2. What do basic statistics (R, S, V, s) physically mean?

3. How to Use Causes and Effects in The Right Ways

4. Maintenance and Improvement

5. Quality of Design and Quality of Conformance

6. Quality Control and Quality Assurance

7. Management by facts

8. Designing Approach and Analytic Approach

9. Seven Tools for Quality Control

10. The Four Principles of Brainstorming

1. Concept of "Management"

In TQM, "Management" is to turn PDCA Cycles based on facts.

P signifies "Plan", D "Do", C "Check", and A "Act".

Everybody is unconsciously applying PDCA Cycles daily. The most important point is to turn PDCA cycles based on facts.

For example, let's suppose you will go on a trip. P: You plan your trip. D: You go to the trip. C: You check your trip plan while you enjoy. A: If you are not satisfied with your plan, You will implement corrective actions for the next trip.

Every person turns PDCA cycles in his/her daily life but it is hardly done based on facts (data). Management requires you to record facts, keep the record and utilize the record whenever necessary. This is "Management".

A PDCA cycle will not move forward if it does not have any ground to touch. Facts (Data) play the role of this ground for PDCA Cycle. On this ground PDCA Cycle can turn and move forward.

2. What do basic statistics (R, S, V, s) physically mean?

Facts (data) are essential to carry out management activities. Data can be classified as qualitative data or quantitative data. Qualitative data is numerical. It is used more frequently than the other.

From original data you can receive some measures. Ones represent central tendency of distribution like the average. Others represent spread of distribution like the standard deviation.

The range (R), the sum of squares (S), the variance (V) and the standard deviation (s) are the most known measures of spread.

Range (R) is difference between the lowest and the highest data. The larger difference, the larger spread. The smaller difference, the smaller spread. Only two data are used when the range is calculated. It doesn't matter how many data you collected.

Because of this, another measure of spread was generated. All the collected data are used for this measure. The average of data and the differences between each data and average are used for the calculation. The difference of data and average is the "variation". As the sum of variations becomes larger, the spread of the distribution becomes larger. The smaller sum represents the smaller variation.

But the sum of all variations always becomes zero (0). Therefore the variations are squared and sum of the squared areas is used instead o the sum of variations. The larger areas, the larger variation and, of course, the smaller the smaller.

The sum of squared variations is the measure called the sum of squares (S). The sum of squares is the sum of squared differences between each original datum and the average of data. Variations are squared and summed to calculate the measure S. The S is used to dissolve and study the variation of the distribution.

The measure S changes according to the number of data (sample size). The more data brings the bigger S. The less data ends up with the smaller S. Because of it one S can not be simply compared with another.

Therefore the average of squares is used. The average of squares can be received by dividing the measure S by the number of data. This is the variance (V). The V can be used to compare and study more than two variations.

Both of the variance and the sum of squares are the squared measures. Their units are squared units, too. For example, if the unit of original data is meter (m), the units of the variance and the sum of squares are squared meter (m²). By extracting the square root of the variance, a measure of spread with the same unit as the original data is received.

This is the standard deviation (s). This is why you extract the square root of the variance to get the standard deviation.

3. How to Use Causes and Effects in The Right Ways

To better your management, it is very important to understand the relationship between a cause and its result. Treatment actions are "A" of PDCA. There are two kinds of treatment action: Corrective Actions and Preventive Actions.

Corrective Actions are the treatments after the occurrence of something. Preventive Actions are often explained by a very famous proverb: Look before you leap.

In summer, mosquitoes show up and bite people. It is a corrective action to apply ointment to the bitten part after being bitten by a mosquito. It is a preventive action to apply the insect repellent. And avoid mosquito bites. Look before you leap! The more effective prevention would be cleaning of ditches, for example. Then mosquitoes will not appear around you.

Here, "being bitten by mosquitoes" is a result. Why mosquitoes bite you? It's because mosquitoes appear and find you. Appearance of mosquitoes is the cause. If the cause "Appearance of mosquitoes" does not exist, the result "Bitten by a mosquito" can not happen. If the cause "Appearance of mosquitoes" can be removed, the bad result "Bitten by a mosquito" can be eliminated.

"Cause and Effect Diagram" is a tool to arrange relationship of causes and an effect. It is a component of the Seven Tools for Quality Control. One of its most important roles is to visualize all already-known causes in a diagram. Please pay attention to visualize all the already-known causes in a figure when you draw a Cause and Effect Diagram. It is also very essential to pursue the causes up to the root-cause level. Root causes should be something that you can take actions directly or give effective treatments. Sometimes I hear people say "We already know the causes. But we don't know what to do or how to give necessary treatments!" In this case, the root causes have NOT been detected AT ALL. If you know the root causes, you should know what to do or how to do. You pursue root causes up to the level that you can certainly take the actions of treatment. Knowing of necessary things to do proves the truth that you finally detected the root causes.

Brainstorming is an effective method to pursue causes. Please apply brainstorming and pursue the root causes up to the level that you can take actions of treatment.

4. Maintenance and Improvement

Management has two faces. One is Maintenance and the other is Improvement. Maintenance Management Activity is to keep your management at the defined level. Improvement Management Activity is to higher your management up to the better level. Both of these two activities are Management Activities and they require you to turn PDCA cycles.

Maintenance Activity is to maintain your management at a certain level. If a result jumps out from the established level, the necessary treatments should be carried out. Causes of abnormality should be pursued, detected and eliminated. On this purpose, "Control Charts" of the Seven Tools for Quality Control are applied. The Control Charts for Maintenance are used on this purpose.

On the Control Chart for Maintenance, the current management level and the control limits are determined. The control limits show the highest and/or lowest control level and hazard necessity of treatment actions. You need to proceed treatment actions immediately when a plot goes out from a control limit.

Standardization of operations is an very important issue for the maintenance of the quality level. The concept of SDCA cycle is known very well: Standardize (S), Do (D), Check (C), and Act (A).

S: Standardize operations, D: Implement standards, C: Check if standards are correctly implemented, and A: Proceed treatment actions if standards are not correctly implemented.

On the contrary, Improvement Activity is to make current management better, bring it up to a better level. A goal is always established. Establishment of a goal is to show at which level you are targeting. After the goal establishment, you turn a PDCA Cycle.

Generate improvement propositions to attain your goal, and plan necessary activities to realize your propositions (P: Plan). Implement planned activities (D: Do). Check if your implementation was successfully proceeded and your goal was attained as you expected (C: Check). When your goal was not attained, you need to take treatment actions (A: Act).

After finishing an improvement activity and receiving good results as expected, the new process is adapted as your new standard. The new process is added to the existing standard or the existing standard is substituted with the new standard. And then, you finish the improvement activity and shift to the maintenance activity. You arrive at the end of a phase, where you finish the improvement activity and start a new maintenance activity.

Maintenance Activity and Improvement Activity should be proceeded by turns. You must avoid proceeding only Improvement Activities continuously. Without Maintenance Activities your improvements will not be standardized and you will end up with a tiring situation like you always continue to tackle with the same problem.

Maintenance Activity and Improvement Activity are inconsistent with each other. Improvement Activity is the activity to improve or modify the standards. Maintenance Activity is the activity to maintain the established standards. Please remember: It is very important to repeat these two inconsistent activities alternately and improve you management step by step.

5. Quality of Design and Quality of Conformance

One of the main emphases of Management Activity is Quality. Quality has been defined in many ways by many people or organizations. But two of them can be considered as the most basic definitions. One is the definition of "Quality of Design". It is "Quality targeted as the goal of manufacturing (JIS Z 8101)". The other is the definition of "Quality of Conformance". It is "Actual Quality of the product manufactured aiming to attain the targeted Quality of Design (JIS Z 8101)".

Quality of Design is described on designs, specification sheets and so on. Quality of Conformance is sometimes called "Quality of Manufacturing". Manufactured products are measured and inspected their conformity with their Quality of Design. No variance exists on Quality of Design but Quality of Conformance always has variance.

Let's study an example of production and sales of ballpoint pens. Many measurements are on the design such as the length of penholder, the diameter of penholder, the length of cap, the diameter of cap, etc. On the design, these measurements are invariable. Necessary equipment and tools such as molds are ready and products, in this example they are ballpoint pens, are manufactured. The penholders and the caps are measured. There are some products which show the exact designed measurements but others are different from the design. Means variances exist among the products. The products are approved as conformity only when they are produced nearly as same as the design and the inspection results show the satisfaction of the required specifications.

Quality of Conformance is conformity of a product with its Quality of Design (Quality of Manufacturing). A product will be approved as conformity when it is conformable with its Quality of Design. However, Passing of specification does not prove "Good Quality", especially when the range of specification is wide. For example, the specification of the penholder length of a ballpoint pen is 135.0�E�A 5.0mm according to its design. Products are manufactured by following the specification. The penholder length will be various. The shortest will be 130.0 mm and the longest 140.0 mm. Imagine what will happen if the specification of the cap also has the similar tolerance. Some caps will be very loose and others will be too small and won't fit with the penholders.

Of course, Quality of Conformance must satisfy Quality of Design. However, the most important thing is to reduce variance of Quality of Conformance. The smaller variance on Quality of Manufacturing proves the better Quality!

In addition, It must be examined carefully if Quality of Design is adequate. Variance does not exist on Quality of Design. But if the targeted Quality of Design is poor at the beginning, customers will never compliment your product quality, even though the variance of Manufacturing Quality is minimum as possible.

For instance, the length or the diameter of a penholder have to be "easy to write" or "easy to hold". The specification of the length and the diameter must be defined presupposing the penholder must be "easy to write" and "easy to hold". "Tireless" may be another precondition of its design.

Designs always have to be able to satisfy customer requirements. If a product is designed ignoring customer requirements, the market will never accept the manufactured product even with excellent Manufacturing Quality. It is very important to better Quality of Design.

In Quality Function Deployment (QFD), an effective system has been proposed to complete the above aim: Seize Customer requirements as Required Quality, establish Planned Quality for Required Quality, and plan Design Quality which satisfy customer requirements.

Quality Control (QC) was imported to Japan in 1950's. At that period, The term "Quality" meant only "Quality of Design" and "Quality of Conformance". Later, the original Quality Control was arranged and improved, and Japanese Quality Management was developed. During the development process, concepts of Planned Quality and Required Quality were born, and quality. These two new concepts deeply contributed for the improvement of Japanese product Quality.

6. Quality Control/Management and Quality Assurance

What does "Quality Control/Management" mean? Which preposition will be put between "Control/Management" and "Quality"? Control "of" Quality? Control "by" Quality? Or, Control "through" Quality? Once people in Japan argued which preposition is most suitable.

Nowadays, the majority understands "Control through Quality" is the best interpretation of the Japanese term Hinshitsu Kanri. Well then, what is controlled through Quality? The answer is "Process". Process is controlled through Quality. Each process to produce product or service must be controlled. On this purpose, samples are collected in a process, necessary data are gathered by measuring the picked samples, and necessary actions are taken toward the process.

According to the information of the process, it will be judged if the actual process can be maintained or treatment actions are required. If the result shows the necessity of treatment, then, some actions are taken. This procedure will allow you to act for the future. Products are the output of the process. Products are the results. They are the output of the past. Examination (or inspection) of the output from the past and removing of the defects is the action toward the past. The action taken toward the process is avoiding of the failure/defect occurrence. This is the action for the future.

In the meantime, what is Quality Assurance? The interpretation "Assurance of Quality" may sound like an action or an activity. However, for better understanding, I recommend to interpret Quality Assurance as an objective. Quality Control activities must be done under the objective "to assure Quality".

According to the definition by Japan Industrial Standards (JIS) Quality Control is "A system of means to practically produce product or service which satisfy customer requirements (JIS Z 8101)". The complete definition continues more. On the other hand, Quality Assurance is defined as "Systematic activities which producer carries out to assure the satisfaction of quality required by customers (JIS Z 8101)". As you see, Quality Control is interpreted as a "system" and Quality Assurance an "activity".

But, in my opinion, Quality Assurance is an "objective". Quality Control is the activity to accomplish the objective. What do you think?

7. Management by facts

As mentioned above, Management is to turn PDCA cycle based on facts (data). Then, how should a fact be presented? What kind of data can be treated as a fact?

Data can be stratified as quantitative data or qualitative data.

Quantitative data can be subdivided into two categories, attribute and variable.

In general, people tend to think that the numerical data represents a fact. If a person is requested to show data, he/she imagines the presentation of something numerical. In fact, a fact can be shown as numerical data that are measured or counted in most of the cases. Why numerical data can be recognized as a fact?

Subjectivity of a person or a group should be eliminated for the presentation of a fact. Representative of a fact should be objective.

Adding to this, representative of a fact must be provable. When something is objective and provable, it is the representative of a fact. Linguistic data can also be the representative of a fact if it is objective and provable.

"This car is wonderful." This is a linguistic data, may be a fact, and may be not. "This car is 4970 mm length." This is another linguistic data. The length of the car can be measured again. This is provable and objective. "Wonderful" is a subjective expression. "Wonderful" does not mean the same to everybody. It is not possible to prove it.

If a linguistic data is provable and objective, it can be handled as qualitative data.

Management has to be carried out based on facts and careful attentions should be paid on the matters above.

8. Designing Approach and Analytic Approach

Grasp the extent of a problem, pursue its root causes and eliminate the detected root causes. This is the Analytic Approach, a way to conduct management activity. There is another way to direct. It's the Designing Approach: Define a should-be figure, plan how to realize it and execute the plan.

The Analytic Approach follows the procedure below:

Conduct Brainstorming, summarize the information collected e.g. in a Cause and Effect Diagram, define the potential root causes, analyze the listed root causes and prove the truly responsible root causes for the extent of a problem. This problem solving procedure is for Maintenance and Improvement.

In the Designing Approach, at first, a should-be figure has to be defined. Quality Function Deployment (QFD) is known as a typical Designing Approach. In QFD customer requirements are grasped as Required Quality. Customers want a company to realize Required Quality as they expect. The level that customers expect is recognized as the should-be figure. Considering the should-be figure, a question is thought through. What or can satisfy customer requirements, and how?

In the past, more problem-solving-type approach was applied in TQM. This is the Analytic Approach. Introduction of Hoshin Kanri and other advanced management concepts gave the birth of target-attaining-type approach. This is the Designing Approach.

Hoshin Kanri (Policy Management) requires the Designing Approach, not the Analytic Approach. Because top management defines policies such as "to be like this until when" and the policies are the should-be figures of the company.

Of course it is possible to apply the problem-solving-type approach to eliminate the gap between a should-be figure and the current condition and accomplish maintenance or improvement. It is because the gap can be considered as an existing problem. Nevertheless, it will be much more effective to apply the most adequate tool at the right place.

9. Seven Tools for Quality Control

Many tools have been proposed to support Management Activities effectively. The Seven Tools for Quality Control is a set of tools to assist the effective implementation of Quality Control activity. It contains (1) Cause and Effect Diagram, (2) Check List, (3) Histogram, (4) Pareto Diagram, (5) Graphs (including Control Charts), (6) Scatter Diagram, and (7) Stratification.

There is an opinion says that Stratification is a way of thinking and not a tool. In this case, Stratification is omitted and substituted by Control Charts. Control Charts are considered separately from other graphs. Another opinion says both Pareto Diagram and Scatter Diagram must be included in Graphs. Anyway, the most general Seven Tools are the ones listed above.

Cause and Effect Diagram was developed by Dr. Kaoru Ishikawa. It is also called Ishikawa Diagram or Fishbone Diagram because of its shape like a fishbone. It is used to show the relationships of an effect and its causes. The effect should be the result of a process. The causes should be the root causes most strongly influential to the effect. The causes are chosen out of many which bring the result. A line is drawn in the middle of the diagram. It is often compared with the backbone of a fish. This line represents the process. When they are added, the causes appear on the process (the fish backbone) as linguistic data.

Check List is a tool applied at the phase of data collection. There are various kinds. One Check List is designed to check the cells correspond with the data categories. You do not need to write each numerical data in figures. One Check List contains drawings or designs to check locations of failure occurrences. There are many others. Please use the most adequate Check List when you collect necessary data.

Histogram is a kind of bar charts. It is applied to view the summary of the collected data, especially when the data is continuous figures. Collected data are divided into 10 to 13 classes. Number of data in each class is counted and summarized on a bar chart. By adding specification limits and other necessary information, you can have a whole view of the process condition (e.g. performance of finished products comparing their specification).

Pareto Diagram was developed by Pareto, an Italian Economist. It is applied to determine priorities. 80-20 principle is the famous Pareto princeple. 80 % problems can be solved by tackling with only 20 % items out of all. Pareto Diagram consists of two graphs. One is bar chart. The items are arranged in the order of the number of problems or the amount of money. Another graph is line chart of accumulations of bars. The bars show the number of every problem and the line shows the transition of the accumulated number of problems. If the bars show the amount of money, the line shows the transition of the accumulated amount of money. This line chart is often called "Pareto Curve". The best proportion of X-axis to Y-axis is 1:2 for easier analysis of Pareto Diagram. In Industrial Engineering field, Pareto Diagram is called ABC Analysis.

There are various graphs such as bar chart, line chart, pie chart, radar chart, etc. Chronological data will fit on a line chat. Bar chart or pie chart is suitable to show the proportions because they are area graphs. Rader chart is often applied to show the balance of several items. To make other people understand easier, the most adequate graph should be adapted.

Control Charts are applied when the stability of a manufacturing process is evaluated based on the chronological data of the process. Control Charts for Control and Control Charts for Analysis are used differently according to the purpose of use. There are various types of Control Charts. The suitable type of Control Chart is chosen and used depending on the type of data. Control Charts in Japan follows the concepts of Shewhart. They are founded on the concept of three sigma limits.

Scatter Diagram is a figure applied to visualize the relation of two variables. One variable is scaled on X-axis and another on Y-axis. The data is plotted on the space consisted by X-axis and Y-axis. The condition of the plots shows the relation of two variables such as "the bigger variable X results the bigger variable Y."

Stratification is stratifying or categorizing of data. Each category has to have a meaning. What is a meaningful category? Can you imagine proper direct actions to each category? Then the categories are meaningful. Categories can be determined by equipment, by operator, by operation method or by raw material. Data are stratified into some groups according to their common characteristic.

These tools must be used in daily works and managed well. Knowing the tools is not enough. There is a big difference between knowing tools and being capable to use tools. Please move a step forward, and handle the tools perfectly!

10. The Four Principles of Brainstorming

Brainstorming is one of the idea generating methods. It was developed by Osborn with his research group. There is a proverb "Two heads are better than one." When you solve a problem or accomplish a task, it has better not do it by yourself. Brainstorming is a method for a group of people to gather everyone's ideas and find the best solution.

Gathering of ideas is not much different from having of a chat if there is not any rule to follow. If you follow the Four Principles of Brainstorming, the result of Brainstorming will be more effective.

The four Principles of Brainstorming are "Don't criticize", " Be free to comment ", " Quantity, not Quality" and "Follow the lead of another".

When Brainstorming is operated, all the members must follow the four principles above and everybody storms his/her own brain to generate excellent ideas.

But there are some people more influential than others in any kind of group formed in company. These influential people sometimes restrict others and the atmosphere becomes difficult to follow the second principle "Be free to comment." To avoid it happens there is a method called "Brain Writing". Ideas or opinions are written down on a piece of paper, not spoken. The 635 method is a typical Brain Writing.

The 635 method requires 6 people. Each member generates and writes down 3 ideas in every 5 minutes. The procedure of the 635 method is as follows:

Prepare six sheets of A4-sized paper. Each sheet will have a table of two columns nine lines. Pass one sheet to each person. Confirm the problem to be solved or the task to be accomplished with other members. Write down 3 ideas on the sheet in 5 minutes. After passing 5 minutes, round the sheets to the next person clockwise. Write down 3 more ideas on the sheet adding to another person's ideas in 5 minutes. Continue to round the sheets and write down ideas until all the cells are filled up. By doing this, the group can generate 108 ideas in 30 minutes.

Is it true?????

Besides this part, there are many others: [TQM Tips - Part 2: Quality Function Deployment], [TQM Tips - Part 3: Services], [TQM Tips - Part 4: New Product Development], etc.