Improve Phase of Six Sigma Green Belt Tutorial

5.1 Improve

Hello and welcome to the fifth lesson of the Certified Six Sigma Green Belt Course offered by Simplilearn. This lesson will focus on the improve phase of the DMAIC process. Let us start with the objectives of this lesson in the next screen.

5.2 Objectives

After completing this lesson, you will be able to: Describe the various concepts of Design of Experiments (DOE). Discuss the cause and effect matrix and the 5 Why tool in root cause analysis. Explain lean and Kaizen techniques. Let us proceed to the first topic of this lesson in the following screen.

5.3 Topic 1 Design Of Experiments

In this topic, we will discuss design of experiments in detail. Let us learn about design of experiments in the following screen.

5.4 Design Of Experiments An Introduction

Design of experiments or DOE consists of a series of planned and scientific experiments that test various input variables and their eventual impact on the output variable. Design of Experiments can be used as a one-stop alternative for analyzing all influencing factors to arrive at a successful model. DOE is applicable where multiple input variables known as Factors affect a single Response Variable. An Output variable is the variable which may get affected due to multiple input variables. DOE is preferred over One Factor at a Time or OFAT experiments because it does not miss interactions. With techniques like Blocking, experimental error can be eliminated. The trials should be randomized to avoid concluding that a factor is significant when the time at which it is measured or the sequence followed may have influenced the response’s result. An example of blocking is highlighted in the table given on the screen. With techniques like Replication, many experiments can be conducted to ensure a robust model. The next screen will focus on the basic terminologies used while conducting experiments.

5.5 Basic Terms

The basic terms used while conducting experiments are Independent variables, Dependent variables, response, factors, levels, treatment, error, and experimental unit. Click each term to know more. Independent Variables are the factors intentionally varied by the experimenter. Dependent variables are the responses that vary as a result of changes made to the independent variable. For example, an experiment is conducted to determine the factors related to consistent plastic part hardness. In a plastic processing process, mold temperature influences the consistency in plastic part hardness. Hence, plastic part hardness becomes the dependent variable and mold temperature is the independent variable. A response is an outcome of an experimental treatment that varies as changes are made to levels and factors. Factors refer to a number of independent variables that are changed during an experiment to validate their impact on the output. They may be quantitative or qualitative. Levels are the values or conditions of the factors that are tested during the experiment. Most experiments test factors at two or three levels. Hence, in the experiment, one level can be 700 degrees mold temperature with filler and the other level can be 900 degrees mold temperature without filler. Treatment is a combination of factor levels whose effect on the response variable is of interest. Suppose oven temperature and the type of raw material are the two factors of interest. You need to determine their effect on the output of the response variable, which is hardness of the plastic compound. The temperature can be varied at 700 degrees and 900 degrees. These are the quantitative factors. Raw material types are attributes, hence plastic with fillers and plastic without fillers become the qualitative factors. A total of six treatments can be considered with 2 factors and two levels. Error is the variation in experimental units that have been exposed to the same treatment. This variability is due to uncontrollable factors. An experimental unit is the quantity of material in the manufacturing industry or the number served in a service system to which one trial of a single treatment is applied. The following table shows the experiment conducted to determine the factors related to plastic part hardness.

5.6 Repetition Vs Replication

Let us compare repetition and replication in this screen. Repetition is running several samples during one experimental setup without changing the setting. Repetition shows the short-term variability in the outcome of the experimental setup. Replication is repeating the entire experiment with a change in setting of experimental conditions in between the trials. This provides information on how the response variables behave in the long-term. In replication, each trial condition is reset at the start of the experiment and it is conducted in a random manner. Repetition and replication provide an estimate of the experimental error. This estimate is used to determine whether the observed differences are statistically significant. Suppose, three parts are manufactured during one trial at 700 degrees using plastic with fillers. This is repetition. After making parts using plastic without fillers, parts at 700 degrees are made using plastic with fillers. This is replication. With combined analysis, experimental error can be determined and the statistical significance in the differences in readings can be found. The next screen will discuss the concept of randomization.

5.7 Randomization

Randomization is running a trial without any order. By running random trials, the time related changes, uncontrollable variables such as variation in raw materials, and tool wear will be avoided. The use of randomization also eliminates the bias in the opinion given by the experts. Randomization is the most reliable method as it helps to draw conclusions without any ambiguity. We will understand the concept of design experiments through an example in the upcoming screen.

5.8 Doe Plastic Molding Example

To understand DOE and the main effects, consider the following example. Suppose the objective of the experiment is to achieve uniform part dimensions at a particular target value to reduce variations. The inputs (X) or Factors that affect the output are cycle time, mold temperature, holding pressure, holding time, and material type. The process is the molding process and the output or the response of the experiment is the part hardness. The components of the DOE in this example will be described in the next screen.

5.9 Components of DOE in the Molding Example

Output response, factors, levels, and interactions are the components of the DOE in the given example. Click each component to learn more. The response variable is the part hardness and is measured as a result of the experiment and is used to judge the effects of factors. Factors of this experimental setup are cycle time, mold temperature, holding pressure, holding time, and material type. Factors can be varied and are called levels. The molding temperature can be set at 600 degrees Fahrenheit or 700 degrees Fahrenheit. Plastic type can be Fillers and No Fillers and the material type has two levels of nylon or acetal. Interactions refer to the degree to which factors depend on one another. Some experiments evaluate the effect of interactions. In the molding example, the interaction between cycle time and molding temperature is critical. The best level for time depends on what temperature is set. If the temperature level is higher, the cycle time may have to be decreased to achieve the same response from the experiment.

5.10 Full Factorial Experiment An Example

Let us understand full factorial experiments through an example. Full factorial experimental design contains all combinations of all levels of all factors. This experimental design ensures no possible treatment combinations get omitted. Hence full factorial designs are often preferred over other designs. The table shown here is for a two-way heat treatment experiment. There are two factors: oven time (x 2) and the temperature (x 1) at which the material is drawn out of the oven. The output (y) of the experiment is the hardness of the material. Each of the factors has two levels. This example illustrates the concepts of main factor and interaction effects. From the table, it is clear that without repetition, the experiment will have four different outcomes based on the changes in levels of factors. Each experimental trial here is repeated to give a total of eight values. Let us now analyze the mean effect. An analysis of the means helps in understanding how a change in temperature at which the material is drawn creates a difference in the average part hardness. This affects the output and is called the main effect. Analysis of means also tells how a change in oven time creates a difference in the average part hardness. This is also the main effect. Analysis of means explains how interaction between temperature and time affects the average part hardness. This is known as the interaction effect. Let us next understand the concept of main effect. For calculating the main effect, the means have to be calculated. Hence, to calculate the main effect of draw temperature, the mean of the hardness values has to be calculated. The values are populated in the corresponding columns of draw temperatures. The columns have been labeled A1 and A2. The value of the mean of A1 is 91 and of A2 is 82. Plotting the data on a graph shows that changing draw temperature changes the average hardness. Similarly, we calculate the mean of hardness values in B1 and B2. The values are 87 and 86 which are plotted on a graph. It can be seen that changing the oven time does not affect the average hardness. Now let us understand how the interaction between temperature and time affects the average part hardness. To check how draw temperature and oven time interact, the mean values are calculated by taking the repetition response. Hence, the cell A1B1 has the mean of the values 90 and 87. The cell A2B1 has the mean of the values 84 and 87. After the mean values are calculated, they are plotted on a graph. The graph shows that to reduce interactions, low temperature and high oven time should be selected to have the desired output of high hardness, also called Brinell hardness. If low hardness is the desired output, the experimental setup should have high draw temperature and high oven time. The ideal case is represented by the parallel lines which give the desired output based on the main effect without being affected by the interaction between the factors. The parallel lines are shown as a dotted line. The mean of the factors are also calculated and shown in the small table.

5.11 Design Of Experiments Runs

In this, we will introduce the concept of runs in design of experiments. The numbers of experiments in a DOE setting is known as Runs. A full factorial experiment without replication on 5 factors and 2 levels is 2 raised to the power of 5, which equals 32 runs. A full factorial experiment with 1 replication on 5 factors and 2 levels is 32 plus 32, which equals 64 runs. A half fractional factorial experiment without replication on 5 factors and 2 levels is 2 raised to the power of 5 minus 1, which equals 16 runs. A half factorial experiment with 1 replication on 5 factors and 2 levels is 16 plus 16, which equals 32 runs. The number of combinations can be determined using the formula LF (pronounced as L to the power F) where L is the number of levels and F is the number of factors. Half fractional factorial is calculated using the formula LF-1 (pronounced as L to the power F minus one). At 3 levels, five factors full factorial experiment would amount to 243 trials and half-factorial experiments would require 81 trials. The difference between full factorial and half fractional factorial experiments can be seen from the number of runs. Let us proceed to the next topic of this lesson in the following screen.

5.12 Topic 2 Root Cause Analysis

In this topic, we will discuss root cause analysis in detail. Let us learn about process input and output variables in the following screen.

5.13 Process Input And Output Variables

Process improvement has a few prerequisites. Before a process can be improved, it must first be measured to assess the level of improvement required. The first step is to know the input variables and output variables, and check for any relationship. The SIPOC map and the Cause and Effect Matrix are very helpful. There are many ways to measure the key process variables. Metrics such as the percent defective, operation costs, elapsed time, backlog quantity, and documentation errors can be used. Critical variables are best identified by the process owners. Once they are identified, cause and effect tools are used to establish the relationship between variables. A cause and effect matrix is shown on the screen. The key process input variables have been listed vertically and the key process output variables horizontally. For each of the output variables, a prioritization number is assigned. Numbers which reflect the effect of each input variable on the output variable are entered in the matrix. The process output priority is multiplied with the input variables to arrive at the results for each input variable. The values are added to determine the results for each input variable. For process input variable one, the output variables are three, four, and seven with a prioritization value of four, seven, and eleven respectively. Therefore, multiplying the output variables with their corresponding prioritization numbers and adding those gives one hundred seventeen, which is around thirty-three percent of the total effect. The process input variables results are compared to each other to determine which input variable has the greatest effect on the output variables. Click the cause and effect matrix template button to view another template. A sample of the Cause Effect Matrix or CE Matrix is shown here. The CE Matrix gives the correlation between Input and Output variables.

5.14 Cause And Effect Matrix Steps To Update

In this screen, we will discuss the steps to update the CE matrix. The steps for updating the Cause and Effect Matrix are: List the input variables vertically under the column process inputs. List the output variables horizontally under the numbers 1 to 15. These output variables are important from the customer’s perspective. One can refer to either the QFD or the CTQ Tree to know the key output variables. Rank the output variables based on customer priority. These numbers can also be taken from the QFD. The input variables with the highest score become the point of focus in the project. Another method to establish the cause effect relation is the Cause and Effect diagram. This is explained in detail in the following screen.

5.15 Cause And Effect Diagram

The Cause and Effect diagram is used to find the root cause and the potential solutions to a problem. A cause and effect diagram breaks down a problem into bite-sized pieces and also displays the possible causes in a graphic manner. It is also known as the fishbone, the 4-M, or the Ishikawa diagram. It is commonly used to examine effects or problems to find out the possible causes and to indicate the possible areas to collect data. The steps involved in the cause and effect diagram are: All the possible causes of the problem or effect selected for analysis are brainstormed. The major causes are classified under the headings of materials, methods, machinery, and manpower. The cause and effect diagram is drawn with the problem at the point of the central axis line and the causes on the diagram are written under the classifications chosen. The next screen illustrates the cause and effect diagram with the help of an example.

5.16 Cause And Effect Diagram - Example

The diagram shows the cause and effect diagram for the possible causes of solder defects on a reflow soldering line. This diagram helps in collecting data and discovering the root cause. During brainstorming, the group looked at all the major causes and then grouped them under the main headings. Under materials, causes like types of solder paste, components, and the components packaging used are considered. The major causes under methods are technology and preventive maintenance. Similarly, operator and schedule are placed under manpower, while tools and oven are grouped under machinery causes. The next screen will discuss another root cause analysis tool in detail.

5.17 The 5 Why Technique

5 Why is one of the tools used to analyze the root cause of a problem. The responsibility of the root cause analysis lies with the 5 why analysis team. The technical experts have a great responsibility as the conclusion will be drawn from the way the drill down of the symptoms is carried out. The 5 Why is a very simple tool as it poses the ‘why’ question to every problem till the root cause is obtained. It is important to know that the 5 Why tool does not restrict the interrogation to five questions. ‘Why’ can be asked as many times as required till the root cause for the problem is found. It can be used along with the cause & effect diagram. The following screen will explain the process of the 5 Why technique.

5.18 The 5 Why Process

The process for the 5 Why technique is: Identify the problem and emphasize the problem statement Arrange for a brainstorming session with the team including subject matter experts, process owners, and team members Explain the purpose and the problem statement Analyze scenarios working backwards from the problem. Ask ‘why’ for the answers obtained until the root cause is found. Normally, reasons like insufficient resources and time become the root causes. If the drill down in brainstorming is carried out in the right direction, it is often found that the root cause is related to the process. Therefore, the occurrence of a problem is often due to the process and not an individual or a team. In the next screen, we will understand the concept of the 5 Why technique with the help of an example.

5.19 5 Why Example

The following example will simplify the concept of the 5 Why technique. Nutri (Pronounce as: nyutree) Worldwide Inc. encountered a problem concerning erroneous deliveries by the Delivery Management team. The problem statement was defined as “delivery of the parcels to incorrect addresses.” The 5 Why technique revealed that the addresses given on the parcels were often incomplete. When probed further, it was found that complete addresses were not collected from the customers. Typically the addresses were given to the operators over phone calls. The operators were often unable to gauge if the addresses were incomplete. The root cause was found to be the unavailability of an official format for capturing delivery addresses from customers. As a counter measure, a template capturing customers’ details such as flat number, street name and number, region, city, and zip code along with the contact number was prepared. This solution helped in eliminating the occurrence of erroneous deliveries and improved their quality of service. Let us proceed to the next topic of this lesson in the following screen.

5.20 Topic 3 Lean Tools

In this topic, we will discuss lean tools in detail. Let us learn about lean techniques in the following screen.

5.21 Lean Techniques

The eight lean techniques are Kaizen, Poka Yoke (Pronounce as: poh-kah yoh-keh), 5S, Just in Time, Kanban, Jidoka, Takt time, and Heijunka. Click each technique to know more. Kaizen (Pronounced as Kaayee-Zen) or continuous improvement is the building block of all lean production methods. Kaizen philosophy implies that all incremental changes routinely applied and sustained over a long period of time results in significant improvements. The second technique is Poka Yoke (Pronounced as Po-ka-yo-keh). It is also known as Mistake Proofing. It is good to do it right the first time and even better to make it impossible to do it wrong the first time. The prompt received to save the word document before closing it without saving, is an example of Poka Yoke. 5S (pronounced as Five-S) is a set of five Japanese words which translate to Sort, Set in order, Shine, Standardize, and Sustain. This is a simple and yet powerful tool of Lean. The Sort principle refers to sorting items according to a rule. The rule could be frequency of use or time of use. After sorting, the objects are set in order. The place for everything is defined and everything is placed accordingly. Cleaning of the area refers to the Shine principle. The fourth step requires formation and circulation of a set of written standards. The last step refers to sustaining the process by following the standards set earlier. 5S is useful as a framework to create and maintain the workplace. Just in Time or JIT (pronounced as J-I-T) is another Lean technique. This technique philosophizes about producing the necessary units, in the necessary quantity, at the necessary time with the required quality. As an item is removed from a shelf of a super mart, the system confirms it and automatically sends a note for replenishment. This kind of technique can be used in an organization to prevent accumulation of inventory. The fifth technique is known as Kanban (pronounced as Can-ban), which means signboard in Japanese. Kanban utilizes visual display cards to signal movement of material between the steps of a product process. This is one of the examples of visual control in Lean. The next technique is Jidoka (pronounced as Jee-do-kaa). It means automation with human touch and is sometimes known as autonomation. Jidoka implements supervisory function in the production line and stops the process as soon as a defect is encountered. The process does not start again till the root cause of the defect is eliminated. Takt (pronounced as Tact) time is the maximum time in which the customer demands need to be met. For example, a customer needs 100 products, and the company has 420 minutes of available production time. TAKT Time equals Time Available divided by Demand. In this case, the company has a maximum of 4.2 minutes per product. This will be the target for the production line. The final technique is Heijunka (pronounced as Hi-jenka) which means production leveling and smoothing. It is a technique to reduce waste occurring due to fluctuating customer demand.

5.22 Cycle Time Reduction

Let us understand the concept of cycle time reduction in this screen. Cycle time reduction refers to the reduction in the time taken for a complete process. Implementing Lean techniques reduces Cycle Time and releases resources faster than any other method. Low cycle time increases productivity and throughput. Lean techniques release resources early, achieving more production with the same machinery. Internal and external waste is reduced and the operational process is simplified with a decrease in product damage. All these factors help in satisfying the customer and staying ahead in competition. The following screen describes the concept of cycle time reduction through an example.

5.23 Cycle Time Reduction Example

The changes brought by implementing Lean techniques on an existing process are illustrated in the given diagram. Things to be noticed are: number of operators used, work allocation to the operators, path or the movement in the process and flow of the process Notice the changes brought about by implementing Lean techniques on the old process. First, the path followed by the material in between the process is considerably reduced. This decreases the cycle time for the entire process. Second, the number of operators is reduced to three, when compared to five in the old process. Operator 1 can now work on process 1 and process 4. Similarly, operator 2 can work on process 2 and process 3. Hence, there is an increased productivity of the operators and the remaining skilled operators can be used in some other process or system. The next screen will introduce the concepts of Kaizen and Kaizen blitz.

5.24 Kaizen And Kaizen Blitz - Introduction

Kaizen means ‘good change’ in Japanese. Kaizen is a continuous improvement method to improve the functions of an organization. The improvements could be in process, productivity, quality, technology, and safety. It brings in small incremental changes to the process. Kaizen blitz is known as Kaizen event or Kaizen workshop. If the event is tightly defined and the scope is evident for implementation, processes can be easily changed and improved. Teams could improve processes through creative problem solving methods in structured workshops over a short timescale. The next screen will provide the differences between Kaizen and Kaizen Blitz.

5.25 Kaizen And Kaizen Blitz Differences

The differences between Kaizen and Kaizen blitz are: Kaizen is a method that brings continuous improvement in the organization, while Kaizen blitz is a workshop or an event that brings in change. Kaizen brings in small incremental changes in the organization. There are no major changes made within the processes. Kaizen blitz is applied when a rapid solution is required. The Kaizen method follows a step-by-step process. It standardizes, measures, and compares the process with the requirement before improving it. Kaizen blitz plans for the event, executes it, arrives at a solution, and follows it through. All the people of the organization are involved in Kaizen, whereas Kaizen blitz is led by the top management and others are invited to participate. The decision making lies with the upper management. In Kaizen, the process is standardized, and measurements are regularly collected and compared before the decision is taken. This relatively delays the process of decision making. In Kaizen blitz, decisions are taken soon and the process change is wrapped in 3 to 5 days. Kaizen is a continuous improvement method, whereas the kaizen blitz is a part of the improving process. Kaizen follows PDCA, in the sense, plan, do, check and act for the improvement process. Kaizen blitz uses PDCA for execution, where the events are planned, conducted, decided, implemented and followed up. The following screen will elaborate on the concepts of Kaizen and Kaizen Blitz through examples.

5.26 Kaizen and Kaizen Blitz Examples

Kaizen and Kaizen Blitz are practiced in many organizations across the world. The examples of Kaizen and Kaizen Blitz method are shown here. Click each tab to know more. The Toyota production system is known for Kaizen practices. In Toyota, if any issue arises in the production line, the line personnel cease all the production until the issue is resolved. Once the solution is implemented, the team resumes the production cycle. A wood window company in the State of Iowa, US uses the Kaizen Blitz method to redesign their shop floor, and replace expensive, nonflexible automation with low cost, highly flexible cellular applications. Eliminating scraps, reorganizing work areas, and reducing inventory are some of the examples of quick implementation through Kaizen Blitz.

5.27 Quiz

Following is the quiz section to check your understanding of the lesson.

5.28 Summary

Let us summarize what we have learned in this lesson. Design of Experiments (DOE) is a structured method that tests input variables and their impact on the output variable. Key concepts of DOE are factors, levels, treatment, errors, repetition, and replication. The Cause and Effect Matrix gives the correlation between Input and Output variables. The 5 Why tool is used to analyze the root cause of a problem. Cycle time reduction, Kaizen, and Kaizen Blitz help in improving processes.

5.29 Thank You

With this, we have come to the end of this lesson. The next lesson will focus on the control phase.

  • Disclaimer
  • PMP, PMI, PMBOK, CAPM, PgMP, PfMP, ACP, PBA, RMP, SP, and OPM3 are registered marks of the Project Management Institute, Inc.

Request more information

For individuals
For business
Phone Number*
Your Message (Optional)
We are looking into your query.
Our consultants will get in touch with you soon.

A Simplilearn representative will get back to you in one business day.

First Name*
Last Name*
Work Email*
Phone Number*
Job Title*