Total Quality Management

Total quality management revolves around three main themes: the customer, the process, and the people. Figure 3-1 shows some basic features of a TQM model. At its core are the company vision and mission and management commitment. They bind the customer, the process, and the people into an integrated whole. A company's vision is quite simply what the company wants to be. The mission lays out the company's strategic focus. Every employee should understand the company's vision and mission so that individual efforts will contribute to the organizational mission. When employees do not understand the strategic focus, individuals and even departments pursue their own goals rather than those of the company, and the company's goals are inadvertently sabotaged. The classic example is maximizing production with no regard to quality or cost.

Management commitment is another core value in the TQM model. It must exist at all levels for the company to succeed in implementing TQM. Top management envisions the strategy and creates policy. Middle management works on implementation. At the operational level, appropriate quality management tools and techniques are used.

Satisfying customer needs and expectations is a major theme in TQM—in fact, it is the driving force. Without satisfied customers, market share will not grow and revenue will not increase. Management should not second-guess the customer. For example, commercial builders should construct general merchandise stores only after they have determined that there is enough customer interest to support them. If consumers prefer specialty stores, specialty stores should be constructed. Direct feedback using a data-driven approach is the best way to identify customer expectations and needs. A company's strategic plan must conform to these needs.

A key principle in quality programs is that customers are both internal and external. The receiving department of a processed component is a customer of that processing unit. Feedback from such internal customers identifies problem areas before the product reaches its finished stage, thus reducing the cost of scrap and rework.

Customer expectations can, to some extent, be managed by the organization. Factors such as the quality of products and services and warranty policies offered by the competitor influence customer expectations directly. The company can, through truthful advertising, shape the public's expectations. For example, if the average life of a lawn mower under specified operating conditions is 15 years, there is no reason to exaggerate it. In service operations, customers know which companies are responsive and friendly. This doesn't need advertising. Customer surveys can help management determine discrepancies between expectations and satisfaction. Taking measures to eliminate discrepancies is known as gap analysis.

The second theme in TQM is the process. Management is responsible for analyzing the process to improve it continuously. In this framework, vendors are part of the extended process, as advocated by Deming. As discussed earlier, integrating vendors into the process improves the vendors' products, which leads to better final products. Because problems can and do span functional areas, self-directed cross-functional teams are important for generating alternative feasible solutions—the process improves again. Technical tools and techniques along with management tools come in handy in the quest for quality improvement. Self-directed teams are given the authority to make decisions and to make appropriate changes in the process.

The third theme deals with people. Human “capital” is an organization's most important asset. Empowerment—involving employees in the decision-making process so that they take ownership of their work and the process—is a key factor in TQM. It is people who find better ways to do a job, and this is no small source of pride. With pride comes motivation. There is a sense of pride in making things better through the elimination of redundant or non-value-added tasks or combining operations. In TQM, managing is empowering.

Barriers restrict the flow of information. Thus, open channels of communication are imperative, and management had better maintain these. For example, if marketing fails to talk to product design, a key input on customer needs will not be incorporated into the product. Management must work with its human resources staff to empower people to break down interdepartmental barriers. From the traditional role of management of coordinating and controlling has developed the paradigm of coaching and caring. Once people understand that they, and only they, can improve the state of affairs, and once they are given the authority to make appropriate changes, they will do the job that needs to be done. There originates an intrinsic urge from within to do things better. Such an urge has a force that supersedes external forms of motivation.

Linking the human element and the company's vision is the fabric we call organizational culture. Culture comprises the beliefs, values, norms, and rules that prevail within an organization. How is business conducted? How does management behave? How are employees treated? What gets rewarded? How does the reward system work? How is input sought? How important are ethics? What is the social responsibility of the company? The answers to these and many other questions define an organization's culture. One culture may embrace a participative style of management that empowers its employees and delights its customers with innovative and timely products. Another culture may choose short-term profit over responsibility to the community at large. Consider, for example, the social responsibility adopted by the General Electric Company. The company and its employees made enormous contributions to support education, the arts, the environment, and human services organizations worldwide.

Vision and Quality Policy

A company's vision comprises its values and beliefs. Their vision is what they want it to be, and it is a message that every employee should not only hear but also believe in. Visions, carefully articulated, give a coherent sense of purpose. Visions are about the future, and effective visions are simple and inspirational. Finally, it must be motivational so as to evoke a bond that creates unison in the efforts of persons working toward a common organizational goal. From the vision emanates a mission statement for the organization that is more specific and goal oriented.

A service organization, IBM Direct, is dedicated to serving U.S. customers who order such IBM products as ES/9000 mainframes, RS/6000 and AS/400 systems, connectivity networks, and desktop software. Its vision for customer service is “to create an environment for customers where conducting business with IBM Direct is considered an enjoyable, pleasurable and satisfying experience.” This is what IBM Direct wants to be. Its mission is “to act as the focal point for post-sale customers issues for IBM Direct customers. We must address customer complaints to obtain timely and complete resolutions. And, through root cause analysis, we must ensure that our processes are optimized to improve our customer satisfaction.” Here, again, the mission statement gets specific. This is how IBM Direct will get to its vision. Note that no mention is made of a time frame. This issue is usually dealt with in goals and objectives.

Framed by senior management, a quality policy is the company's road map. It indicates what is to be done, and it differs from procedures and instructions, which address how it is to be done, where and when it is to be done, and who is to do it. A beacon in TQM leadership, Xerox Corporation is the first major U.S. corporation to regain market share after losing it to Japanese competitors. Xerox attributes its remarkable turnaround to its conversion to TQM philosophy. The company's decision to rededicate itself to quality through a strategy called Leadership Through Quality has paid off. Through this process, Xerox created a participatory style of management that focuses on quality improvement while reducing costs. It encouraged teamwork, sought more customer feedback, focused on product development to target key markets, encouraged greater employee involvement, and began competitive benchmarking. Greater customer satisfaction and enhanced business performance are the driving forces in its quality program, the commitment to which is set out in the Xerox quality policy: “Quality is the basic business principle at Xerox.”

Another practitioner of TQM, the Eastman Chemical Company, manufactures and markets over 400 chemicals, fibers, and plastics for over 7000 customers around the world. A strong focus on customers is reflected in its vision: “to be the world's preferred chemical company.” A similar message is conveyed in its quality goal: “to be the leader in quality and value of products and services.” Its vision, values, and goals define Eastman's quality culture. The company's quality management process is set out in four directives: “focus on customers; establish vision, mission, and indicators of performance; understand, stabilize, and maintain processes; and plan, do, check, act for continual improvement and innovation.”

Eastman Chemical encourages innovation and provides a structured approach to generating new ideas for products. Cross-functional teams help the company understand the needs of both its internal and external customers. The teams define and improve processes, and they help build long-term relationships with vendors and customers. Through the Eastman Innovative Process, a team of employees from various areas—design, sales, research, engineering, and manufacturing—guides an idea from inception to market. People have ownership of the product and of the process. Customer needs and expectations are addressed through the process and are carefully validated. One outcome of the TQM program has been the drastic reduction (almost 50%) of the time required to launch a new product. Through a program called Quality First, employees team with key vendors to improve the quality and value of purchased materials, equipment, and services. Over 70% of Eastman's worldwide customers have ranked the company as their best supplier. Additionally, Eastman has received an outstanding rating on five factors that customers view as most important: product quality, product uniformity, supplier integrity, correct delivery, and reliability. Extensive customer surveys led the company to institute a no-fault return policy on its plastic products. This policy, believed to be the only one of its kind in the chemical industry, allows customers to return any product for any reason for a full refund.

Balanced Scorecard

The balanced scorecard (BSC) is a management system that integrates measures derived from the organization's strategy. It integrates measures related to tangible as well as intangible assets. The focus of BSC is on accomplishing the company's mission through the development of a communication and learning system. It translates the mission and strategy to objectives and measures that span four dimensions: learning and growth, internal processes, customers, and financial (Kaplan and Norton 1996). Whereas traditional systems have focused only on financial measures (such as return on investment), which is a short-term measure, BSC considers all four perspectives from a long-term point of view. So, for example, even for the financial perspective, it considers measures derived from the business strategy, such as sales growth rate or market share in targeted regions or customers. Figure 3-2 shows the concept behind the development of a balanced scorecard.

Measures in the learning and growth perspective that serve as drivers for the other three perspectives are based on three themes. First, employee capabilities, which include employee satisfaction, retention, and productivity, are developed. Improving satisfaction typically improves retention and productivity. Second, development of information systems capabilities is as important as the system for procuring raw material, parts, or components. Third, creation of a climate for growth through motivation and empowerment is an intangible asset that merits consideration.

For each of the four perspectives, diagnostic and strategic measures could be identified. Diagnostic measures relate to keeping a business in control or in operation (similar to the concept of quality control). On the contrary, strategic measures are based on achieving competitive excellence based on the business strategy. They relate to the position of the company relative to its competitors and information on its customers, markets, and suppliers. Strategic measures could be of two types: outcome measures and performance measures. Outcome measures are based on results from past efforts and are lagging indicators. Examples are return on equity or employee productivity. Performance measures reflect the uniqueness of the business strategy and are lead indicators, examples of which are sales growth rate by segment or percentage revenue from new products. Each performance measure has to be related to an outcome measure through a cause-and-effect type of analysis, which will therefore reflect the financial drivers of profitability. It could also identify the specific internal processes that will deliver value to targeted customers if the company strategy is to expand its market share for a particular category of customers.

In the learning and growth perspective, employee satisfaction, a strategic lag indicator, could be measured on an ordinal scale of 1 to 5. Another lag indicator could be the revenue per employee, a measure of employee productivity. A performance measure, a lead indicator, could be the strategic job coverage ratio, which is the ratio of the number of employees qualified for strategic jobs to the organizational needs that are anticipated. This is a measure of the degree to which the company has reskilled its employees. Under motivation, an outcome measure could be the number of suggestions per employee or the number of suggestions implemented.

When considering the internal processes perspective, one is required to identify the critical processes that will enable the meeting of customer or shareholder objectives. Based on the expectations of specific external constituencies, they may impose demands on internal processes. Cycle time, throughput, and costs associated with existing processes are examples of diagnostic measures. In the strategic context, a business process for creating value could include innovation, operations, and post-sale service. Innovation may include basic research to develop new products and services or applied research to exploit existing technology. Time to develop new products is an example of a strategic outcome measure. Under post-sale service, measures such as responsiveness (measured by time to respond), friendliness, and reliability are applicable.

The strategic aspect of the customer perspective deals with identifying customers to target and the corresponding market segments. For most businesses, core outcome measures are market share, degree of customer retention, customer acquisition, customer satisfaction, and customer profitability from targeted segments. All of these are lagging measures and do not indicate what employees should be doing to achieve desired outcomes. Thus, under performance drivers (leading indicators), measures that relate to creating value for the customer are identified. These may fall in three broad areas: product/service attributes, customer relationship, and image and reputation. In association with product/service attributes, whereas lead time for existing products may be a diagnostic measure, time to serve targeted customers (e.g., quick check-in for business travelers in a hotel) is a strategic performance measure. Similarly, while quality of product/services (as measured by, say, defect rates) is considered as a “must,” some unique measures such as service guarantees (and the cost of such), which offer not only a full refund but also a premium above the purchase price, could be a performance measure.

Under the financial perspective, the strategic focus is dependent on the stage in which the organization currently resides (i.e., infancy, dominancy, or maturity). In the infancy stage, companies capitalize on significant potential for growth. Thus, large investments are made in equipment and infrastructure which may result in negative cash flow. Sales growth rate by product or market segment could be a strategic outcome measure. In the dominancy phase, where the business dwells mainly on the existing market, traditional measures such as gross margin or operating income are valid. Finally, for those in the mature stage, a company may not invest in new capabilities with a goal of maximizing cash flow. Unit costs could be a measure. For all three phases, some common themes are revenue growth, cost reduction, and asset utilization (Kaplan and Norton 1996).

Several features are to be noted about the balanced scorecard. First, under strategic measures, the link between performance measures and outcome measures represent a cause-and-effect relationship. However, based on the outcome measures observed and a comparison with the strategic performance expected, this feedback may indicate a choice of different performance measures. Second, all of the measures in the entire balanced scorecard represent a reflection of a business's performance. If such a performance does not match the performance expected based on the company strategy, a feedback loop exists to modify the strategy. Thus, the balanced scorecard serves an important purpose in linking the selection and implementation of an organization's strategy.

Performance Standards

One intended outcome of a quality policy is a desirable level of performance: that is, a defect-free product that meets or exceeds customer needs. Even though current performance may satisfy customers, organizations cannot afford to be complacent. Continuous improvement is the only way to stay abreast of the changing needs of the customer. The tourism industry, for instance, has seen dramatic changes in recent years; options have increased and customer expectations have risen. Top-notch facilities and a room filled with amenities are now the norm and don't necessarily impress the customer. Meeting and exceeding consumer expectations are no small challenges. Hyatt Hotels Corporation has met this challenge head-on. Its “In Touch 100” quality assurance initiative provides a framework for its quality philosophy and culture. Quality at Hyatt means consistently delivering products and services 100% of the time. The In Touch 100 program sets high standards—standards derived from guest and employee feedback—and specifies the pace that will achieve these standards every day. The core components of the quality assurance initiative are standards, technology, training, measurements, recognition, communication, and continuous improvement (Buzanis 1993).

Six Sigma Quality

Although a company may be striving toward an ultimate goal of zero defects, numerical standards for performance measurement should be avoided. Setting numerical values that may or may not be achievable can have an unintended negative emotional impact. Not meeting the standard, even though the company is making significant progress, can be demoralizing for everyone. Numerical goals also shift the emphasis to the short term, as long-term benefits are sacrificed for short-term gains.

So, the question is: How do we measure performance? The answer is: by making continuous improvement the goal and then measuring the trend (not the numbers) in improvement. This is also motivational. Another effective method is benchmarking; this involves identifying high-performance companies or intra-company departments and using their performance as the improvement goal. The idea is that although the goals may be difficult to achieve, others have shown that it can be done.

Quantitative goals do have their place, however, as Motorola, Inc. has shown with its concept of six sigma quality. Sigma (σ) stands for the standard deviation, which is a measure of variation in the process. Assuming that the process output is represented by a normal distribution, about 99.73% of the output is contained within bounds that are three standard deviations (3σ) from the mean. As shown in Figure 3-3, these are represented as the lower and upper tolerance limits (LTL and UTL). The normal distribution is characterized by two parameters: the mean and the standard deviation. The mean is a measure of the location of the process. Now, if the product specification limits are three standard deviations from the mean, the proportion of nonconforming product is about 0.27%, which is approximately 2700 parts per million (ppm); that is, the two tails, each 1350 ppm, add to 2700 ppm. On the surface, this appears to be a good process, but appearances can be deceiving. When we realize that most products and services consist of numerous processes or operations, reality begins to dawn. Even though a single operation may yield 97.73% good parts, the compounding effect of out-of-tolerance parts will have a marked influence on the quality level of the finished product. For instance, for a product that contains 1000 parts or has 1000 operations, an average of 2.7 defects per product unit is expected. The probability that a product contains no defective parts is only 6.72% (e^−2.7, using the Poisson distribution discussed in a later chapter)! This means that only about 7 units in 100 will go through the entire manufacturing process without a defect (rolled throughput yield)—not a desirable situation.

For a product to be built virtually defect free, it must be designed to specification limits that are significantly more than ±3σ from the mean. In other words, the process spread as measured by ±3σ has to be significantly less than the spread between the upper and lower specification limits (USL and LSL). Motorola's answer to this problem is six sigma quality; that is, process variability must be so small that the specification limits are six standard deviations from the mean. Figure 3-4 demonstrates this concept. If the process distribution is stable (i.e., it remains centered between the specification limits), the proportion of nonconforming product should be only about 0.001 ppm on each tail.

In real-world situations, the process distribution will not always be centered between the specification limits; process shifts to the right or left are not uncommon. It can be shown that even if the process mean shifts by as much as 1.5 standard deviations from the center, the proportion nonconforming will be about 3.4 ppm. Comparing this to a 3σ capability of 2700 ppm demonstrates the improvement in the expected level of quality from the process. If we consider the previous example for a product containing 1000 parts and we design it for 6σ capability, an average of 0.0034 defect per product unit (3.4 ppm) is expected instead of the 2.7 defects expected with 3σ capability. The cumulative yield (rolled throughput yield) from the process will thus be about 99.66%, a vast improvement over the 6.72% yield in the 3σ case.

Establishing a goal of 3σ capability is acceptable as a starting point, however, because it allows an organization to set a baseline for improvement. As management becomes more process oriented, higher goals such as 6σ capability become possible. Such goals may require fundamental changes in management philosophy and the organizational culture.

Although the previous description of six sigma has defined it as a metric, on a broader perspective six sigma may also be viewed as a philosophy or a methodology for continuous improvement. When six sigma is considered as a philosophy, it is considered as a strategic business initiative. In this context, the theme of identification of customer needs and ways to satisfy the customer is central. Along the lines of other philosophies, continuous improvement is integral to six sigma as well.

As a methodology, six sigma may be viewed as the collection of the following steps or phases: define, measure, analyze, improve, and control. Within each phase there are certain tools that could be utilized. Some of the tools are discussed in this chapter. In the define phase customer needs are translated to specific attributes that are critical to meeting such needs. Typically, these are categorized in terms of critical to quality, delivery, or cost. Identification of these attributes will create a framework for study. For example, suppose that the waiting time of customers in a bank is the problem to be tackled. The number of tellers on duty during specific periods in the day is an attribute that is critical to reducing the waiting time, which might be a factor of investigation.

The measure phase consists of identifying metrics for process performance. This includes the establishment of baseline levels as well. In our example, a chosen metric could be the average waiting time prior to service, in minutes, while the baseline level could be the current value of this metric, say 5 minutes. Key process input and output variables are also identified. In the define phase, some tools could be the process flowcharts or its detailed version, the process map. To identify the vital few from the trivial many, a Pareto chart could be appropriate. Further, to study the various factors that may affect an outcome, a cause-and-effect diagram may be used. In the measure phase, one first has to ensure that the measurement system itself is stable. The technical name for such a study is gage repeatability and reproducibility. Thereafter, benchmark measures of process capability could be utilized, some of which are defects per unit of product, parts per million nonconforming, and rolled throughout yield, representing the proportion of the final product that has no defects. Other measures of process capability are discussed later in a separate chapter.

In the analyze phase, the objective is to determine which of a multitude of factors affects the output variable (s) significantly, through analysis of collected data. Tools may be simple graphical tools such as scatterplots and multivari charts. Alternatively, analytical models may be built linking the output or response variable to one of more independent variables through regression analysis. Hypothesis testing on selected parameters (i.e., average waiting time before and after process improvement) could be pursued. Analysis-of-variance techniques may be used to investigate the statistical significance of one or more factors on the response variable.

The improve phase consists of identifying the factor levels of significant factors to optimize the performance measure chosen, which could be to minimize, maximize, or achieve a goal value. In our example, the goal could be to minimize the average waiting time of customers in the bank subject to certain resource or other process constraints. Here, concepts in design of experiments are handy tools.

Finally, the control phase deals with methods to sustain the gains identified in the preceding phase. Methods of statistical process control using control charts, discussed extensively in later chapters, are common tools. Process capability measures are also meaningful in this phase. They may provide a relative index of the degree to which the improved product, process, or service meets established norms based on customer requirements.

QFD Process

Figure 3-5 shows a QFD matrix, also referred to as the house of quality. The objective statement delineates the scope of the QFD project, thereby focusing the team effort. For a space shuttle project, for example, the objective could be to identify critical safety features. Only one task is specified in the objective. Multiple objectives are split into separate QFDs in order to keep a well-defined focus.

The next step is to determine customer needs and wants. These are listed as the “whats” and represent the individual characteristics of the product or service. For example, in credit-card services, the “whats” could be attributes such as a low interest rate, error-free transactions, no annual fee, extended warranty at no additional cost, customer service 24 hours a day, and a customers' advocate in billing disputes. The list of “whats” is kept manageable by grouping similar items. On determination of the “whats” list, a customer importance rating that prioritizes the “whats” is assigned to each item. Typically, a scale of 1–5 is used, with 1 being the least important. Multiple passes through the list may be necessary to arrive at ratings that are acceptable to the team. The ratings serve as weighting factors and are used as multipliers for determining the technical assessment of the “hows.” The focus is on attributes with high ratings because they maximize customer satisfaction. Let's suppose that we have rated attributes for credit-card services as shown in Table 3-1. Our ratings thus imply that our customers consider error-free transactions to be the most important attribute and the least important to be charging no annual fee.

The customer plays an important role in determining the relative position of an organization with respect to that of its competitors for each requirement or “what.” Such a comparison is entered in the section on “customer assessment of competitors.” Thus, customer perception of the product or service is verified, which will help identify strengths and weaknesses of the company. Different focus groups or surveys should be used to attain statistical objectivity. One outcome of the analysis might be new customer requirements, which would then be added to the list of “whats,” or the importance ratings might change. Results from this analysis will indicate what dimensions of the product or service the company should focus on. The same rating scale that is used to denote the importance ratings of the customer requirements is used in this analysis.

Consider, for example, the customer assessment of competitors shown in Table 3-2, where A represents our organization. The ratings are average scores obtained from various samples of consumers. The three competitors (companies B, C, and D) are our company's competition, so the maximum rating scores in each “what” will serve as benchmarks and thus the acceptable standard toward which we will strive. For instance, company C has a rating of 4 in the category “customer service 24 hours a day” compared to our 2 rating; we are not doing as well in this “what.” We have identified a gap in a customer requirement that we consider important. To close this gap we could study company C's practices and determine whether we can adopt some of them. We conduct similar analyses with the other “whats,” gradually implementing improved services. Our goal is to meet or beat the circled values in Table 3-2, which represent best performances in each customer requirement. That is, our goal is to become the benchmark.

Customer Requirements (“Whats”)	Competitive Assessment of Companies
	A	B	C	D
Low interest rate	3	2		2
Error-free transactions	4		3	3
No annual fee			2	3
Extended warranty at no additional cost	2	2	1
Customer service 24 hours a day	2	2		3
Customers' advocate in billing disputes		2	3	3

Coming up with a list of technical descriptors—the “hows”—that will enable our company to accomplish the customer requirements is the next step in the QFD process. Multidisciplinary teams whose members originate in various departments will brainstorm to arrive at this list. Departments such as product design and development, marketing, sales, accounting, finance, process design, manufacturing, purchasing, and customer service are likely to be represented on the team. The key is to have a breadth of disciplines in order to “capture” all feasible “hows.” To improve our company's ratings in the credit-card services example, the team might come up with these “hows”: software to detect errors in billing, employee training on data input and customer services, negotiations and agreements with major manufacturers and merchandise retailers to provide extended warranty, expanded scheduling (including flextime) of employee operational hours, effective recruiting, training in legal matters to assist customers in billing disputes, and obtaining financial management services.

Target goals are next set for selected technical descriptors or “hows.” Three symbols are used to indicate target goals: ↑ (maximize or increase the attained value), ↓ (minimize or decrease the attained value), and (achieve a desired target value). Table 3-3 shows how our team might define target goals for the credit-card services example. Seven “hows” are listed along with their target goals. As an example, for how 1, creating a software to detect billing errors, the desired target value is zero: that is, no billing errors. For how 2, it is desirable to maximize or increase the effect of employee training to reduce input errors and interact effectively with customers. Also, for how 4, the target value is to achieve customer service 24 hours a day. If measurable goals cannot be established for a technical descriptor, it should be eliminated from the list and the inclusion of other “hows” considered.

“Hows”	1	2	3	4	5	6	7
Target goals		↑	↑		↑	↑	↑
Legend
Number	Technical descriptors or “hows”
1	Software to detect billing errors
2	Employee training on data input and customer services
3	Negotiations with manufacturers and retailers (vendors)
4	Expanded scheduling (including flextime) of employees
5	Effective recruiting
6	Legal training
7	Financial management services
Symbol	Target goal
↑	Maximize or increase attained value
↓	Minimize or decrease attained value
	Achieve a target value

The correlation matrix of the relationship between the technical descriptors is the “roof” of the house of quality. In the correlation matrix shown in Figure 3-6, four levels of relationship are depicted: strong positive, positive, negative, and strong negative. These indicate the degree to which the “hows” support or complement each other or are in conflict. Negative relationships may require a trade-off in the objective values of the “hows” when a technical competitive assessment is conducted. In Figure 3-6, which correlates the “hows” for our credit-card services example, how 1, creating a software to detect billing errors, has a strong positive relationship (++) with how 2, employee training on data input and customer services. The user friendliness of the software will have an impact on the type and amount of training needed. A strong positive relationship indicates the possibility of synergistic effects. Note that how 2 also has a strong positive relationship with how 5; this indicates that a good recruiting program in which desirable skills are incorporated into the selection procedure will form the backbone of a successful and effective training program.

Following this, a technical competitive assessment of the “hows” is conducted along the same lines as the customer assessment of competitors we discussed previously. The difference is that instead of using customers to obtain data on the relative position of the company's “whats” with respect to those of the competitors, the technical staff of the company provides the input on the “hows.” A rating scale of 1–5, as used in Table 3-2, may be used. Table 3-4 shows how our company's technical staff has assessed technical competitiveness for the “hows” in the credit-card services example. Our three competitors, companies B, C, and D, are reconsidered. For how 1 (creating a software to detect billing errors), our company is doing relatively well, with a rating of 4, but company B, with its rating of 5, is doing better; company B is therefore the benchmark against which we will measure our performance. Similarly, company C is the benchmark for how 2; we will look to improve the quality and effectiveness of our training program. The other assessments reveal that we have room to improve in hows 3, 4, and 5, but in hows 6 and 7 we are the benchmarks. The circled values in Table 3-4 represent the benchmarks for each “how.”

Company	Technical Descriptors (“Hows”)
	1	2	3	4	5	6	7
A	4	3	2	3	4
B		3	1		1	2	3
C	3		2	2		3	2
D	2	2		1	3	3	4

The analysis shown in Table 3-4 can also assist in setting objective values, denoted by the “how muches,” for the seven technical descriptors. The achievements of the highest-scoring companies are set as the “how muches,” which represent the minimum acceptable achievement level for each “how.” For example, for how 4, since company B has the highest rating, its achievement level will be the level that our company (company A) will strive to match or exceed. Thus, if company B provides customer service 16 hours a day, this becomes our objective value. If we cannot achieve these levels of “how muches,” we should not consider entering this market because our product or service will not be as good as the competition's.

In conducting the technical competitive assessment of the “hows,” the probability of achieving the objective value (the “how muches”) is incorporated in the analysis. Using a rating scale of 1–5, 5 representing a high probability of success, the absolute scores are multiplied by the probability scores to obtain weighted scores. These weighted scores now represent the relative position within the industry and the company's chances of becoming the leader in that category.

The final step of the QFD process involves the relationship matrix located in the center of the house of quality (see Figure 3-5). It provides a mechanism for analyzing how each technical descriptor will help in achieving each “what.” The relationship between a “how” and a “what” is represented by the following scale: 0 ≡ no relationship; 1 ≡ low relationship; 3 ≡ medium relationship; 5 ≡ high relationship. Table 3-5 shows the relationship matrix for the credit-card services example. Consider, for instance, how 2 (employee training on data input and customer services). Our technical staff believes that this “how” is related strongly to providing error-free transactions, so a score of 5 is assigned. Furthermore, this “how” has a moderate relationship with providing customer service 24 hours a day and serving as customers' advocate in billing disputes, so a score of 3 is assigned for these relationships. Similar interpretations are drawn from the other entries in the table. “Hows” that have a large number of zeros do not support meeting the customer requirements and should be dropped from the list.

The cell values, shown in parentheses in Table 3-5, are obtained by multiplying the rated score by the importance rating of the corresponding customer requirement. The absolute score for each “How” is calculated by adding the values in parentheses. The relative score is merely a ranking of the absolute scores, with 1 representing the most important. It is observed that how 5 (effective recruiting) is most important because its absolute score of 57 is highest.

The analysis can be extended by considering the technical competitive assessment of the “hows.” Using the rating scores of the benchmark companies for each technical descriptor—that is, the objective values (the “how muches”) from the circled values in Table 3-4—our team can determine the importance of the “hows.” The weighted absolute scores in Table 3-5 are found by multiplying the corresponding absolute scores by the technical competitive assessment rating. The final scores demonstrate that the relative ratings of the top three “hows” are the same as before. However, the rankings of the remaining technical descriptors have changed. Hows 4 and 7 are tied for last place, each with an absolute score of 140 and a relative score of 6.5 each. Management may consider the ease or difficulty of implementing these “hows” in order to break the tie.

Our example QFD exercise illustrates the importance of teamwork in this process. An enormous amount of information must be gathered, all of which promotes cross-functional understanding of the product or service design system. Target values of the technical descriptors or “hows” are then used to generate the next level of house of quality diagram, where they will become the “whats.” The QFD process proceeds by determining the technical descriptors for these new “whats.” We can therefore consider implementation of the QFD process in different phases. As Figure 3-7 depicts, QFD facilitates the translation of customer requirements into a product whose features meet these requirements. Once such a product design is conceived, QFD may be used at the next level to identify specific characteristics of critical parts that will help in achieving the product designed. The next level may address the design of a process in order to make parts with the characteristics identified. Finally, the QFD process identifies production requirements for operating the process under specified conditions. Use of quality function deployment in such a multiphased environment requires a significant commitment of time and resources. However, the advantages—the spirit of teamwork, cross-functional understanding, and an enhanced product design—offset this commitment.

Benchmarking

As discussed earlier, the practice of identifying best practices in industry and thereby setting goals to emulate them is known as benchmarking. Companies cannot afford to stagnate; this guarantees a loss of market share to the competition. Continuous improvement is a mandate for survival, and such fast-paced improvement is facilitated by benchmarking. This practice enables an organization to accelerate its rate of improvement. While innovation allows an organization to “leapfrog” its competitors, it does not occur frequently and thus cannot be counted on. Benchmarking, on the other hand, is doable. To adopt the best, adapt it innovatively, and thus reap improvements is a strategy for success.

Specific steps for benchmarking vary from company to company, but the fundamental approach is the same. One company's benchmarking may not work at another organization because of different operating concerns. Successful benchmarking reflects the culture of the organization, works within the existing infrastructure, and is harmonious with the leadership philosophy. Motorola, Inc., winner of the Malcolm Baldrige Award for 1988, uses a five-step benchmarking model: (1) Decide what to benchmark; (2) select companies to benchmark; (3) obtain data and collect information; (4) analyze data and form action plans; and (5) recalibrate and start the process again.

AT&T, which has two Baldrige winners among its operating units, uses a nine-step model: (1) Decide what to benchmark; (2) develop a benchmarking plan; (3) select a method to collect data; (4) collect data; (5) select companies to benchmark; (6) collect data during a site visit; (7) compare processes, identify gaps, and make recommendations; (8) implement recommendations; and (9) recalibrate benchmarks.

A primary advantage of the benchmarking practice is that it promotes a thorough understanding of the company's own processes—the company's current profile is well understood. Intensive studies of existing practices often lead to identification of non-value-added activities and plans for process improvement. Second, benchmarking enables comparisons of performance measures in different dimensions, each with the best practices for that particular measure. It is not merely a comparison of the organization with a selected company, but a comparison with several companies that are the best for the measure chosen. Some common performance measures are return on assets, cycle time, percentage of on-time delivery, percentage of damaged goods, proportion of defects, and time spent on administrative functions. The spider chart shown in Figure 3-9 is used to compare multiple performance measures and gaps between the host company and industry benchmark practices. Six performance measures are being considered here. The scales are standardized: say, between 0 and 1, 0 being at the center and 1 at the outer circumference, which represents the most desired value. Best practices for each performance measure are indicated, along with the companies that achieve them. The current performance level of the company performing the benchmarking is also indicated in the figure. The difference between the company's level and that of the best practice for that performance measure is identified as the gap. The analysis that focuses on methods and processes to reduce this gap and thereby improve the company's competitive position is known as gap analysis.

Another advantage of benchmarking is its focus on performance measures and processes, not on the product. Thus, benchmarking is not restricted to the confines of the industry in which the company resides. It extends beyond these boundaries and identifies organizations in other industries that are superior with respect to the measure chosen. It is usually difficult to obtain data from direct competitors. However, companies outside the industry are more likely to share such information. It then becomes the task of management to find ways to adapt those best practices innovatively within their own environment.

In the United States, one of the pioneers of benchmarking is Xerox Corporation. It embarked on this process because its market share eroded rapidly in the late 1970s to Japanese competition. Engineers from Xerox took competitors' products apart and looked at them component by component. When they found a better design, they sought ways to adapt it to their own products or, even better, to improve on it. Similarly, managers from Xerox began studying the best management practices in the market; this included companies both within and outside the industry. As Xerox explored ways to improve its warehousing operations, it found a benchmark outside its own industry: L. L. Bean, Inc., the outdoor sporting goods retailer.

L. L. Bean has a reputation of high customer satisfaction; the attributes that support this reputation are its ability to fill customer orders quickly and efficiently with minimal errors and to deliver undamaged merchandise. The backbone behind this successful operation is an effective management system aided by state-of-the-art operations planning that addresses warehouse layout, workflow design, and scheduling. Furthermore, the operations side of the process is backed by an organizational culture of empowerment, management commitment through effective education and training, and a motivational reward system of incentive bonuses.

Figure 3-10 demonstrates how benchmarking brings the “soft” and “hard” systems together. Benchmarking is not merely identification of the best practices. Rather, it seeks to determine how such practices can be adapted to the organization. The real value of benchmarking is accomplished only when the company has integrated the identified best practices successfully into its operation. To be successful in this task, soft and hard systems must mesh. The emerging organizational culture should empower employees to make decisions based on the new practice.

For benchmarking to succeed, management must demonstrate its strategic commitment to continuous improvement and must also motivate employees through an adequate reward and recognition system that promotes learning and innovative adaptation. When dealing with hard systems, resources must be made available to allow release time from other activities, access to information on best practices, and installation of new information systems to manage the information acquired. Technical skills, required for benchmarking such as flowcharting and process mapping, should be provided to team members through training sessions. The team must also identify performance measures for which the benchmarking will take place. Examples of such measures are return on investment, profitability, cycle time, and defect rate.

Several factors influence the adoption of benchmarking; change management is one of them. Figure 3-11 illustrates factors that influence benchmarking and the subsequent outcomes that derive from it. In the current environment of global competition, change is a given. Rather than react haphazardly to change, benchmarking provides an effective way to manage it. Benchmarking provides a road map for adapting best practices, a major component of change management. These are process-oriented changes. In addition, benchmarking facilitates cultural changes in an organization. These deal with overcoming resistance to change. This is a people-oriented approach, the objective being to demonstrate that change is not a threat but an opportunity.

The ability to reduce process time and create a model of quick response is important to all organizations. The concept of time-based competition is linked to reductions in cycle time, which can be defined as the interval between the beginning and ending of a process, which may consist of a sequence of activities. From the customer's point of view, cycle time is the elapsed time between placing an order and having it fulfilled satisfactorially. Reducing cycle time is strongly correlated with performance measures such as cost, market share, and customer satisfaction. Detailed flowcharting of the process can identify bottlenecks, decision loops, and non-value-added activities. Reducing decision and inspection points, creating electronic media systems for dynamic flow of information, standardizing procedures and reporting forms, and consolidating purchases are examples of tactics that reduce cycle time. Motorola, Inc., for example, reduced its corporate auditing process over a three-year period from an average of seven weeks to five days.

Technological development is another impetus for benchmarking. Consider the micro-electronics industry. Its development pace is so rapid that a company has no choice but to benchmark. Falling behind the competition in this industry means going out of business. In this situation, benchmarking is critical to survival.

Quality Auditing

The effectiveness of management control programs may be examined through a practice known as quality auditing. One reason that management control programs are implemented is to prevent problems. Despite such control, however, problems can and do occur, so, quality audits are undertaken to identify problems.

In any quality audit, three parties are involved. The party that requests the audit is known as the client, the party that conducts the audit is the auditor, and the party being audited is the auditee. Auditors can be of two types, internal or external. An internal auditor is an employee of the auditee. External auditors are not members of the auditee's organization. An external auditor may be a single individual or a member of an independent auditing organization.

Quality audits fulfill two major purposes. The first purpose, performed in the suitability quality audit, deals with an in-depth evaluation of the quality program against a reference standard, usually predetermined by the client. Reference standards are set by several organizations, including the ANSI/ASQ, ISO, and British Standards Institute (BSI). Some ISO standards are discussed later in this chapter. The entire organization may be audited or specific processes, products, or services may be audited. The second purpose, performed in the conformity quality audit, deals with a thorough evaluation of the operations and activities within the quality system and the degree to which they conform to the quality policies and procedures defined.

Quality audits may be categorized as one of three types. The most extensive and inclusive type is the system audit. This entails an evaluation of the quality program documentation (including policies, procedures, operating instructions, defined accountabilities, and responsibilities to achieve the quality function) using a reference standard. It also includes an evaluation of the activities and operations that are implemented to accomplish the quality objectives desired. Such audits therefore explore conformance to quality management standards and their implementation to specified norms. They encompass the evaluation of the phases of planning, implementation, evaluation, and comparison. An example of a system audit is a pre-award survey, which typically evaluates the ability of a potential vendor to provide a desired level of product or service.

A second type of quality audit (not as extensive as the system audit) is the process audit, which is an in-depth evaluation of one or more processes in the organization. All relevant elements of the identified process are examined and compared to specified standards. Because a process audit takes less time to conduct than a system audit, it is more focused and less costly. If management has already identified a process that needs to be evaluated and improved, the process audit is an effective means of verifying compliance and suggesting places for improvement. A process audit can also be triggered by unexpected output from a process. For industries that use continuous manufacturing processes, such as chemical industries, a process audit is the audit of choice.

The third type of quality audit is the product audit, which is an assessment of a final product or service on its ability to meet or exceed customer expectations. This audit may involve conducting periodic tests on the product or obtaining information from the customer on a particular service. The objective of a product audit is to determine the effectiveness of the management control system. Such an audit is separate from decisions on product acceptance or rejection and is therefore not part of the inspection system used for such processes. Customer or consumer input plays a major role in the decision to undertake a product audit. For a company producing a variety of products, a relative comparison of product performance that indicates poor performers could be used as a guideline for a product audit.

Audit quality is heavily influenced by the independence and objectivity of the auditor. For the audit to be effective, the auditor must be independent of the activities being examined. Thus, whether the auditor is internal or external may have an influence on audit quality. Consider the assessment of an organization's quality documentation. It is quite difficult for an internal auditor to be sufficiently independent to perform this evaluation effectively. For such suitability audits, external auditors are preferable. System audits are also normally conducted by external auditors. Process audits can be internal or external, as can product audits. An example of an internal product audit is a dock audit, where the product is examined prior to shipment. Product audits conducted at the end of a process line are also usually internal audits. Product audits conducted at the customer site are typically external audits.

Vendor audits are external. They are performed by representatives of the company that is seeking the vendor's services. Knowledge of product and part specifications, contractual obligations and their secrecy, and purchase agreements often necessitate a second-party audit where the client company sends personnel from its own staff to perform the audit. Conformity quality audits may be carried out by internal or external auditors as long as the individuals are not directly involved in the activities being audited.

Methods for conducting a quality audit are of two types. One approach is to conduct an evaluation of all quality system activities at a particular location or operation within an organization, known as a location-oriented quality audit. This audit examines the actions and interactions of the elements in the quality program at that location and may be used to interpret differences between locations. The second approach is to examine and evaluate activities relating to a particular element or function within a quality program at all locations where it applies before moving on to the next function in the program. This is known as a function-oriented quality audit. Successive visits to each location are necessary to complete the latter audit. It is helpful in evaluating the overall effectiveness of the quality program and also useful in tracing the continuity of a particular function through the locations where it is applicable.

The utility of a quality audit is derived only when remedial actions in deficient areas, exposed by the quality audit, are undertaken by company management. A quality audit does not necessarily prescribe actions for improvement; it typically identifies areas that do not conform to prescribed standards and therefore need attention. If several areas are deficient, a company may prioritize those that require immediate attention. Only on implementation of the remedial actions will a company improve its competitive position. Tools that help identify critical areas, find root causes to problems, and propose solutions include cause-and-effect diagrams, flowcharts, and Pareto charts; these are discussed later in the chapter.

Vendor Selection and Certification Programs

As discussed in Chapter 2, the modern trend is to establish long-term relationships with vendors. In an organization's pursuit of continuous improvement, the purchaser (customer) and vendor must be integrated in a quality system that serves the strategic missions of both companies. The vendor must be informed of the purchaser's strategies for market-share improvement, advance product information (including changes in design), and delivery requirements. The purchaser, on the other hand, should have access to information on the vendor's processes and be advised of their unique capabilities.

Cultivating a partnership between purchaser and vendor has several advantages. First, it is a win–win situation for both. To meet unique customer requirements, a purchaser can then redesign products or components collaboratively with the vendor. The vendor, who makes those particular components, has intimate knowledge of the components and the necessary processes that will produce the desired improvements. The purchaser is thus able to design its own product in a cost-effective manner and can be confident that the design will be feasible to implement. Alternatively, the purchaser may give the performance specification to the vendor and entrust them with design, manufacture, and testing. The purchaser thereby reduces design and development costs, lowers internal costs, and gains access to proprietary technology through its vendor, technology that would be expensive to develop internally. Through such partnerships, the purchaser is able to focus on its areas of expertise, thereby maintaining its competitive edge. Vendors gain from such partnerships by taking ownership of the product or component from design to manufacture; they can meet specifications more effectively because of their involvement in the entire process. They also gain an expanded insight into product and purchaser requirements through linkage with the purchaser; this helps them better meet those requirements. This, in turn, strengthens the vendor's relationship with the purchaser.

Vendor Rating and Selection

Maintaining data on the continual performance of vendors requires an evaluation scheme. Vendor rating based on established performance measures facilitates this process. There are several advantages in monitoring vendor ratings. Since the quality of the output product is a function of the quality of the incoming raw material or components procured through vendors, it makes sense to establish long-term relationships with vendors that consistently meet or exceed performance requirements. Analyzing the historical performance of vendors enables the company to select vendors that deliver their goods on time. Vendor rating goes beyond reporting on the historical performance of the vendor. It ensures a disciplined material control program. Rating vendors also helps reduce quality costs by optimizing the cost of material purchased.

Measures of vendor performance, which comprise the rating scheme, address the three major categories of quality, cost, and delivery. Under quality, some common measures are percent defective as expressed by defects in parts per million, process capability, product stoppages due to poor quality of vendor components, number of customer complaints, and average level of customer satisfaction. The category of cost includes such measures as scrap and rework cost, return cost, incoming-inspection cost, life-cycle costs, and warranty costs. The vendor's maintenance of delivery schedules is important to the purchaser in order to meet customer-defined schedules. Some measures in this category are percent of on-time deliveries, percent of late deliveries, percent of early deliveries, percent of underorder quantity, and percent of overorder quantity.

Which measures should be used are influenced by the type of product or service, the customer's expectations, and the level of quality systems that exists in the vendor's organization. For example, the Federal Express Corporation, winner of the 1990 Malcolm Baldrige National Quality Award in the service category, is the name in fast and reliable delivery. FedEx tracks its performance with such measures as late delivery, invoice adjustment needed, damaged packages, missing proof of delivery on invoices, lost packages, and missed pickups. For incoming material inspection, defectives per shipment, inspection costs, and cost of returning shipment are suitable measures. For vendors with statistical process control systems in place, measuring process capability is also useful. Customer satisfaction indices can be used with those vendors that have extensive companywide quality systems in place.

Vendor performance measures are prioritized according to their importance to the purchaser. Thus, a weighting scheme similar to that described in the house of quality (Figure 3-5) is often used. Let's consider a purchaser that uses rework and scrap cost, price, percent of on-time delivery, and percent of underorder quantity as its key performance measures. Table 3-6 shows these performance measures and the relative weight assigned to each one. This company feels that rework and scrap costs are most important, with a weight of 40. Note that price is not the sole determinant; in fact, it received the lowest weighting.

Table 3-6 shows the evaluation of vendors, A, B, and C. For each performance measure, the vendors are rated on a scale of 1–5, with 1 representing the least desirable performance. A weighted score is obtained by adding the products of the weight and the assigned rating for each performance measure (weighted rating). The weighted scores are then ranked, with 1 denoting the most desirable vendor. From Table 3-6 we can see that vendor B, with the highest weighted score of 350, is the most desirable.

Vendor evaluation in quality programs is quite comprehensive. Even the vendor's culture is subject to evaluation as the purchaser seeks to verify the existence of a quality program. Commitment to customer satisfaction as demonstrated by appropriate actions is another attribute the purchaser will examine closely. The purchaser will measure the vendor's financial stability; the purchaser obviously prefers vendors that are going to continue to exist so the purchaser will not be visited with the problems that follow from liquidity or bankruptcy. The vendor's technical expertise relating to product and process design is another key concern as vendor and purchaser work together to solve problems and to promote continuous improvement.

Health Care Analytics and Big Data

Health-care-related data, be it on patients, physicians, hospitals and providers, research and evidence-based findings, and knowledge expansion through breakthroughs are compounding at an aggressive rate beyond comprehension. It may physically be impossible to keep up with all of the data on an individual basis. The concept of big data is a reality. The increased volume and velocity of such data are critical issues. Furthermore, a major challenge is the ability to integrate data from a variety of sources, some not necessarily compatible to each other, into a common platform on a dynamic basis and analyze the information to provide value to the provider, patient, and organization. This becomes a formidable task for health care analytics.

Application of health care analytics to big data may yield beneficial results to entities at several levels. Figure 3-12 shows some sequential steps in decision making, using big data, utilizing health care analytics. At the micro-level, data collected from data warehouses and other sources will often require some form of cleansing to make it compatible for decision making. In the simplest format, process data on a patient such as waiting time to registration, time to bed occupancy, and waiting time to see a physician may be kept on individual patients. On the other hand, a laboratory in a facility conducting blood tests may keep records on turnaround time by day and time of day and the order number, which is linked to a particular patient. Hence, these two data sets could be linked by the patient number, a commonality present in both. Further, electronic medical records (EMRs) on patients and facilities may be kept in a variety of formats. Data on some patients on some of the various process attributes could be missing, implying some form of data cleansing could be necessary.

Once data are structured in a format suitable for processing, a visual means for summarizing the information is usually a first step in the analysis. Development of dashboards such as average length of stay and mortality or morbidity rates is one form of visualization. Alternatively, summary graphs such as the histogram of length of stay of patients on a monthly basis is another form of data visualization.

At the next stage, say at the patient level, health care analytics could utilize the data to develop predictive models by the type of illness. Here, based on patient characteristics, risk-adjusted models, for example, could be developed for diabetic patients. Models using methods of regression analysis may be utilized in this context. Such information could be helpful to physicians for diagnosis as well as monitoring of patients. It could lead to changes in medications as well as patient advisement.

At the macro-level, analysis of various forms of health care data at the aggregate level could provide a status report on population health by state, region, country, or the world. Such information could shape the development of health care policies at the national level. Formulation of such policies based on prescriptive analytics of data could be utilized in population health management. They typically involve optimization techniques based on stated objectives and constraints. For example, for diabetic patients, guidelines for a maximum level of hemoglobin A1C levels could be prescribed along with those for cholesterol levels.

Uniqueness of Health Care

Challenges in Health Care Quality

Given the unique structure of the health care delivery system, there are some challenges to improving quality and reducing costs concurrently. A few of these are discussed.

Adoption and Integration of Health Information Technology (HIT)

As data continue to amass at an ever-increasing rate, the use of electronic health records (EHRs) or EMRs will be the only feasible option that will support decision making in an informed, timely, and error-free manner. The federal government through the CMS is strongly supporting this initiative.

While data input to EHR/EMR by physicians and providers is important, a critical issue is the development of a compatible health information technology platform that can assimilate and integrate various forms and types of data from different sources. Figure 3-13 demonstrates the challenges involved in this concept.

Presently, in creating EHR/EMR records there is not a single standard. Further, clinical data for the same patient who has multiple providers not in the same network may be stored in different formats. Additionally, operational data from health care facilities, laboratories, and pharmacies could be in varying formats. Financial data, important for billing and collection and estimation of profitability of the organization, could be in yet another format. The challenge is to create an HIT platform that can aggregate and integrate these structures into a compatible platform that is suitable for health care decision making.

Health Care Decision Support Systems

One of the major challenges in decision making in health care in the twenty-first century is the development of an adequate health care decision support system (HCDSS). With the rate at which knowledge in the field is expanding, it may not be feasible for individuals to keep up with this information and utilize it in their decision making without appropriate technology support. While it is true that a decision support system can only recommend actions based on historical evidence, the importance of the physician to integrate that information along with current knowledge, which is ongoing and dynamic, will always exist. Figure 3-14 displays the challenges and the benefits from the development of a HCDSS.

Aggregating and integrating information from various data warehouses as well as ongoing research and discoveries to create an integrated and dynamic knowledge base will require a coordinated and dedicated effort on a continual basis. This notion will support the concept of continuous quality improvement but may have the following barriers. High investment costs and lack of standards, such that most applications do not communicate well, leading to high costs of interfacing, are a couple. Maintaining privacy of patient records is another barrier as merging of information takes place (Bates and Gawande 2003).

An ongoing challenge is the development of an appropriate search engine, which is the backbone of a HCDSS. The search engine must be able to handle the volume of information and support timely decision making at the point of service (POS) level. It must be able to integrate the discoveries in the fields of genomics, proteomics, and pharmacogenomics, for example, to assist the health care provider.

With the adoption of an adequate search engine, the creation of a suitable HCDSS will occur. Methods of health care analytics will be utilized in the formulation of decisions to be depicted by the decision support system. The benefits from such a HCDSS will be realized at all levels. At the aggregate level, it may lead to improved population health. Such a DSS could project the amount of vaccination that would be necessary to stop the spread of a certain influenza virus in the population. At the patient and physician levels, major benefits may occur. Reductions in medication errors could occur as possible interactions between drugs to be taken by the patient are reported to the provider in a timely manner. Also, there is less chance of an error in the calculation of weight-based doses of medication. Additionally, as current information on the patient is input to the system, the DSS could provide a rapid response on certain adverse events such as nosocomial infections. As another example, remote monitoring of intensive care unit (ICU) patients could happen, which could aid in timely decisions. On a more general basis, with the reduction of medical errors and updates on waiting times in various units within the facilities, decisions recommended will improve the effectiveness and efficiency of the facilities.

Pareto Diagrams

Pareto diagrams are important tools in the quality improvement process. Alfredo Pareto, an Italian economist (1848–1923), found that wealth is concentrated in the hands of a few people. This observation led him to formulate the Pareto principle, which states that the majority of wealth is held by a disproportionately small segment of the population. In manufacturing or service organizations, for example, problem areas or defect types follow a similar distribution. Of all the problems that occur, only a few are quite frequent; the others seldom occur. These two problem areas are labeled the vital few and the trivial many. The Pareto principle also lends support to the 80/20 rule, which states that 80% of problems (nonconformities or defects) are created by 20% of causes. Pareto diagrams help prioritize problems by arranging them in decreasing order of importance. In an environment of limited resources, these diagrams help companies decide the order in which they should address problems.

Figure 3-15 shows a Pareto diagram of reasons for airline customer dissatisfaction. Delays in arrival is the major reason, as indicated by 40% of customers. Thus, this is the problem that the airlines should address first.

Flowcharts

Flowcharts, which show the sequence of events in a process, are used for manufacturing and service operations. They are often used to diagram operational procedures to simplify a system, as they can identify bottlenecks, redundant steps, and non-value-added activities. A realistic flowchart can be constructed by using the knowledge of the personnel who are directly involved in the particular process. Valuable process information is usually gained through the construction of flowcharts. Figure 3-16 shows a flowchart for patients reporting to the emergency department in a hospital. The chart identifies where delays can occur: for example, in several steps that involve waiting. A more detailed flowchart would allow pinpointing of key problem areas that contribute to lengthening waiting time.

Further, certain procedures could be modified or process operations could be combined to reduce waiting time. A detailed version of the flowchart is the process map, which identifies the following for each operation in a process: process inputs (e.g., material, equipment, personnel, measurement gage), process outputs (these could be the final results of the product or service), and process or product parameters (classified into the categories of controllable, procedural, or noise). Noise parameters are uncontrollable and could represent the in-flow rate of patients or the absenteeism of employees. Through discussion and data analysis, some of the parameters could be classified as critical. It will then be imperative to monitor the critical parameters to maintain or improve the process.

Cause-and-Effect Diagrams

Cause-and-effect diagrams were developed by Kaoru Ishikawa in 1943 and thus are often called Ishikawa diagrams. They are also known as fishbone diagrams because of their appearance (in the plotted form). Basically, cause-and-effect diagrams are used to identify and systematically list various causes that can be attributed to a problem (or an effect) (Ishikawa 1976). These diagrams thus help determine which of several causes has the greatest effect. A cause-and-effect diagram can aid in identifying the reasons why a process goes out of control. Alternatively, if a process is stable, these diagrams can help management decide which causes to investigate for process improvement. There are three main applications of cause-and-effect diagrams: cause enumeration, dispersion analysis, and process analysis.

Cause enumeration is usually developed through a brainstorming session in which all possible types of causes (however remote they may be) are listed to show their influence on the problems (or effect) in question. In dispersion analysis, each major cause is analyzed thoroughly by investigating the subcauses and their impact on the quality characteristic (or effect) in question. This process is repeated for each major cause in a prioritized order. The cause-and-effect diagram helps us analyze the reasons for any variability or dispersion. When cause-and-effect diagrams are constructed for process analysis, the emphasis is on listing the causes in the sequence in which the operations are actually conducted. This process is similar to creating a flow diagram, except that a cause-and-effect diagram lists in detail the causes that influence the quality characteristic of interest at each step of a process.

Using Minitab, for each Branch or main cause, create a column and enter the subcauses in the worksheet. Then, execute the following: Stat > Quality Tools > Cause-and-Effect. Under Causes, enter the name or column number of the main causes. The Label for each branch may be entered to match the column names. In Effect, input the brief problem description. Click OK. Figure 3-17 shows the completed cause-and-effect diagram.

Scatterplots

The simplest form of a scatterplot consists of plotting bivariate data to depict the relationship between two variables. When we analyze processes, the relationship between a controllable variable and a desired quality characteristic is frequently of importance. Knowing this relationship may help us decide how to set a controllable variable to achieve a desired level for the output characteristic. Scatterplots are often used as follow-ups to a cause-and-effect analysis.

Using Minitab, choose the commands Graph > Scatterplot. Select Simple and click OK. Under Y, enter the column number or name, in this case, “Tool wear.” Under X, enter the column number or name, in this case, “Depth of cut.” Click OK.

The resulting scatterplot is shown in Figure 3-18. It gives us an idea of the relationship that exists between depth of cut and amount of tool wear. In this case the relationship is generally nonlinear. For depth-of-cut values of less than 3.0 mm, the tool wear rate seems to be constant, whereas with increases in depth of cut, tool wear starts increasing at an increasing rate. For depth-of-cut values above 4.5 mm, tool wear appears to increase drastically. This information will help us determine the depth of cut to use to minimize downtime due to tool changes.

Multivariable Charts

In most manufacturing or service operations, there are usually several variables or attributes that affect product or service quality. Since realistic problems usually have more than two variables, multivariable charts are useful means of displaying collective information.

Several types of multivariate charts are available (Blazek et al. 1987). One of these is known as a radial plot, or star, for which the variables of interest correspond to different rays emanating from a star. The length of each ray represents the magnitude of the variable.

Several process characteristics can be observed from Figure 3-19. First, from time 1 to time 2, an improvement in the process performance is seen, as indicated by a decline in the percentage nonconforming. Next, we can examine what changes in the controllable variables led to this improvement. We see that a decrease in temperature, an increase in both pressure and manganese content, and a basically constant level of silicon caused this reduction in the percentage nonconforming.

Other forms of multivariable plots (such as standardized stars, glyphs, trees, faces, and weathervanes) are conceptually similar to radial plots. For details on these forms, refer to Gnanadesikan (1977).

Matrix and Three-Dimensional Plots

Investigating quality improvement in products and processes often involves data that deal with more than two variables. With the exception of multivariable charts, the graphical methods discussed so far deal with only one or two variables. The matrix plot is a graphical option for situations with more than two variables. This plot depicts two-variable relationships between a number of variables all in one plot. As a two-dimensional matrix of separate plots, it enables us to conceptualize relationships among the variables. The Minitab software can produce matrix plots.

Using Minitab, the data are entered for the three variables in a worksheet. Next, choose Graph > Matrix Plot and Matrix of Plots Simple. Under Graph variables, input the variable names or column numbers. Click OK. The resulting matrix plot is shown in Figure 3-20. Observe that seal strength tends to increase linearly with temperature up to a certain point, which is about 210°C. Beyond 210°C, seal strength tends to decrease. The relationship between seal strength and pressure decreases with pressure. Also, the existing process conditions exhibit a relationship between temperature and pressure that decreases with pressure. Such graphical aids provide us with some insight on the relationship between the variables, taken two at a time.

Three-dimensional scatterplots depict the joint relationship of a dependent variable with two independent variables. While surface-plots demonstrate two-variable relationships, they do not show the joint effect of more than one variable on a third variable. Since interactions do occur between variables, a three-dimensional scatterplot is useful in identifying optimal process parameters based on a desired level of an output characteristic.

Failure Mode and Effects Criticality Analysis

Failure mode and effects criticality analysis (FMECA) is a disciplined procedure for systematic evaluation of the impact of potential failures and thereby to determine a priority of possible actions that will reduce the occurrence of such failures. It can be applied at the system level, at the design level for a product or service, at the process level for manufacturing or services, or at the functional level of a component or subsystem level.

In products involving safety issues, say the braking mechanism in automobiles, FMECA assists in a thorough analysis of what the various failure modes could be, their impact and effect on the customer, the severity of the failure, the possible causes that may lead to such a failure, the chance of occurrence of such failures, existing controls, and the chance of detection of such failures. Based on the information specified above, a risk priority number (RPN) is calculated, which indicates a relative priority scheme to address the various failures. The risk priority number is the product of the severity, occurrence, and detection ratings. The larger the RPN, the higher the priority. Consequently, recommended actions are proposed for each failure mode. Based on the selected action, the ratings on severity, occurrence, and detection are revised. The severity ratings typically do not change since a chosen action normally influences only the occurrence and/or detection of the failure. Only through fundamental design changes can the severity be reduced. Associated with the action selected, a rating that measures the risk associated with taking that action is incorporated. This rating on risk is a measure of the degree of successful implementation. Finally, a weighted risk priority number, which is the product of the revised ratings on severity, occurrence, and detection and the risk rating, is computed. This number provides management with a priority in the subsequent failure-related problems to address.

Several benefits may accrue from using failure modes and effects criticality analysis. First, by addressing all potential failures even before a product is sold or service rendered, there exists the ability to improve quality and reliability. A FMECA study may identify some fundamental design changes that must be addressed. This creates a better product in the first place rather than subsequent changes in the product and/or process. Once a better design is achieved, processes to create such a design can be emphasized. All of this leads to a reduction in development time of products and, consequently, costs. Since a FMECA involves a team effort, it leads to a thorough analysis and thereby identification of all possible failures. Finally, customer satisfaction is improved, with fewer failures being experienced by the customer.

It is important to decide on the level at which FMECA will be used since the degree of detailed analysis will be influenced by this selection. At the system level, usually undertaken prior to the introduction of either a product or service, the analysis may identify the general areas of focus for failure reduction. For a product, for example, this could be suppliers providing components, parts manufactured or assembled by the organization, or the information system that links all the units. A design FMECA is used to analyze product or service designs prior to production or operation. Similarly, a process FMECA could be used to analyze the processing/assembly of a product or the performance of a service. Thus, a hierarchy exists in FMECA use.

After selection of the level and scope of the FMECA, a block diagram that depicts the units/operations and their interrelationships is constructed, and the unit or operation to be studied is outlined. Let us illustrate use of FMECA through an example. Consider an OEM with one supplier who assembles to-order computers, as shown in Figure 3-22. There is flow of information and goods taking place in this chain. We restrict our focus to the OEM, where failure constitutes not meeting customer requirements regarding order quantity, quality, and delivery date.

Now, functional requirements are defined based on the selected level and scope. Table 3-10 lists these requirements based on customer's order quantity, quality, and delivery date. Through group brainstorming, potential failures for each functional requirement are listed. There could be more than one failure mode for each function. Here, for example, a failure in not meeting order quantity could occur due to the supplier and/or the OEM, as shown in Table 3-10. The impact or effects of failures are then listed. In the example, it leads to customer dissatisfaction. Also, for failures in order quantity or delivery date, another effect could be the creation of back orders, if permissible.

The next step involves rating the severity of the failure. Severity ratings are a measure of the impact of such failures on the customer. Such a rating is typically on a discrete scale from 1 (no effect) to 10 (hazardous effect). The Automotive Industry Action Group (AIAG) has some guidelines for severity ratings (El-Haik and Roy 2005), which other industries have adapted correspondingly. Table 3-11 shows rating scores on severity, occurrence, and detection, each of which is on a discrete scale of 1–10. AIAG has guidelines on occurrence and detection ratings as well, with appropriate modifications for process functions (Ehrlich 2002). For the example we indicate a severity rating of 6 on failure to meet order quantity or delivery date, while a rating of 7 is assigned to order quality, indicating that it has more impact on customer dissatisfaction, as shown in Table 3-10.

Causes of each failure are then listed, which will lead to suggestion of remedial actions. The next rating relates to the occurrence of failures; the larger the rating the more likely the possible happening. The guidelines in Table 3-11 could be used to select the rated value. Here, we deem that a failure in not meeting order quantity or delivery date at the supplier to be more likely (rating of 7 in occurrence) relative to that of the OEM (rating of 4). Further, in not meeting specified quality, we believe that this is less likely to happen at the supplier (supplier rating is 5) and even more remote at the OEM (rating of 3). Existing controls, to detect failures, are studied. Finally, the chance of existing controls detecting failures is indicated by a rating score. In this example, there is a moderately high chance (rating of 4) of detecting lack of capacity at the supplier's location through available capacity/inventory reports, whereas detecting the same at the OEM through capacity reports has a very high chance (rating of 2). A similar situation exists for detecting lack of quality, with the OEM having a high chance of detection (rating of 3) through matching of customer orders with product bar codes relative to that of the supplier (rating of 4) through process control and capability analysis. Finally, in Table 3-10, a risk priority number (RPN) is calculated for each failure mode and listed. Larger RPN values indicate higher priority to the corresponding failure modes. Here, we would address capacity issues at the supplier (RPN of 168), and not meeting due dates due to the supplier (RPN of 168) first, followed by lack of quality at the supplier (RPN of 140).

Continuing on with the FMECA analyses, the next step involves listing specific recommended actions for addressing each failure mode. Table 3-12 presents this phase of the analysis. The objectives of the recommended actions are to reduce the severity and/or occurrence of the failure modes and to increase their detection through appropriate controls such as evaluation techniques or detection equipment. Whereas severity can be reduced only through a design change, occurrence may be reduced through design or process improvements. Of the possible recommended actions, Table 3-12 lists the action taken for each failure mode and the revised ratings on severity, occurrence, and detection. In this example, with no design changes made, the severity ratings are the same as before. However, the occurrence has been lowered by the corresponding action taken. Also, for certain failure modes (e.g., lack of quality), through monitoring critical to quality (CTQ) characteristics, the chances of detection are improved (lower rating compared to those in Table 3-10). An RPN is calculated, which could then be used for further follow-up actions. Here, it seems that failure to meet order quantity due to lack of capacity at the supplier and the OEM is still a priority issue.

Associated with each action, a rating on a scale of 1–5 is used to indicate the risk of taking that action. Here, risk refers to the chance of implementing the action successfully, with a 1 indicating the smallest risk. Using this concept, observe from Table 3-12 that it is easier to add overtime at the OEM (risk rating of 1) compared to that at the supplier (risk rating of 3), since the OEM has more control over its operations. Similarly, for the OEM it is easier to add overtime than it is to reduce downtime (risk rating of 4). Finally, the last step involves multiplying the revised RPN by the risk factor to obtain a weighted risk priority number. The larger this number, the higher the priority associated with the failure mode and the corresponding remedial action. Management could use this as a means of choosing areas to address.

Features of ISO 9000 a

There are eight key management principles based on which the revisions of ISO 9000 and the associated ANSI/ISO/ASQ Q9000 have been incorporated. They reflect the philosophy and principles of total quality management, as discussed earlier in the chapter.

With the primary focus on an organization being to meet or exceed customer needs, a desirable shift in the standards has been toward addressing customer needs. This blends with our discussion of the philosophies of quality management that equally emphasize this aspect. Senior management must set the direction for the vision and mission of the company, with input from all levels in order to obtain the necessary buy-in of all people. As stated in one of the points of Deming's System of Profound Knowledge, optimization of the system (which may consist of suppliers and internal and external customers) is a desirable approach. Further, emphasis on improving the process based on observed data and information derived through analyses of the data, as advocated by Deming, is found to exist in the current revision of the standards. Finally, the principle of continuous improvement, advocated in the philosophy of total quality management, is also embraced in the standards.

ISO 9001 and ISO 9004 are a consistent pair. They are designed for use together but may be used independently, with their structures being similar. Two fundamental themes, customer-related processes and the concept of continual improvement, are visible in the new revision of the standards.

The standards have evolved to a focus on developing and managing effective processes from documenting procedures. An emphasis on the role of top management is viewed along with a data-driven process of identifying measurable objectives and measuring performance against them. Concepts of quality improvement discussed under the Shewhart (Deming) cycle of plan–do–check–act are integrated in the standards.

The new version of ISO 9000 standards follows a high-level structure with uniform use of core texts and terms. The focus on a process-oriented approach is adopted along with an inclusion of topics on risk management, change management, and knowledge management.

Other Industry Standards

Various industries are adopting to standards, similar to ISO 9000, but modified to meet their specific needs. A few of these standards are listed:

1. ISO/TS 16949. In the United States, the automotive industry comprised of the big three companies—Daimler, Ford, and General Motors adopted the ISO/TS 16949—Quality Management Systems—Particular Requirements for Automotive Production and Relevant Service Organizations, through the Automotive Industry Action Group (AIAG), standards, thereby eliminating the conflicting requirements for suppliers. Previously, each company had its own requirements for suppliers.
2. AS 9100. The aerospace industry, following a process similar to that used in the automotive industry, has developed the standard AS 9100, Quality Management Systems—Requirements for Aviation, Space, and Defense Organizations. These standards incorporate the features of ISO 9001 as well as the Federal Aviation Administration (FAA) Aircraft Certification System Evaluation Program and Boeing's massive D1-9000 variant of ISO 9000. Companies such as Boeing, Rolls-Royce Allison, and Pratt & Whitney use AS 9100 as the basic quality management system for their suppliers.
3. TL 9000. This is the standard developed in the telecommunications service industry to seek continuous improvement in quality and reliability. The Quality Excellence for Suppliers of Telecommunications (QuEST) Leadership Forum, formed by leading telecommunications service providers such as BellSouth, Bell Atlantic, Pacific Bell, and Southwestern Bell was instrumental in the creation of this standard. The membership now includes all regional Bell operating companies (RBOCs), AT&T, GTE, Bell Canada, and telecommunications suppliers such as Fujitsu Network Communications, Lucent Technologies, Motorola, and Nortel Networks. The globalization of the telecommunications industry has created a need for service providers and suppliers to implement common quality system requirements. The purpose of the standard is to effectively and efficiently manage hardware, software, and services by this industry. Through the adoption of such a standard, the intent is also to create cost- and performance-based metrics to evaluate efforts in the quality improvement area.
4. ISO 13485. The ISO 13485 standard is applicable to medical devices manufacturers and is a stand-alone standard.
5. Anticipated developments. OHSAS 18001—Occupational Health and Safety Zone is a set of international occupational health and safety management standards to help minimize risks to employees. ISO 45001—Occupational Health and Safety Management Standard is set to replace OHSAS 18001.