Saturday, June 3, 2017

Total Quality Management - Six Sigma Methodology

Critical To Quality (CTQ):
CTQs are the internal critical quality parameters that relate to the wants and needs of the customer. They are not the same as CTCs (Critical to Customer), and the two are often confused.
CTCs are what is important to the customer; CTQs are what’s important to the quality of the process or service to ensure the things that are important to the customer.
A quality function deployment (QFD) or CTQ tree relates the CTQs to the CTCs. For instance, car door sound when closing might be a CTC, while the dimensional tolerances and cushioning needed to produce those conditions are CTQs for the auto maker.
CTQ trees:
(Critical-To-Quality trees) are the key measurable characteristics of a product or process whose performance standards or specification limits must be met in order to satisfy the customer. They align improvement or design efforts with customer requirements.
CTQs are used to decompose broad customer requirements into more easily quantified elements. CTQ trees are often used as part of six sigma methodology to help prioritize such requirements.
CTQs represent the product or service characteristics as defined by the customer/user. Customers may be surveyed to elicit quality, service and performance data. They may include upper and lower specification limits or any other factors. A CTQ must be an actionable, quantitative business specification.
CTQs reflect the expressed needs of the customer. The CTQ practitioner converts them to measurable terms using tools such as DFMEA. Services and products are typically not monolithic. They must be decomposed into constituent elements (tasks in the cases of services).
Defining CTQ outputs: A key Step in the Design Process:
After starting a project and gathering the voice of the customer (VOC), it is time to define the critical-to-quality outputs (CTQs). 
CTQs are the key measurable characteristics of a product or process whose performance standards or specification limits must be met in order to satisfy the customer. These outputs represent the product or service characteristics defined by the customer (internal or external). They may include the upper and lower specification limits or any other factors related to the product or service. 
Typically, a CTQ must be interpreted from a qualitative customer statement to an actionable, quantitative business specification. Establishing CTQs is vital for a company to meet customer needs and keep up with the competition.
VOC becomes CTQs:
The flowchart in Figure 1 provides an overview of the requirements necessary to translate the VOC into usable CTQs. Operational definitions of the flowchart steps are:
  • Characteristics of product or service: A word or phrase that describes some aspect of the product or service. Example: dry cleaning process time.
  • Measures and operational definitions: A definition of how the product or service’s characteristic is to be quantified. There may be several ways of quantifying a given characteristic. Example: the unit used to measure time between when the cleaner receives clothes and when the clothes are ready for pickup (hours).
  • Target value(s): The aim for a product or service. If there were no variation in the product or service, this is the value that would always be achieved; it is the desired level of performance. Example: clothes ready for pickup in 24 hours.
  • Specification limits: How much variation is the customer willing to tolerate in the delivery of the product or service? Specification limits are performance limits that are acceptable to the customer. Example: Upper specification limit for dry cleaning process time is 28 hours.
  • Defect rate(s): This is how often the producer is willing to produce a product or service outside the specification limits. Example: 3.4 defects per million opportunities.
Data Quality and CTQs:

Although it is often overlooked, data quality is an important consideration in the design effort. The impact of poor data quality can be very serious. From an organizational perspective, it may create extra costs, rework, low productivity; drive the “wrong” decisions (because of outdated data); and prompt a sense of frustration or lack of trust. From a project perspective, it could result in project delays and impairment of testing. Project teams need to assure that data associated with their designs is both accurate and complete. This may be accomplished by defining CTQs for data quality.
Possible data quality CTQs include:
  • Access restriction
  • Age
  • Availability
  • Completeness
  • Definition and format
  • Encryption
  • Timeliness
Types of Data:
Data can be discrete or continuous. When possible, practitioners should collect continuous data because it can be recorded at many different points. Examples include length, size, time, temperature and cost. Continuous data can be broken down into smaller parts, meaning practitioners can get more information about what they are measuring than from attribute data.

Setting Measurements:
The design of a product or service starts with quantified requirements. Practitioners need to develop measures for which targets and limits can be established. There may be several ways to quantify a given characteristic. Practitioners should try to pick measures that can be used as inputs to design and avoid measures that are only relevant after the product or service is being produced or offered (i.e., customer satisfaction, complaints). Also, it is important to consider how the characteristic will be measured. Practitioners must avoid measurement systems that, in themselves, introduce variation into the process.

Choosing the Right Metrics:
Practitioners can save a lot of frustration by choosing the right metrics up front. This will not eliminate the need to evaluate the metrics during the design process, but it will cut down on the overall project duration. The selected metrics need to be solution independent and support the product or service as an indicator of customer needs. But keep in mind that all customers are not created equal – the project may require more than one measure per customer need. Again, also choose continuous metrics if possible.

Developing Targets and Establishing Specification Limits:
Unfortunately, there is no specific recipe for setting targets and specifications. This is a function of business know-how and technical expertise, so practitioners should use the business or subject-matter experts to assist them with brainstorming and developing these requirements. There are many variables to consider, as shown in Figure 2. (Note: Current or projected capability to achieve a performance level should not be the primary basis for establishing targets. To ensure success in the market, the customer and competitive information should be the primary drivers).

Elements of the House of Quality:
One of the most powerful tools used in defining CTQs is the Quality Function Deployment (QFD), also known as the house of quality. This is a structured methodology and mathematical tool used to identify and quantify customers’ requirements and translate them into key critical parameters. QFD helps practitioners to prioritize actions to improve their process or product to meet customers’ expectations.
As Don Clausing and John Hauser write in their article The House of Quality about QFD: “None of this is simple. An elegant idea ultimately decays into process, and processes will be confounding as long as human beings are involved. But that is no excuse to hold back. If a technique like the house of quality can help break down functional barriers and encourage teamwork, serious efforts to implement it will be many times rewarded.”
The QFD was originally developed by Yoji Akao in 1966 when he combined his work in quality assurance and quality control points with function deployment used in value engineering. Akao described QFD as a “method to transform user demands into design quality, to deploy the functions forming quality, and to deploy methods for achieving the design quality into subsystems and component parts, and ultimately to specific elements of the manufacturing process.”
Figure 3 shows the design of the house of quality.

QFD is designed to help planners focus on characteristics of a new or existing product or service from the viewpoints of market segments, company or technology-development needs. The technique yields graphs and matrices.
Basic steps in the creation of the QFD include:
  1. Identify customer needs and wants (collect VOC).
  2. Identify the engineering characteristics of products or services that meet VOC.
  3. Set development targets and test methods for the products or services.
Once again, the QFD helps transform VOC into engineering characteristics (and appropriate test methods) for a product or service, prioritizing each product or service characteristic while simultaneously setting development targets for the product or service, all of which are necessary in defining CTQs.
One of the biggest advantages of QFD is that the process requires groups of cross-functional representatives to work together to understand customer expectations in a way that focuses on customer requirements by using and strengthening functional teamwork. It provides flexible and easy-to-assimilate documentation and uses competitive positioning and marketing potential to prioritize design goals. Finally, it translates soft customer requirements into measurable goals.
Benefits experienced when using the QFD include a reduction in design, a reduction in design changes and a reduction in start-up costs.
Lessons Learned When Using a QFD:
QFD is more of an art than a science. The big benefit comes from the discussion the process generates. Practitioners might be surprised to find that even with the simplest process, a QFD requires a lot of effort. Many entries may look obvious, even after they are written down; however, if there are no “tough spots,” it probably is not being done right. Practitioners must always focus on the end customer and remember that “charts” are not the objective. Most importantly, QFD is a valuable decision support tool; it is not a decision maker.
QFD is an organizing tool – the bulk of the effort lies in gathering the inputs to the house of quality. The QFD should be performed via a cross-functional team and communicated to all involved in the design. Although QFD takes time, it will ultimately save time spent reworking “defective” designs and assist in balancing time commitment with benefits.
Mitigating Potential Impacts:
How does the inability to meet major CTQs in the design – or of not considering a CTQ – impact the customer or a company? Potential customer impacts include an increase in product or service variability, non-functional products or services, delays in delivery time and cost of the product or service, as well as a decrease in value to the customer.
Potential internal impacts include increased rework and costs, and loss of profit margin, customers (or return customers), and growth opportunities for not keeping up with its competition, which can lead to barriers to entry in other markets. Therefore, defining CTQ requirements should be at the top of a project’s priorities.
Six Sigma Project:

Defects vs Defectives: 
A product may have many defects – imperfections. But a product is not defective unless the defects prevent the product from functioning. If a product is not usable, it is considered defective.
Process View of Operations:


KANO Model Example: Airline

A popular Airline in your city is having problem meeting the sales target on their First-class flights from and to some destinations that they believe is critical to meeting their overall profit target for the year.

Several of their clients have complained about the quality of services they get on these routes on the business and first class categories of this airline and are ready to switch to another competitor that in their view offers better service quality.

Your company have just been contracted to come up with a solution on how they could improve the quality of service on this route. 

During one of your brainstorming sessions in your company, your group have come up with a list of all possible features on how to improve the quality of services on this route but doesn’t know which of these ideas to prioritize.

You’ve been assigned the task of designing a questionnaire using the Kano Model.

As a student of QPLS1x SixSigma: Define and Measure, who just took the topic “Understanding Customer Expectations - Kano Model”. You've got some feedback from the customers that have flown with the airline recently on the First-class category.

1) First, consider the functional and dysfunctional responses for each feature below. (You can add more and discuss this in the discussion forum):
For example: On-board shower
F: Functional question: If the suites have an On-board shower spa, how do you feel?
D: Dysfunctional question: If the suites does not have an On-board shower spa, how do you feel?
The feedback are as shown below for 11 characteristics. 
·      Complimentary Chauffeur-drive service. Response:  F: Like, D: Dislike
·      Additional legroom. Response:  F: Expect/Must, D: Do not care
·      Expedited baggage service. Response:  F: Like, D: Can live with it
·      LED mood lighting. Response:  F: Do not care, D: Dislike
·      Social seating arrangement. Response:  F: Can live with it, D: Dislike
·      New selection of wine, beer, spirit and cocktail. Response:  F: Like, D: Do not care
·      Live events with fellow travellers on the 55-inch LCD TV screen. Response:  F: Do not care, D: Can live with it
·      Complimentary gift package on board. Response:  F: Like, D: Do not care
·      Privacy suite doors. Response:  F: Can live with it, D: Dislike
·      Private cinema. Response:  F: Like, D: Can live with it
·      Personal dining service. Response:  F: Like, D: Dislike
2.)  Second, Categorize them by using the Kano Evaluation Table.
Example: On-board Shower Spa : in the table we see that this is given the following ratings
Functional Question (F) = Expect/Must
Dysfunctional Question (D) = Dislike
You can see from the table that this translates to E = Expected/Basic Quality(E), as shown on the table.
Complete this exercise for the following 11 characteristics.
3) Finally, based on your results, make a recommendation to the airline company! (Assume that most of the feedback follow the same pattern as the one above).
Candle case study:


During the Define phase the (target) customer’s expectations must be determined and the outputs which are Critical to Quality must be identified. Additionally, it is necessary to gain a basic understanding of the Process Flow. In a full process improvement project, the project team would be established to draw up a formal project charter, including setting up a project timeline and cost estimation.
In this first part of our Candle case study, we will examine the Define phase in order to:
    • Specify the customer expectations for quality parameters for candles.
    • Identify the process output which is critical to quality for candles according to the European Candle Association.
    • Gain a basic understanding of the production flow (video KOPSCHITZ KERZEN).

In this section we will begin our case study of a candle-making process. Here we will review the first step of the DMAIC cycle, Define, and will look at a few examples of defining customer expectations for candles.

Candles have been providing light for mankind for thousands of years and have always been central to religious festivals and ceremonies the world over. Their origins have been traced back to the ancient Romans, who developed wicked candles using papyrus dipped in animal fat or beeswax. The Egyptians used forms of wicked candles over 3,000 years ago and early civilisations in China, Japan and India used various forms of paper and wax to form their candles.
Rendered animal fat or tallow candles were used commonly in Europe in the middle ages, with the clean-burning, sweet-smelling beeswax candles reserved for only those rich families who could afford them. Candlemakers – or chandlers – made and sold candles. More significant developments came only in the 19th century when the extraction of stearic acid from animal fatty acids was discovered in France, leading to the development of stearin wax to provide hard, durable and clean-burning candles. Candle production became mechanized, bringing down the costs. Paraffin wax became available in the mid 19th century, providing a high-quality wax that was cheap to produce. Combining the paraffin wax and stearin provided economical, durable and clean-burning candles.
Interestingly, although the light bulb quickly replaced the humble candle, new demand surged fin the 20th century for candles as decorative pieces and to provide atmosphere. Candles produced in many different shapes and colors and new types of wax were developed. Today we enjoy candles for many reasons.


During the Measure phase, you will measure and evaluate the Critical to Quality Outputs identified in the Define phase (in this case study, we will be following the quality standards of the European Candle Association). This will establish the baseline process performance with respect to the goal of the process improvement project. You and your project team then need to understand the production process in as much detail as possible by creating a detailed process map. The process map is carefully reviewed in order to identify the parameters which are likely to influence the critical to quality output. In an actual project, a data collection plan would be developed and implemented for these parameters.
In this Candle case study, the learning objectives for Measure are:
    • use the given process flow to map the process for candle production.
    • establish the baseline process performance with respect to the customer expectations.


During this step of our project, the project team will analyze the data for the key process parameters identified in Measure which are likely to influence the critical to quality output. In Analyze, the objective is to establish the root cause of the process deviation or process problem by confirming (or not) the effect of these parameters on the output. In an actual project, this phase could include additional work to identify potential causes and to develop Cause and Effect diagrams to help prioritize the work.
The learning objective for Analyze for the Candle case study is:
    • Using the key parameters which are likely to influence the critical to quality output, perform correlation analysis to determine root cause.


During this step of the project, your project team will propose a hypothesis for a process change in order to effect the process improvement. You will identify a solution and demonstrate that the solution does effect the desired process improvement.
For the Candle production, the learning objectives for Improve are:
    • Establish a hypothesis to be tested.
    • Analyze the data for the new process to show that the implemented changes result in an improved process and quantify this improvement.


In the final step of the project of the DMAIC process improvement cycle, the goal is to ensure that the improvement gains are maintained over the long term. Here it is important to implement standardized documentation, employee training and on-going process monitoring.
In this Candle case study, the learning objective for Control is:
    • Establish the appropriate X-bar and Range control charts for the production process to prevent the problem from occurring again and to monitor the production performance.
In each section of the case study, we will go through each phase of the DMAIC project which will help us accomplish the corresponding learning objectives given above for each DMAIC phase.

Kopschitz Kerzen Intro and Process Flow:

European Candle Association & RAL Quality Mark:
The European Candle Association (ECA) is an alliance of the biggest and most reputable candle manufacturing companies in Europe. Together, its members represent more than two thirds of the European candle production and they produce high-quality candles in 11 European countries.

(Source:, accessed on 14.02.2016)
The European Candles Association pursues in particular the following main goals:
    • Maintenance and further development of safety, health and environmental protection.
    • Maintaining and increasing quality of the candles.
    • Information about and participation in legal developments and proposals.
    • Identification of new raw materials and development of new technologies.
As the collective voice of the European candle industry it is the association's explicit intention to actively seek and further improve the communication with all market participants on all aspects of high-quality candles as a product. The ECA would like to include all interested national and international parties, from raw material suppliers to retailers, from other associations to official bodies, from end consumers to consumer organizations.
This way, manufacturers, retailers and consumers shall continue to gain pleasure from one of the oldest and most atmospheric consumer goods in the world – the candle.

The „Quality Mark for Candles” is the only independent quality standard for candles at the moment. It requires both, high quality raw materials as well as impeccable burning characteristics.
You can be absolutely sure to buy a quality candle when it shows this quality mark.
The “Quality Mark for Candles” is awarded by the European Quality Association for Candles to its members if their candles fulfill all requirements regarding raw materials, dimensions, weight, color and design and if they burn perfectly:
    • bright, calm flame
    • ideal wick curvature
    • no sooting
    • no dripping
    • adherence to the burning time
    • minimum wax remainder
To be able to guarantee the high quality standard, the manufacturers pledge them-selves to continuously monitor their products. The following items are part of the testing and monitoring program:

Initial inspection:

The compliance with all requirements regarding raw materials and candles are to be checked by a state-approved testing organization or an independent expert.


A continuous and verifiable self-monitoring of all graded products is to be performed by the manufacturer.

External monitoring:

The compliance with all requirements regarding raw materials and candles as well as the correct performance of the self-monitoring is to be checked by an independent expert without prior notice.

External candle tests:

Appearance, dimensions and burning behavior are to be tested by an independent and accredited laboratory at least once year.
As we discussed in the lecture:
• A random variable is a variable, which assumes for its values the outcomes of random experiment.

• A random variable assumes only real numbers.

• The set of all possible values of a random variable X is called its range space.

• Events can be defined in the range space and their probabilities can be found. 
There are two kinds of random variables:
• A discrete random variable takes a finite (or countably infinite) number of possible values.  

• A continuous random variable takes an infinite number of possible values; it takes values in an interval.

How do we use the concept of random variables? Random variables are used to represent populations with variability in them. When we are trying to make a decision which concerns an element in a population (for example, we are a purchaser for a retailer and need to order in advance the ski jackets for this coming winter season), but we do not know the outcome of the event (we don’t know what the demand for each style and size jacket will be during the season), then we would like to understand the behavior of the possible outcomes of the event (what is the most likely season demand for each size and style combination).
Probability distributions are models used to describe the behavior of random variables and if we know the shape of the probability distribution, this can help us to make decisions about the possible outcomes for our random variable.

We learned that there are two types of probability distributions:
    • Probability Mass Function used to represent discrete random variables

    • Probability Density Function used to represent continuous random variables

These functions are used to calculate the probability that the random variable will take on a certain value or a certain range of values. 
We often find that continuous random variables are normally distributed, which means that their probability density function, pdf, follows a bell shape centered around a value μ, the mean and with a variation given by the standard deviation, σ.
As for all probability density functions, denoted by f(x), the following applies:
 1) f(x) ≥ 0, for all x
The pdf is positive for all possible values of the random variable (no “negative” probabilities)
 2) ∫ f(x)dx = 1
The function integrates to 1.0 over all possible values of the random variable (the random variable must take a value given by the function)
3) P(a ≤ X ≤ b) = ∫ f(x)dx, integrated over the interval from a to b 
The integral of the function value over any interval gives the probability that the random variable lies in that interval.
The standard deviation is the distance from the mean to the point of inflection of the normal curve, as shown below.
Take a look at the distribution of the students' heights and answer the following multiple choice questions. 

The standard deviation is the distance from the mean to the point of inflection (a change of curvature from convex to concave at a particular point on a curve.) of the normal curve.
Exercise - Central Limit Theorem: Tossing a die (1 toss)

We will throw a die (“die” is the singular of “dice”) one time and look at the number that appears. Our random variable, X, is the number that appears: a discrete random variable with 6 possible results: 1, 2, 3, 4, 5, 6. The probability that we get a 1 is 1/6. The probability that we get a 2 is also 1/6, etc. The Probability Mass Function (pmf) looks like this:
Each possible value of X has a probability of 1/6 (=0.16667) of occurring. You can use your equation for the Mean (μ) for a discrete random variable (also called “Expected Value”): 
μx = Σx x . p(x)
and calculate the mean of this distribution:
μx = 1.(1/6) + 2.(1/6) + 3.(1/6) + 4.(1/6) + 5.(1/6) + 6.(1/6) = 3.5
Likewise you can calculate the standard deviation:
σx = Σx (x - μx)2 . p(x)

= (1 – 3.5)2 . (1/6) + (2 – 3.5)2 . (1/6) + (3 – 3.5)2 . (1/6) + (4 – 3.5)2 . (1/6) + (5 – 3.5)2 . (1/6) +(6 – 3.5)2 . (1/6) 

= (-2.5)2 . (1/6) + (-1.5)2 . (1/6) + (0.5)2 . (1/6) + (0.5)2 . (1/6) + (-1.5)2 . (1/6) + (-2.5)2 . (1/6) + (-3.5)2 . (1/6) 
= 2.92
Exercise - Central Limit Theorem: Tossing a die (2 tosses)
Now we will toss the die 2 times.

Our random variable X is still the value we see on the die, but we will now calculate the average, X́ = (X_first toss + X_second toss)/2
For example, if we roll first a 4 and then a 2, our average value is (4+2)/2 = 3.
Here we have a “sample” size of n = 2.
Below you can see the Probability Mass Function (pmf):

Each possible value of X́ is given, and we see that the pmf looks more triangular. Not every value is equally probable. For example, to get a value of X́=2 you need to roll a 1 twice. This is a probability of (1/6)(1/6) = 1/36. But to get a value of X́=1.5 you could roll first a 1 and then a 2 or first a 2 and then a 1. This is a probability of (1/6)(1/6) + (1/6)(1/6) = 2/36. And so forth.
HANWAG Hiking Shoes (Baseline Process Performance):