There are two topics I will address in this section of my web page. First, I will briefly tell “My Story” of how I wound up studying, teaching, and practicing Six Sigma management and Lean thinking. Second, I will explain “My Point of View” about Six Sigma management and Lean thinking.


My Story

My journey on the pathway toward quality began in 1969. I was a doctoral student at New York University’s Graduate School of Business Administration and President of the Statistics Club. (People still kid me about being president of the Statistics Club.) I was called into Dr. Deming’s office to collect an anonymous gift to be given to a Japanese student studying Statistics. I had no idea why he wanted a Japanese student, but “mine was not to question why.” Much later I discovered how famous Dr. Deming was in Japan. I only knew that he was VERY well known for his work in sampling theory and practice.

Dr. Deming really shocked me when I first met him. I had really long hair at the time and his first words to me were: “Why don’t you get a haircut!” Fast forwarding eleven years, I was teaching at the University of Miami in 1980 when the NBC white paper “If Japan Can, Why Can’t We?” was aired on television. It caused quite a stir in the United States. The dean at the time, Dr. Carl McKenry, called Dr. Deming and asked if he would come give a seminar. While Carl was on the phone with Dr. Deming he said: “We have one of your former students down here.” Dr. Deming replied: “Who is it?” Carl said: “Howard Gitlow.” Dr. Deming replied: “Tell me, do you pay him enough to get a hair cut?” When I heard that story, I just couldn’t believe that he remembered me. Much later I realized that this was typical of Dr. Deming.  

Dr. Deming came to the University of Miami in 1980 to give a seminar and I was assigned to be his assistant. I sat with him on the dais and helped him. I listened very closely to what he had to say about management and had a personal epiphany. I realized that I would at best be a decent theoretical statistician, but I thought I could be a better management theorist. Well, it is now 27 years later and I am still studying management theory.

Dr Deming was very kind to me. He would take me to visit with his clients. I learned a lot by watching a master statistician ply his profession.

In the late 1980s, the top management of Florida Power & Light Company decided to transform their style of management from traditional Management by Objective, to Japanese Total Quality Control, and also to try to be the first non-Japanese company to ever win the Deming Prize. 

Back to top

My Point of View

The good news is that management is now a science. It has three component parts: (1) Macro Model (Dashboards), (2) Micro Model (Projects), and (3) Management Model (Management style), see figure 1. It is the form of professional management that you will learn in the University of Miami’s Six Sigma programs.

Figure 1: Model for Professional Management (Six Sigma Management)

Model for Professional Management (Six Sigma Management)  


Back to top

The macro model is a dashboard management system. A dashboard is a tool used by management to clarify and assign accountability for the “critical few” key objectives, key indicators and projects / tasks needed to steer an organization toward its mission statement.  Dashboards have several benefits, they include: 

Monitoring deployment of the mission statement throughout an organization using a cascading and interlocking system of key objectives and indicators.

Balancing management’s attention between financial, customer, process, and employee key objectives.

Increasing communication between, and within, the levels of an organization. 

Linking all processes (jobs) to the mission statement.

Dramatically reducing the number of daily crises that take you away from your regular work and frequently cause uncompensated overtime. These daily crises are frequently nothing more that the typical malfunctioning of the processes within your organization.

Developing and testing hypotheses about the effectiveness of potential process improvements.


The structure of a dashboard can be seen in table 1 below. The president’s key objectives and indicators emanate from the mission statement (see row 1 and columns 1 and 2 of Table 1).  Direct reports identify their key objectives and indicators by studying the president’s key indicators (column 2 of Table 1) that relate to their area of responsibility.  The outcome of these studies is to identify the key objectives and indicators (see columns 3 and 4 of Table 1) required to improve the president’s key indicator(s) (see column 2 of Table 1) to achieve a desirable state for presidential key objective(s) (see column 1 of Table 1).  This process is cascaded throughout the entire organization until processes are identified which must be improved or innovated with Six Sigma projects or tasks (see column 5 of Table 1).


Table 1: Generic Dashboard (Macro Model)

Mission Statement: --------------------------------------------------------------------------



Direct Reports

Potential Six Sigma Projects

(Micro Model)


Business Objectives

Business Indicators

Area Objectives

Area Indicators


Business objectives must be achieved to attain the mission statement.

One or more business indicators show progress toward each business objective.

Area objectives are established to move each business indicator in the proper direction.

One or more area indicators show progress toward each area objective.

Six Sigma projects are used for improving or innovating processes to move indicators in the proper direction.









The University of Miami provides an excellent example of a dashboard. Table 2 shows the dashboard between a third level manager (Vice President of Business Services) and a fourth level manager (Chief of Campus Police) at the University. More on this dashboard later.


Table 2: Portion of the University of Miami Dashboard






Back to top

Micro Model - Six Sigma DMAIC Model

The micro model includes the DMAIC model for improving an existing product, service or process, the DMADV model for inventing a new product, service or process, or innovating an existing product, service or process, and Lean thinking for eliminating waste in processes.

The DMAIC model is the Six Sigma model used to improve existing products, services or processes. It has 5 phases, they are: Define, Measure, Analyze, Improve, and Control. Each phase is briefly explained below with an example about an accounting reporting process.

(1) Define phase: The define phase involves preparing a project charter (rationale for the project), understanding the relationships between Suppliers-Inputs-System-Outputs-Customers (called SIPOC analysis), and analyzing Voice of the Customer data to identify the critical to quality (CTQs) characteristics important to customers, and developing a project objective.

For example, a Six Sigma team was assigned by top management to review the production of a monthly report by the accounting department as a potential Six Sigma project.  This involved identifying the need for the project (relative to other potential projects), the costs and benefits of the project, the resources required for the project and the time frame of the project.  As a consequence of doing a SIPOC analysis and a Voice of the Customer analysis, the team determined that management wants a monthly accounting report to be completed in 7 days (the normal time to complete is seven days). They also determined that the report should never be completed in less than 4 days (the relevant information is not available before then) and never later than 10 days.  Team members identified the project objective as follows: Reduce (direction) the variability in the cycle time (measure) to produce an error free accounting report (system) from the current level of 7 days plus or minus 3 days to 7 days plus or minus 1½ days (target) by January 10, 2003 (deadline). 

(2) Measure phase: The measure phase involves developing operational definitions for each Critical-To-Quality (CTQ) variable, determining the validity of the measurement system for each CTQ, and establishing baseline capabilities for each CTQ.

Referring to the accounting report example, the team members created an operational definition of variability in cycle time such that all relevant personnel agreed upon the definition (for example, they clearly identified the start and stop points needed to compute cycle time).  Next, they performed a statistical study to determine the validity of the measurement system for variability in cycle time.  Finally, team members collected baseline data about variability in cycle time and statistically analyzed it to get a clear picture of the current situation.

(3) Analyze phase: The analyze phase involves identifying the upstream variables (Xs) for each CTQ using a flowchart.  Upstream variables are the factors that affect the performance of a CTQ. Also, the analyze phase involves conducting a risk analysis (called Failure Modes and Effect Analysis - FMEA) for each X in respect to it's potential impact on the CTQ, operationally defining each X, performing studies to determine the validity of the measurement system for each X, establishing baseline capabilities for each X, and understanding the effect of each X on the CTQ.

Referring to the accounting report example, team members identify all input and system variables (Xs) that impact the CTQ, “variability in cycle time,” 

They are:
X1 = number of days from request to receipt for line item A data,
X2 = number of days from request to receipt for line item B data,
X3 = number of days from request to receipt for line item C data,
X4 = number of days from request to receipt for line item D data,
X5 = number of days to reformat the line item data to prepare the report,
X6 = number of days to prepare the report,
X7 = accounting clerk preparing the report (Mary or Joe),
X8 = number of errors in the report,
X9 = number of days to correct the report,
X10 = accounting supervisor correcting the report (Harry or Sue), and
X11 = number of signatures required before the report is released.

For example, the number of signatures required before releasing the report (X11) may affect the average time to produce the report, or the accounting clerk preparing the report (X7) may dramatically affect the variability in cycle time to produce the report. Next, team members operationally define the Xs and perform statistical studies to determine the validity of their measurement systems.  Fourth, team members collect baseline data to determine the current status of each X using control charts. Finally, team members study the data and develop hypotheses about the relationships between the Xs and the CTQ.  In this case, separate histograms of the CTQ for each level of each X indicated that: X1 (number of days from request to receipt for line item A data), X3 (number of days from request to receipt for line item C data), X7 (accounting clerk preparing the report (Mary or Joe)), and X10 (accounting supervisor performing the corrections to the report (Harry or Sue)) may be important to the reduction of variability in the cycle time (CTQ).  The other Xs did not substantially affect the CTQ.

(4) Improve phase: The improve phase involves designing experiments to understand the relationships between the CTQs and the Xs, determining the levels of the critical Xs that optimize the CTQs, developing action plans to formalize the level of the Xs that optimize the CTQs, and conducting a pilot test of the revised system.

Back to the accounting report example, team members conducted an experiment to identify the levels of the critical Xs identified in the analyze phase to minimize variation in the time to produce the accounting report.  The experiment revealed that team members had to work with the personnel responsible for line items A and C to decrease the average and standard deviation of days to forward the line items to the department preparing the report.  Further, the experiment revealed that there is an interaction between the clerk preparing the report and the supervisor correcting the report.  The analysis showed that if Mary prepared the report, it was best for Sue to correct the report, or if Joe prepared the report, it was best for Harry to correct the report.  A pilot run of the revised accounting report system showed it generated a stable normal distribution of days to produce the report with a mean of 7 days and a standard deviation of ½ day. 

(5) Control phase: The control phase involves avoiding potential problems with the Xs with risk management and mistake proofing, standardizing successful system revisions, controlling the critical Xs, documenting the revised system, developing a control plan, turning the revised system over to the process owner for continuous turning of the PDSA cycle, disbanding the Six Sigma team, and celebrating the team's success. Risk management involves developing a plan to minimize the risk of increasing variation in cycle time.  Mistake proofing involves installing systems that have a low probability of producing errors in the production of the accounting report, from incoming data to submitted report.

In the accounting report example, team members identify potential problems and methods to avoid them with X1, X3, X7 and X10 using risk management and mistake proofing techniques. For example, they establish procedures to ensure the pairing of clerks and supervisors, and data collection methods to identify and resolve future problems in the reporting system.  The new system is standardized and fully documented in training manuals.  At this point, team members turn the revised system over to the process owner (supervisor of the Accounting Department) and celebrate their success.  The process owner continues to work toward improvement of the revised system beyond its (supervisor of the Accounting Department) current level of output, that being, the distribution of days to produce the report has been improved to have an average of 7 days with a standard deviation of ½ day, and is a predictable normal distribution.  This translates to a report being early or late about once every 24,500 years!  The team chose not to wait around for an error to occur.

Returning the University of Miami auto theft DMAIC project team, the auto theft key indicator (SEC7) can be seen in Figure 2.  It shows the number of auto thefts by month (SEC7) before September 1999 (month 24).  September 1999 (month 24) is the month that a Six Sigma project team implemented a change to the process for patrolling parking lots. The team members determined that the number of auto thefts per month is a stable process. Using Pareto analysis they determined that most auto thefts occur in 2 campus parking lots between 7:00 a.m. and 7:00 p.m.  The Police Chief redeployed the police force to heavily patrol the 2 problematic lots between 7:00 a.m. and 7:00 p.m. in September 1999.  Subsequently, as shown in the line graph in Figure 2, there was a drastic reduction in the number of auto thefts by month (SEC7).

Figure 2: Line Graph Obtained from Minitab of Auto Thefts on Campus Per Month Before and After September 1999

When the project team began, the University was experiencing about 93 auto thefts per year on average. After the project was completed and the findings implemented, the University experiences only 2 to 3 auto thefts annually. Wow!  

Back to top

Micro Model - Six Sigma DMADV Model

The DMADV model is the Design for Six Sigma (DFSS) model used to create major new features of existing products, services, or processes, or to create entirely new products, services, or processes. It has 5 phases, they are: Define, Measure, Analyze, Design, and Verify/Validate. Each phase is briefly described below using an example 6 of designing a new dormitory at University of Miami.

(1) Define phase: The Define Phase of the DMADV model has five components: establishing the background and business case, assessing the risks and benefits of the project, forming the team, developing the project plan, and writing the project objective.

The University of Miami has become a strong, private, doctoral granting university with academic integrity.  Rapid growth in student enrollment, a policy that stipulates that all incoming freshmen must live on campus (unless they live with their family), and the wish of the President for a more residential campus, created more demand than supply for on-campus housing.  A Design for Six Sigma project team was assembled to develop the business case and project charter for the building of a new residential dormitory on campus.

The project charter is: To create a design for a high-class living facility that encourages learning and community (product) aimed at executives-in-residence, MBA students, as well as junior and senior undergraduate business students (market segments) to increase (direction) the number of on campus residents (measure of success) by 280 students (target) by July 15, 2008 (deadline). The project leaders were Scott Widener (Master Black Belt) and Adam Johnson (Black Belt).

(2) Measure phase: The Measure Phase of a Design for Six Sigma project has three steps: segmenting the market, designing and conducting a survey of stakeholder segments, and, using the survey results as Quality Function Deployment inputs to find Critical to Quality Characteristics (CTQs). Team members use Quality Function Deployment matrix to understand the relationships between the needs and wants of stakeholders and the features of the product, service or process design.

In the dormitory example, the Dean of the School of Business Administration identified three distinct market segments for the new on-campus housing.  These market segments are: executives-in-residence, regular MBA students, and junior and senior  undergraduate business students.  Executives-in-residence are individuals that come to campus for one or two weeks to attend a concentrated class.  Currently, no regular MBA students live on campus. Team members developed a survey using the features identified from focus groups. The survey was then completed by a sample of regular MBA and undergraduate business students. Table 3 shows the results of the survey broken down by market segment.

Table3: Kano Survey Results Broken Down by Market Segment



One-Way (O) – User satisfaction is proportional to the performance of the feature; the less performance the less user satisfaction, and the more performance, the more user satisfaction.

Must-Be (M) – User satisfaction is not proportional to the performance of the feature; the less performance, the less user satisfaction, but high performance creates feelings of indifference to the feature.

Attractive (A) – User satisfaction is not proportional to the performance of the feature; low levels of performance create feelings of indifference to the feature, but high levels of performance create feelings of delight to the feature.

Indifferent (I) – User does not care about the feature.

Questionable (Q) – Users response does not make sense (e.g., delighted if feature is present and delighted if feature is absent).

Reverse (R) – User offers responses opposite the expected responses (e.g., “do not like it” if feature is present and “delighted” if feature is absent).

Next team members created a Quality Function Deployment matrix crossing the needs and wants of stakeholders (rows) from the survey with the dormitory features (columns), see table 4.

Table 4: Overall QFD

The bottom of table 4 indicates the importance of each feature (column) to the stakeholder's needs (row) for the dormitory, for example,
simple occupancy rooms is the most important feature (normalized weight = 8.46%).

(3) Analyze phase: The Analyze Phase contains four steps: design generation, design analysis, risk analysis, and model design. The aim of these four steps in the Analyze Phase is to develop high level designs.  In addition to this, the designs will be evaluated using risk analysis.  Finally, nominal (desired) values are established for all CTQs in the Analyze Phase for the “best” design.

     Five room designs were developed in the Analyze phase, they are:

Undergraduate Preferences – Includes only the features that are deemed as “One-Way”, “Attractive”, or “Must-Be” via the undergraduate responses in the survay.

Graduate Preferences – Includes only the features that are deemed as “One-Way”, “Attractive”, or “Must-Be” via the graduate responses in the survay.

Eaton Hall – Includes only the features of the nicest dormitory rooms currently available on campus.

Business Suite – Includes only the features and services that have large contributions to business student education..

Luxury Suite – Includes all of the features that were deemed as “One-Way”, “Attractive”, or “Must-Be” by any of the market segments via the survey.

Note that the five designs do not consider common area designs, just the rooms themselves.  However, all designs will share the same common area design within the building.

The five designs are graded on six criteria determined by project team members through brainstorming using a Pugh Matrix, with Eaton Hall serving as a baseline.  The six criteria are:

Willingness of Customer to Pay More – Luxuries come at a price that must be evaluated with respect to customer price sensitivity.  This information was determined by the survey. 

Low Repair Frequency – This is a general comparison to the baseline that answers the question:  Will this design increase the frequency of needed repairs over that of the baseline?

Ease of Repair – This is a general comparison to the baseline that answers the question:  Will this design introduce CTQs that will unduly burden current employees in repair and maintenance work?

Replacement Frequency – Does the design introduce many CTQs that need yearly replacement?

Ease to Clean and Common Maintenance – Do any of the introduced CTQs that require an inordinate amount of maintenance and cleaning?  As an example of this criterion, fish tanks would score a low grade on this criterion, as they require significant upkeep, whereas plastic plants would score high, as they only require an occasional dusting.

Low Cost/Benefit Ratio – This criterion considers the cost of the design and tries to match the soft benefit of appreciation of current university students and the value as a selling point to future students.

The results led to the realization that the “Graduate Preferences” concept is the best concept, with Undergraduate and Luxury concepts being possible substitutes.

A Risk Analysis revealed seven potentially serious hazards with the “Graduate Preferences” design, they are:

Single Occupancy Rooms – Potential lack of help in disabling circumstances

Kitchenette – Potential Fire

Microwave – Potential Fire

Appliance Rental Service – Potential Fire

Individual Bathrooms – Potential lack of help in disabling circumstances

Full Size Bathtub -- Potential lack of help in disabling circumstances

Accessible Roof – Potential Falls

Finally, a model of the “Graduate Preferences” design was created with Broderbund’s 3D Home Architect 4.0 and is depicted in Figure 3.

Figure 3: Room Layout

(4) Design phase: The Design Phase of a Design for Six Sigma project has three steps: constructing a detailed design of the “best” design from the Analyze Phase; developing and estimating the capabilities of the Critical to Process elements (CTPs) in the design; and preparing a verification plan to enable a smooth transition among all affected departments.

Table 5 shows the features of the final design for the Graduate Preference design.

Table 5: Final Design Features


Finally, a residential floor design was developed, given the constraints placed by the dimensions of the plot of land, and the integration of common areas and other required items into the design, such as stairs, elevators, and trash disposal.

(5) Verify/Validate phase: The intent of the Verify/Validate Phase is to facilitate buy-in of process owners, to design a control and transition plan, and to conclude the DMADV project. 

In the dormitory example, the process owners and all stakeholders were kept intimately involved in the project.  A summarized checklist of the findings of this project was developed and should serve as a guide for the engineers and architects who will further develop the project. All bids must include historical process capabilities of the bidding parties.

All bids must include historical process capabilities of the bidding parties.  These process capabilities may include:

 defects per units constructed, 
 timeliness of deliveries, 
 timeliness of construction, 
 defects per units fabricated, 
 rework time per initial man hours invested, and 
 fines per project that result from construction regulation violations. 

A preventive maintenance system per manufacturer recommendations must be implemented after construction. Occupancy indicator control charts must also be implemented, they include:

number of applications per semester,

percentage of vacant rooms by semester,

number of students on a waiting list by semester, and

time between residents by semester.

The final part of the Verify Phase is to maintain communication between the champion and the process owner.  These lines of communication will alleviate any confusion or other unforeseen problems that will inevitably develop.  It will ensure that the conceptual design is not compromised by outside forces and neglect.  

Back to top

Micro Model - Lean Management

Lean management is a management style that promotes reducing waste through the elimination of non-value added activities for example, unnecessary 
(for example, unnecessary complexity)
, eliminating work in process and inventory, and increasing productive flexibility and speed of employees and equipment.  Non-value added activities in a process include any step that: (1) customers are not willing to pay for, (2) do not change the product or service, (3) contain errors, defects, or omissions, (4) require preparation or setup, (5) involve control or inspection, (6) involve over-production, special processing and inventory, or (6) involve waiting and delays. Value added activities include steps that customers are willing pay for because they positively change the product or service in the view of the customer.  Lean management has several components, they are: Value Streams, Total Productive Maintenance (TPM), Quick Changeovers, Poka-Yoke, and the 5Ss. Each component is explained briefly below. An example of lean thinking using the 5Ss follows.

A value stream describes all the value added and non-value added process steps and decisions necessary to move a product or service from supplier to customer. These steps include design and redesign, raw material flows, sub-component flows, information flows, production and service flows, and people flows, to name a few steps. A lean value stream is a value stream in which an upstream process only makes what a downstream process needs, when it needs it. There is no (or little) inventory. Takt time indicates the pace at which every step in the process must produce one unit of output to meet customer demand per time period, for example, per shift. Continuous flow refers to a process with a batch size equal to one.  Each unit passes immediately from step to step without any waiting time in between steps.  Processes with a batch size of one and no waiting between the steps in the process is the “holy grail” of production and service. A pull system initiates production/service in a given step in a process using a request from the next downstream step in the process.

Total Productive Maintenance (TPM) is a theory useful for maintaining plants and equipment with total involvement from all employees. Its objectives are to dramatically increase production and employee morale by: (1) decreasing waste, (2) reducing costs, (3) decreasing batch sizes, (4) increasing production velocity, and (5) increasing quality.

Quick changeover (Single Minute Exchange of Die – SMED) is a technique that team members can use to analyze, and then reduce: (1) the time it takes to setup equipment (including tools and dies) and people (for example, shift to shift setup for cashiers in a supermarket), (2) the resources required for a changeover, and (3) the materials necessary for a changeover. It creates the opportunity in a value stream to effectively and efficiently institute small batch sizes, or even one-piece flows. One piece flows occur when one unit at a time flows through a value stream, as opposed to batch flows when a batch of units flows through a value stream.  Internal activities are the steps in a process that can only be performed while the process is idle. External activities are the steps in a process that can be performed while the process is in operating mode.                   

Poka-yoke (pronounced POH-kah YOH-kay) is Japanese for mistake-proofing devices. These devices are used to prevent the causes of defects and/or defective output (called errors), or to inexpensively inspect each item that is produced to determine whether it is conforming or defective. A poka-yoke device is any mechanism that prevents a mistake from being made or makes the mistake obvious at a glance.

The 5Ss form a system for tidying up and maintaining a process. Each of the 5Ss is discussed below. Seiri means throwing away unnecessary things, and putting the remaining necessary  “things” in order, that is, organizing “things” using specific rules. Once an employee has internalized the rules for throwing away unnecessary things and for organizing necessary “things,” he or she will quickly be able to find "things". Seiton means tidily putting “things” in their proper place which is determined with seiri. Putting things away requires following three rules: (1) deciding where things belong, (2) deciding how things should be put away, and (3) deciding when things should be put away. Following the put-away rules leaves "things" where they can be quickly found next time they are needed. Seiso is an attitude that considers a dirty and untidy work place intolerable.There are three broad levels of cleaning. First, there is the overall cleaning of everything. Secondly, there is the cleaning of specific items, tools, machines and workplaces. Thirdly, there is the cleaning at the detail level, getting to grime in screw threads, corners and crevices. Seiketsu is visual management. Visual management leverages location, distance, shape, brightness, color, and contrast so that something stands out when we are looking for it. Visual controls include work instructions, hazard warnings, indicators of where things are kept, equipment and tool designations, cautions and reminders, and indicators and plans of what happens when. Whenever people need reminding, a visual control should be there to help them. Shitsuke draws together the other four Ss ensuring they are used properly. People make mistakes, forget, and state things incorrectly. We also get stuck in habits which are not helpful with our work. Habits are, however, very useful things, and if we can align them with the work disciplines of the 5Ss, we can forge them into a complete disciplined approach to management training is an important method of changing how people think and act.

An example of Lean thinking using the 5Ss. A company produces plastic cups for packaging food products. A thin film of plastic is produced by an extruder which goes into a thermoforming machine where cups take form. After thermoforming, a cardboard sleeve is added to the cup to make it strong enough to hold the food stuff and keep its form. The company was experiencing problems in their production process. Variation in the thickness of the plastic film from the extrusion process caused problems in the thermoforming process. A plastic film that is too thin will “pop” when a vacuum is created in the thermoforming mold. This situation forces the company to produce a thicker plastic film (waste of raw materials) to avoid constant stops of the thermoforming machine (waste of time). The Green Belt project leader was Dario Prevecarius.

(1) Define phase: Six Sigma team members prepared a business case for the project which results in a S.M.A.R.T. project objective. S.M.A.R.T. is an acronym for Specific, Measurable, Actionable, Results Oriented, and Time Bound. The project objective is to reduce (direction) the CDSpread (measure of variation) of plastic films produced (process) to less than 0.03 (target) by June 30, 2006 (deadline).

(2) Measure Phase: Team members operationally define the CDSpread (CTQ) of plastic film. It is defined below.

Criteria: Reset the full spectrum machine before producing a plastic roll. After the roll has been produced, print the Summarized Report from the full spectrum machine. The report contains the CDSpread for the roll.

Test: Record the CDSpread from the summarized or electronic report.

Decision: If CDSpread ≥ 0.03, then the roll is considered to be regular quality. If CDSpread < 0.03, then the roll is considered to be high quality.

Additionally, team members conduct a measurement systems analysis and collect baseline data for the CTQ. Baseline data was drawn from 15 successive rolls of plastic. The distribution of CDSpread approximated a normal distribution. Figure 4 shows a dot plot of the CDSpread with an average CDSpread = 0.06, and 0 rolls having CDSpread < 0.03.

Figure 4: Dotplot of CDSpread

Figure 5 is an individual and moving range chart of the CDSpread data. It shows the CDSpread is out of statistical control (2 out of 3 points in zone A or beyond), and hence, not stable and predictable.

Figure 5: I-MR chart of CDSpread


(3) Analyze phase: Team Members created a flowchart (see figure 6) of the plastic roll extrusion process with the Xs that potentially can affect the stability, shape, spread and center of CDSpread.

Figure 6: Process Map of Plastic Roll Production Process

X1 = Employee is properly trained in standard methods (yes, no)

X2 = Information tag is complete (complete, incomplete)

X3 = Information tag reliable (reliable, unreliable)

X4 = Request for production format (Errors, No-Errors)

Next, the team members performed a Failure Modes and Effects Analysis (FMEA) to identify the high risk Xs that have a high probability of effecting CDSpread. From the FMEA, the Xs with the highest Risk Priority Number are:

X1 = Employee is trained (yes, no)

X2 = Information tag is complete (complete, incomplete)

X3 = Information tag is reliable (reliable, unreliable)

Team members developed operational definitions for X1, X2,and X3, conducted measurement systems analyses for the Xs, and collected baseline data for each high risk X. Finally, they developed hypotheses concerning the relationships between for the Xs the high risk Xs and the CTQ (CDSpread), they are:

Properly training employees will decrease the number of mistakes, not only in the inventory department, but also on the extrusion department. Once employees realize how crucial the inventory keeping process is (quality in, quality out), they will act more carefully, production will be smoother, and CDSpread will decrease.

Complete tags will help employees in the warehouse to sort the materials that will go into the extruder, and to decrease the number of mistakes in the proportion of new and recycled materials, and consequently, decrease CDSpread.

Accurate information in the tags will help employees in the warehouse to identify the materials that will go into the extruder, and to decrease the number of mistakes by mixing two different grades of materials, and consequently, decrease CDSpread.

(D) Improve phase. Due to the characteristics of the plant, it was difficult to run an experiment between the Xs and CDSpread. The general manager felt that it was not reasonable to mix different raw materials and/or in suboptimal percentages on purpose to experiment on the effect on CDSpread. Expense and scheduling were given as reasons for not conducting experiments. Team members hypothesized that the application of the “5S” techniques would have a very high likelihood of setting the optimal levels of the Xs, and consequently, optimizing CDSpread. Before and after pictures are shown in figures 7 through 12.

Figures 7 through 12: Before and After Pictures Using the “5S”s”






Team members created a revised flowchart for the process which incorporates the 5Ss. It is shown in figure 13.

Figure 13: Revised Flowchart

Next, team members pilot tested the revised process. Figure 14 shows that the 5S technique was implemented in the week of April 10, 2006. Everything was ready on April 14th, and on April 16th an out of control point appeared in the moving range chart (data point 15). This is a sign that there was a change in the process output, directly related to the 5S techniques. 


Figure 14: Before and After I-MR Chart for CDSpread


 As you can see, the process was dramatically improved by application of the 5S techniques.

(5) Control phase: Team members established a risk abatement plan to decrease the risk of failure of the revised process that included control guidelines for the process owner and present the benefits and costs of the revised process. Finally, table 6 shows the costs and benefits of the project.

Table 6: Costs and Benefits of the Project





$100,000 annual savings on raw material

$1500 (paint, new tags, barcode inventory system)


Reduce arguing between the extrusion and thermoforming departments.

Time spent on the project


The macro (dashboard) model and the micro model (DMAIC, DMADV or LEAN projects) both require data based management if they are to work effectively in an organization. Data based management requires that employees report the actual data about their outputs (key indicators). If a punitive management style, such as Management by Objectives (M.B.O.) is being used to mange the employees in an organization, then the employees may distort the data about their outputs to avoid negative performance reviews. For this reason, it is imperative that the macro model and the micro model are used in a positive and nurturing managerial environment. Such an environment was described by Dr. W. Edwards Deming and is briefly discussed below.

Back to top

Management Style - System of Profound Knowledge (Dr. Deming)

The management model is the management style required to properly utilize the macro and micro models. It was developed by Dr. W. Edwards Deming and is called the “System of Profound Knowledge.” Its aim is to “promote joy in work. Joy in work is the vehicle by which management can create a positive and nurturing environment in which employees can perform their jobs, and have the energy and desire for improving or innovating their jobs. Deming’s theory of management is based on four paradigms, or belief systems, that an individual or group uses to interpret data about conditions and circumstances.  You can think of each of Deming’s paradigms as a shift in assumptions for the practice of management, designed to create the environment required to promote joy in work, and hence, release the power contained in intrinsic motivation.

Paradigm 1. People are best inspired by a mix of intrinsic and extrinsic motivation, not only by extrinsic motivation. Intrinsic motivation comes from the sheer joy of performing an act. It releases human energy that can be focused into improvement and innovation of a system. It is management's responsibility to create an atmosphere that fosters intrinsic motivation.  This atmosphere is a basic element of Deming's theory of management.  Extrinsic motivation comes from the desire for reward or the fear of punishment. It restricts the release of energy from intrinsic motivation by judging, policing, and destroying the individual. Management based on extrinsic motivation will "squeeze out from an individual, over his lifetime, his innate intrinsic motivation, self-esteem, dignity, and build into him fear, self-defense."

Paradigm 2. Manage using both a process and results orientation, not only a results orientation. Management's job is to create an environment in which employees have the direction and energy to use the DMAIC model, the DMADV model, and Lean thinking to improve and innovate the processes that create results, not just to manage results. This paradigm shift allows management to define the capabilities of processes, and consequently, to predict and plan the future of a system to achieve organizational optimization. This type of optimization requires that managers make decisions based on facts, not on guesswork and opinion. It is critical that top management change the culture of their organization from “management by guts” (called KKD in Japan) to “management by data.”  It is easy to refute an argument based on guesswork or opinion, but it is difficult to refute an argument based on solid, scientific data. Managers must consider visible figures, as well as unknown and unknowable figures (for example, the cost of an unhappy customer or the benefit of a prideful employee).

Paradigm 3. Management's function is to optimize the entire system so that everyone wins, not to maximize only their component of the system. Managers must understand that individuals, organizations, and systems of organizations are interdependent. Optimization of one component may cause sub-optimization of another component. Management's job is to optimize the entire system towards its aim.  This may require the managers of one or more components of a system to knowingly sub-optimize their component of the system to optimize the entire system.

Paradigm 4. Cooperation works better than competition, if the aim of the system is not to win. In a cooperative environment, everybody wins. Customers win products and services they can brag about. The firm wins returns for investors and secure jobs for employees. Suppliers win long-term customers for their products. The community wins an excellent corporate citizen.

In a competitive environment, most people lose. The costs resulting from competition are huge.  They include the costs of rework, waste, and redundancy, as well as the costs for warranty, retesting, re-inspection, customer dissatisfaction, schedule disruptions, and destruction of the individual's joy in work. Individuals and organizations cannot reap the benefits of a win-win point of view when they are forced to compete.

Is competition ever the preferred paradigm?  The answer is “yes,” if and only if the aim of the system is to win.  If the aim of the system is anything other than to win, for example to improve or have fun, then competition is not the preferred paradigm.  Cooperation is the preferred paradigm in all systems with non-competitive aims.

In my opinion, if managers practice these four paradigms, they will reap enormous benefits.

In summary, the purpose of Six Sigma management is to “promote joy in work” for all employees so that they have the energy to participate in the improvement and innovation projects identified from the organizational dashboard!!!  

Back to top