardization

Shitsuke

Self-Discipline (or, Sustain)

Figure 3.6 Implementation of 5S at a tech support department.

Figure 3.6 shows a picture of a tech support department before and after 5S implementation. A typical 5S implementation can be accomplished through the following five steps3:

1. Plan a course of action: Obtain 5S materials, coordinate activities with all departments involved, select a team, and develop a schedule.

2. Educate the work group: This is key, so that workers in the department know the objectives, and what will be involved. Ideally, the work group should be involved in the 5S implementation, but that is not necessary—it could be done by another group responsible for productivity improvements.

3. Evaluate the work area: Map and photograph the area, define boundaries, and conduct 5S appraisals (see Table 3.2).

4. Initiate the 5S: Sort unnecessary items (e.g., use red tags, use it or lose it auctions), simplify access, sweep, etc.

5. Measure results and maintain the workplace neat (this is the last S).

Layout Redesign

The idea behind layout redesign is to reduce transportation and motion waste. Traditionally, work areas have been designed around functional layouts, where each physical location of a facility is dedicated to a particular type of equipment—that is, each physical location in the facility performs a particular function. Figure 3.7 illustrates a typical functional layout in a manufacturing plant, with five different functions (saw, grinder, lathe, press, and heat treat areas), and each function confined to a single different area in the layout. The flow lines represent the flow of a typical product in that layout: saw → lathe → grinder → heat treat → lathe → press. Note how the product moves significantly across different physical areas of the plant, which indicates a considerable amount of transportation and motion waste.

Table 3.2 Example of a 5S appraisal sheet

Sorting

Yes

No

1. Do all teams know this program is in place?

   

2. Have criteria been defined to distinguish between necessary and unnecessary items?

   

3. Have all unnecessary items been removed from the area?

   

Simplifying access

   

4. Is there a visually marked, specified place for everything?

   

5. Is everything in its specified place?

   

Sweeping

   

6. Are work and break areas, offices, and rooms clean and orderly?

   

7. Are cleaning guidelines and schedules visible?

   

Standardizing

   

8. Are current processes documented?

   

Source: Peterson and Smith (1998).

Figure 3.7 A typical functional layout.

Suppose now that the layout is then redesigned as a cellular layout, and that is shown in Figure 3.8. That layout has two cells: A and B, with the only heat treatment station being shared between cells A and B. This new layout minimizes transportation and motion waste, as the typical travel time of a product in the facility is shortened considerably. The drawback from the cellular layout is the loss of “pooling” among identical machines. For example, suppose that in Figure 3.8 the firm produces two products, A and B, where product type A is processed in cell A, and product type B is processed in cell B. If the saw in cell A is busy and the saw in cell B is idle, then a product of type A will have to wait until the saw in cell A becomes available. This is despite the fact that the other saw (in cell B) is available, because cell B only processes products of type B. This situation would not arise in the functional layout of Figure 3.7. In most applications, however, a typical manufacturing plant processes multiple (sometimes hundreds) of product types, and production cells are designed so that a cell works on a family of products that share similar processing requirements.

Figure 3.8 Reorganization of layout from Figure 3.7 into a cellular layout.

Although the examples in Figures 3.7 and 3.8 are from a manufacturing environment, the idea of cellular layouts can certainly be applied to services. Examples in a hospital environment include: (i) creating self-contained administrative units, (iii) providing data entering devices for nurses and doctors located near patients (as opposed to only in offices), (iv) optimizing patient flow, and (iv) optimizing staff flow through reconfiguration of work stations.

3.3 Six-Sigma: Similarities and Differences with Lean

Six-sigma is another process improvement methodology that is focused in reducing variation in processes, and it can be summarized on the DMAIC steps as shown in Figure 3.9. As can be seen, six-sigma is heavily focused on decision-making using data and statistical tools, which clearly demand a significant amount of training. In addition to tools, six-sigma is very disciplined, and involves full-time personnel with full knowledge of all the tools (black belts), as well as workers with a good knowledge of the tools, who work on these initiatives part time (perhaps 10–20 hours a week; the green belts). A black belt may have 4–6 weeks of training in the methodology plus experience in implementing it in actual projects. A green belt may have 25 hours of training in the methodology.

Figure 3.9 Six-sigma’s DMAIC process and some tools that can be used in each step.

Six-sigma and lean are both focused on continuous, incremental improvement of existing processes, as opposed to radical redesign. Some of the tools are common, such as the use of process flow charts, run charts, and data collection. The tools in lean, as seen before, are more straightforward and do not require extensive statistical training as is the case with six-sigma. The two are complementary in that reducing variation in processes reduces the amount of waste. This is shown in Figure 3.10, where the “bell shape” curve represents the histogram of measurements for a particular variable of interest x (say, weight of a can of tuna), , σA and σB represent the average value, standard deviation of the value for process A, and standard deviation of the value for process B for that variable. With six-sigma implementation, the variability in the process is reduced (σB < σA) and as a result fewer units fall outside the engineering specification limits.

Figure 3.10 Relationship between variability reduction and waste reduction.

3.4 Why is Lean Green?

Lean is green because waste reduction is associated with lower resource consumption, whether in the form of energy or raw materials; in addition, solid waste is reduced. For example, the Wausau Equipment Company, which produces machinery and equipment for the agribusiness industry, implemented lean at one of its plants and observed the following results4:

Parts reworked decreased by 70 percent: this clearly reduces overall energy consumption, materials use, and solid waste.

Throughput (production per unit time) increased by 35 percent: this represents a productivity improvement, which reduces overall energy consumption, and solid waste.

Thus, being lean is a first step toward being green.