Operation Management Paper

operations_management_6e_ch17.pdf

Home >Business & Finance homework help >Operations Management homework help >Operation Management Paper

ee eBook Collection

Introduction Quality is the only one of the five ‘operations performance criteria’ to have its own dedicated chapter in this book. There are two reasons for this. First, in some organizations a separate function is devoted exclusively to the management of quality. Second, quality is a key concern of almost all organizations. High-quality goods and services can give an organization a considerable competitive edge. Good quality reduces the costs of rework, waste, complaints and returns and, most importantly, generates satisfied customers. Some operations managers believe that, in the long run, quality is the most important single factor affecting an organization’s performance relative to its competitors.

Chapter 17 Quality management

Key questions ➤ What is quality and why is it so

important?

➤ How can quality problems be diagnosed?

➤ What steps lead towards conformance to specification?

➤ What is total quality management (TQM)?

Figure 17.1 This chapter covers quality management

Check and improve your understanding of this chapter using self assessment questions and a personalised study plan, audio and video downloads, and an eBook – all at www.myomlab.com.

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka joka

Part Three Planning and control496

The first Four Seasons Hotel opened over 45 years ago. Since then the company has grown to 81 properties in 34 countries. Famed for its quality of service, the hotel group has won countless awards including the prestigious Zagat survey and numerous AAA Five Diamond Awards, and it is also one of only 14 organizations that have been on the Fortune magazine’s list of ‘100 Best Companies to Work For’ every year since it launched in 1998, thus ranking as ‘top hotel chain’ internationally. From its inception the group has had the same guiding principle, ‘to make the quality of our service our competitive advantage’. The company has what it calls its Golden Rule: ‘Do to others (guests and staff ) as you would wish others to do to you.’ It is a simple rule, but it guides the whole organization’s approach to quality.

‘Quality service is our distinguishing edge and the company continues to evolve in that direction. We are always looking for better, more creative and innovative ways of serving our guests’, says Michael Purtill, the General Manager of the Four Seasons Hotel Canary Wharf in London. ‘We have recently refined all of our operating standards across the company, enabling us to further enhance the personalized, intuitive service that all our guests receive. All employees are empowered to use their creativity and judgement in delivering exceptional service and making their own decisions to enhance our guests’ stay. For example, one morning an employee noticed that a guest had a flat tyre on their car and decided of his own accord to change it for them, which was very much appreciated by the guest.

The golden rule means that we treat our employees with dignity, respect and appreciation. This approach encourages them to be equally sensitive to our guests’ needs and offer sincere and genuine service that exceeds expectations. Just recently one of our employees accompanied a guest to the hospital and stayed there with him for the entire afternoon. He wanted to ensure that the guest wasn’t alone and was given the medical attention he needed. The following day that same employee took the initiative to return to the hospital (even though it was his day off ) to visit and made sure that that guest’s family in America was kept informed about his progress. We ensure that we have an ongoing focus on recognizing these successes and publicly praise and celebrate all individuals who deliver these warm, spontaneous, thoughtful touches.

At Four Seasons we believe that our greatest asset and strength is our people. We pay a great deal of

attention to selecting the right people with an attitude that takes great pride in delivering exceptional service. We know that motivated and happy employees are essential to our service culture and are committed to developing our employees to their highest potential. Our extensive training programmes and career development plans are designed with care and attention to support the individual needs of our employees as well as operational and business demands. In conjunction with traditional classroom-based learning, we offer tailor-made internet-based learning featuring exceptional quality courses for all levels of employee. Such importance is given to learning and development that the hotel has created two specialized rooms, designated for learning and development. One is intended for group learning and the other is equipped with private computer stations for internet-based individual learning. There is also a library equipped with a broad variety of hospitality-related books, CDs and DVDs that can be taken home at any time. This encourages our employees to learn and develop at an individual pace. This is very motivating for our employees and in the same instance their development is invaluable to the growth of our company. Career-wise, the sky is the limit and our goal is to build lifelong, international careers with Four Seasons.

Our objective is to exceed guest expectations and feedback from our guests and our employees is an invaluable barometer of our performance. We have created an in-house database that is used to record all guest feedback (whether positive or negative). We also use an online guest survey and guest comment cards which are all personally responded to and analysed to

Operations in practice Quality at the Four Seasons Canary Wharf1

S o

u rc

e :

F o

u r

S e a so

n s

H o

te ls

, P

h o

to g

ra p

h e r

R o

b e rt

M ill

e r

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

identify any potential service gaps. We continue to focus on delivering individual personalized experiences and our Guest History database remains vital in helping us to achieve this. All preferences and specific comments about service experience are logged on the database.

Every comment and every preference is discussed and planned for, for every guest, for every visit. It is our culture that sets Four Seasons apart: the drive to deliver the best service in the industry that keeps our guests returning again and again.’

Chapter 17 Quality management 497

What is quality and why is it so important?

It is worth revisiting some of the arguments which were presented in Chapter 2 regarding the benefits of high quality. This will explain why quality is seen as being so important by most operations. Figure 17.2 illustrates the various ways in which quality improvements can affect other aspects of operations performance. Revenues can be increased by better sales and enhanced prices in the market. At the same time, costs can be brought down by improved efficiencies, productivity and the use of capital. So, a key task of the operations function must be to ensure that it provides quality goods and services, to both its internal and external customers.

Figure 17.2 Higher quality has a beneficial effect on both revenues and costs Source: Based on Gummerson, E. (1993)2

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

The operation’s view of quality

There are many definitions of quality; here we define it as ‘consistent conformance to customers’ expectations’.

The use of the word ‘conformance’ implies that there is a need to meet a clear specification. Ensuring a product or service conforms to specification is a key operations task. ‘Consistent’ implies that conformance to specification is not an ad hoc event but that the product or service meets the specification because quality requirements are used to design and run the processes that produce products and services. The use of ‘customers’ expectations’ recognizes that the product or service must take the views of customers into account, which may be influenced by price. Also note the use of the word ‘expectations’ in this definition, rather than ‘needs’ or ‘wants’.

Customers’ view of quality

Past experiences, individual knowledge and history will all shape customers’ expectations. Furthermore, customers may each perceive a product or service in different ways. One person may perceive a long-haul flight as an exciting part of a holiday; the person on the next seat may see it as a necessary chore to get to a business meeting. So quality needs to be understood from a customer’s point of view because, to the customer, the quality of a particular product or service is whatever he or she perceives it to be. If the passengers on a skiing charter flight perceive it to be of good quality, despite long queues at check-in or cramped seating and poor meals, then the flight really is of good perceived quality.3 Also customers may be unable to judge the ‘technical’ specification of the service or product and so use surrogate measures as a basis for their perception of quality.4 For example, a customer may find it difficult to judge the technical quality of dental treatment, except insofar as it does not give any more trouble. The customer may therefore perceive quality in terms the attire and demeanour of the dentist and technician, décor of the surgery, and how they were treated.

Reconciling the operation’s and the customer’s views of quality

The operation’s view of quality is concerned with trying to meet customer expectations. The customer’s view of quality is what he or she perceives the product or service to be. To create a unified view, quality can be defined as the degree of fit between customers’ expecta- tions and customer perception of the product or service.5 Using this idea allows us to see the customers’ view of quality of (and, therefore, satisfaction with) the product or service as the result of the customers comparing their expectations of the product or service with

There are many definitions of quality

Customer expectations

Customer perception

Part Three Planning and control498

Figure 17.3 Perceived quality is governed by the magnitude and direction of the gap between customers’ expectations and their perceptions of the product or service

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

their perception of how it performs. This is not always straightforward; see the short case ‘Tea and Sympathy’. Also, if the product or service experience was better than expected then the customer is satisfied and quality is perceived to be high. If the product or service was less than his or her expectations then quality is low and the customer may be dissatisfied. If the product or service matches expectations then the perceived quality of the product or service is seen to be acceptable. These relationships are summarized in Figure 17.3.

Both customers’ expectations and perceptions are influenced by a number of factors, some of which cannot be controlled by the operation and some of which, to a certain extent, can be managed. Figure 17.4 shows some of the factors that will influence the gap between expectations and perceptions. This model of customer-perceived quality can help us under- stand how operations can manage quality and identifies some of the problems in so doing. The bottom part of the diagram represents the operation’s ‘domain’ of quality and the top part the customer’s ‘domain’. These two domains meet in the actual product or service, which is provided by the organization and experienced by the customer. Within the operation’s domain, management is responsible for designing the product or service and providing a specification of the quality to which the product or service has to be created. Within the customer’s domain, his or her expectations are shaped by such factors as previous experi- ences with the particular product or service, the marketing image provided by the organization and word-of-mouth information from other users. These expectations are internalized as a set of quality characteristics.

A customer’s view of quality is shaped by the gap between perception and expectation

Chapter 17 Quality management 499

Defining quality in terms of perception and expectation can sometimes reveal some surprising results. For example, Tea and Sympathy is a British restaurant and café in the heart of New York’s West Village. Over the last ten years it has become a fashionable landmark in a city with one of the broadest range of restaurants in the world. Yet it is tiny, around a dozen tables packed into an area little bigger than the average British sitting room. Not only expatriate Brits but also native New Yorkers and celebrities queue to get in. As the only British restaurant in New York, it has a novelty factor, but also it has become famous for the unusual nature of its service. ‘Everyone is treated in the same way’, says Nicky Perry, one of the two ex-Londoners who run it, ‘We have a firm policy that we don’t take any shit.’ This robust attitude to the treatment of customers is reinforced by ‘Nicky’s Rules’ which are printed on the menu.

1 Be pleasant to the waitresses – remember Tea and Sympathy girls are always right.

2 You will have to wait outside the restaurant until your entire party is present – no exceptions.

3 Occasionally, you may be asked to change tables so that we can accommodate all of you.

4 If we don’t need the table you may stay all day, but if people are waiting it’s time to naff off.

Short case Tea and Sympathy6

5 These rules are strictly enforced. Any argument will incur Nicky’s wrath. You have been warned.

Most of the waitresses are also British and enforce Nicky’s Rules strictly. If customers object they are thrown out. Nicky says that she has had to train ‘her girls’ to toughen up. ‘I’ve taught them that when people cross the line they can tear their throats out as far as I’m concerned. What we’ve discovered over the years is that if you are really sweet, people see it as a weakness. People get thrown out of the restaurant about twice a week and yet customers still queue for the genuine shepherd’s pie, a real cup of tea, and of course the service.’

S o

u rc

e :

C o

rb is

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Part Three Planning and control500

Figure 17.4 The customer’s domain and the operations domain in determining the perceived quality, showing how the gap between customers’ expectations and their perception of a product or service could be explained by one or more gaps elsewhere in the model Source: Adapted from Parasuraman, A. et al. (1985) A conceptual model of service quality and implications for future research, Journal of Marketing, vol. 49, Fall, pp. 41–50. Reproduced with permission from the American Marketing Association.

Magic Moments is a small, but successful wedding photography business. Its owner, Richard Webber, has seen plenty of changes over the last twenty years. ‘In the past, my job involved taking a few photos during the wedding ceremony and then formal group shots outside. I was rarely at a wedding for more than two hours. Clients would select around 30 photos to go in a standard wedding album. It was important to get the photos right,

Short case Quality at Magic Moments

S o

u rc

e :

A la

m y

Im a g

e s

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Chapter 17 Quality management 501

because that was really the only thing I was judged on. Now it’s different. I usually spend all day at a wedding, and sometimes late into the evening as well. This creates a very different dynamic with the wedding party, as you’re almost like another guest. Whilst the bride and groom are still my primary concern, other guests at the wedding are also important. The challenge is to find the right balance between getting the best photos possible whilst being as discreet as possible. I could spend hours getting the perfect picture, but annoy everyone in the process. It’s difficult, because clients judge you on both the technical quality of your work and the way you interact with everyone on the day. The product has changed too. Clients receive a CD or memory stick with around 500 photos taken during the day. Also I can give them a choice of 10 albums in different sizes, ranging from 30 to 100 photos. This year, I have started offering photo books which allow a much greater level of customization and have proved popular for younger couples. For the future, I’m considering offering albums with wedding items such as invitations, confetti and menus, and individual paintings created from photographs. Obviously I would have to

outsource the paintings. I’m also going to upgrade our web site, so wedding guests can order photos and related products online. This will generate revenue and act as a good marketing tool. My anxiety is that advertising this additional service at the wedding will be seen as being too commercial, even if it’s actually of benefit to guests.

One of the biggest problems for the business is the high level of demand in the summer months. Weekends in June, July and August are often booked up two years in advance. One option is to take on additional photographers during busy periods. However, the best ones are busy themselves. The concern is that the quality of the service I offer would deteriorate. A large part of the business is about how one relates to clients and that’s hard to replicate. Having been to so many weddings, I often offer clients advice on various aspects of their wedding, such as locations, bands, caterers and florists. However, with development, wedding planning is clearly an area that could be profitable to the business. Of course, another option is to move beyond weddings into other areas, such as school photos, birthdays, celebrations, or studio work.’

Diagnosing quality problems7

Figure 17.4 shows how quality problems can be diagnosed. If the perceived quality gap is such that customers’ perceptions of the product or service fail to match their expectations of it, then the reason (or reasons) must lie in other gaps elsewhere in the model as follows.

Gap 1: The customer’s specification–operation’s specification gap. Perceived quality could be poor because there may be a mismatch between the organization’s own internal quality specification and the specification which is expected by the customer. For example, a car may be designed to need servicing every 10,000 kilometres but the customer may expect 15,000-kilometre service intervals.

Gap 2: The concept–specification gap. Perceived quality could be poor because there is a mismatch between the product or service concept (see Chapter 5) and the way the organiza- tion has specified quality internally. For example, the concept of a car might have been for an inexpensive, energy-efficient means of transportation, but the inclusion of a climate con- trol system may have both added to its cost and made it less energy-efficient.

Gap 3: The quality-specification–actual-quality gap. Perceived quality could be poor because there is a mismatch between actual quality and the internal quality specification (often called ‘conformance to specification’). For example, the internal quality specification for a car may be that the gap between its doors and body, when closed, must not exceed 7 mm. However, because of inadequate equipment, the gap in reality is 9 mm.

Gap 4: The actual-quality–communicated-image gap. Perceived quality could be poor because there is a gap between the organization’s external communications or market image and the actual quality delivered to the customer. This may be because the marketing function has set unachievable expectations or operations is not capable of the level of quality expected by the customer. For example, an advertising campaign for an airline might show a cabin attendant offering to replace a customer’s shirt on which food or drink has been spilt, whereas such a service may not in fact be available should this happen.

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Part Three Planning and control502

Conformance to specification

Conformance to specification means producing a product or providing a service to its design specification. It is usually seen as the most important contribution that operations manage- ment can make to the customer’s perception of quality. We shall examine how it can be achieved in the remainder of this chapter by describing quality management as six sequential steps. This chapter, and Chapters 18, 19 and 20, will deal with these steps.

Step 1 Define the quality characteristics of the product or service. Step 2 Decide how to measure each quality characteristic. Step 3 Set quality standards for each quality characteristic. Step 4 Control quality against those standards. Step 5 Find and correct causes of poor quality. Step 6 Continue to make improvements.

Step 1 – Define the quality characteristics

Much of the ‘quality’ of a product or service will have been specified in its design. But not all the design details are useful in controlling quality. For example, the design of a television may specify that its outer cabinet is made with a particular veneer. Each television is not checked, however, to make sure that the cabinet is indeed made from that particular veneer. Rather it is the consequences of the design specification which are examined – the appearance of the cabinet, for example. These consequences for quality planning and control of the design are called the quality characteristics of the product or service. Table 17.1 shows a list of the quality characteristics which are generally useful.

Table 17.1 Quality characteristics for a car, a bank loan and an air journey

Quality characteristic

Functionality – how well the product or service does its job

Appearance – the sensory characteristics of the product or service: its aesthetic appeal, look, feel, etc.

Reliability – the consistency of the product’s or service’s performance over time

Durability – the total useful life of the product or service

Recovery – the ease with which problems with the product or service can be resolved

Contact – the nature of the person-to-person contact which might take place

Air journey (customer transformation process)

Safety and duration of journey, onboard meals and drinks, car and hotel booking services

Décor and cleanliness of aircraft, lounges and crew

Keeping to the published flight times

Keeping up with trends in the industry

Resolution of service failures

Knowledge, courtesy and sensitivity of airline staff

Bank loan (information transformation process)

Interest rate, terms and conditions

Aesthetics of information, web site, etc.

Keeping promises (implicit and explicit)

Stability of terms and conditions

Resolution of service failures

Knowledge and courtesy of branch and call centre staff

Car (material transformation process)

Speed, acceleration, fuel consumption, ride quality, road-holding, etc.

Aesthetics, shape, finish, door gaps, etc.

Mean time to failure

Useful life (with repair)

Ease of repair

Knowledge and courtesy of sales staff

Quality characteristics

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Quality characteristics of the total package

Many services are a whole package of several elements, each of which will have its own quality characteristics. Some aspects of quality may be influenced by two or more elements within the total package. To understand the quality characteristics of the whole package therefore it is necessary to understand the individual characteristics within and between each element of the package. For example, Figure 17.5 shows some of the quality characteristics for a web-based online grocery shopping service. To judge this service it is necessary to consider the web site through which information is transmitted and orders are placed, the products that are sold through the site and the delivery service that transports purchases to the customer. Identifying where each characteristic of quality lies is useful because it is the first step towards understanding which part of the total service should be given responsibility for maintaining each aspect of quality.

Step 2 – Decide how to measure each characteristic

These characteristics must be defined in such a way as to enable them to be measured and then controlled. This involves taking a very general quality characteristic such as ‘appearance’ and breaking it down, as far as one can, into its constituent elements. ‘Appearance’ is diffi- cult to measure as such, but ‘colour match’, ‘surface finish’ and ‘number of visible scratches’ are all capable of being described in a more objective manner. They may even be quantifiable. Other quality characteristics pose more difficulty. The ‘courtesy’ of airline staff, for example,

Chapter 17 Quality management 503

Figure 17.5 Some quality characteristics for an online grocery shopping service

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

has no objective quantified measure. Yet operations with high customer contact, such as airlines, place a great deal of importance on the need to ensure courtesy in their staff. In cases like this, the operation will have to attempt to measure customer perceptions of courtesy.

Variables and attributes

The measures used by operations to describe quality characteristics are of two types: variables and attributes. Variable measures are those that can be measured on a continu- ously variable scale (for example, length, diameter, weight or time). Attributes are those which are assessed by judgement and are dichotomous, i.e. have two states (for example, right or wrong, works or does not work, looks OK or not OK). Table 17.2 categorizes some of the measures which might be used for the quality characteristics of the car and the airline journey.

Step 3 – Set quality standards

When operations managers have identified how any quality characteristic can be measured, they need a quality standard against which it can be checked; otherwise they will not know whether it indicates good or bad performance. The quality standard is that level of quality which defines the boundary between acceptable and unacceptable. Such standards may well be constrained by operational factors such as the state of technology in the factory, and the cost limits of making the product. At the same time, however, they need to be appropriate to the expectations of customers. But quality judgements can be difficult. If one airline passenger out of every 10,000 complains about the food, is that good because 9,999 passengers out of 10,000 are satisfied? Or is it bad because, if one passenger complains, there must be others who, although dissatisfied, did not bother to complain? And if that level of complaint is similar for other airlines, should it regard its quality as satisfactory?

Variables

Attributes

Part Three Planning and control504

Table 17.2 Variable and attribute measures for quality characteristics

Quality characteristic

Functionality

Appearance

Reliability

Durability

Recovery

Contact

Attribute

Was the food acceptable?

Is the crew dressed smartly?

Were there any complaints?

Generally, is the airline updating its services in a satisfactory manner?

Do customers feel that staff deal satisfactorily with complaints?

Did customers feel that the staff were helpful (yes or no)?

Airline journey

Variable

Number of journeys which actually arrived at the destination (i.e. didn’t crash!)

Number of seats not cleaned satisfactorily

Proportion of journeys which arrived on time

Number of times service innovations lagged competitors

Proportion of service failures resolved satisfactorily

The extent to which customers feel well treated by staff (1 to 5 scale)

Attribute

Is the ride quality satisfactory?

Is the colour to specification?

Is the reliability satisfactory?

Is the useful life as predicted?

Is the serviceability of the car acceptable?

Did customers feel well served (yes or no)?

Car

Variable

Acceleration and braking characteristics from test bed

Number of blemishes visible on car

Average time between faults

Life of the car

Time from fault discovered to fault repaired

Level of help provided by sales staff (1 to 5 scale)

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Step 4 – Control quality against those standards

After setting up appropriate standards the operation will then need to check that the pro- ducts or services conform to those standards: doing things right, first time, every time. This involves three decisions:

1 Where in the operation should they check that it is conforming to standards? 2 Should they check every product or service or take a sample? 3 How should the checks be performed?

Chapter 17 Quality management 505

Starting with less than half a hectare of watercress beds, Vitacress now has more than 300,000 sq m of growing beds and employs over 1,000 full-time staff, with farms in the south of the UK and in Portugal. The group is the world’s leading watercress producer. Now owned by RAR Group, a private Portuguese company, it supplies baby leaf salads in the UK and is a major European grower and packer of salads and speciality vegetables for major supermarkets’ own-label products. Since 2003 it has also sold premium products under its own ‘Vitacress’ brand.

The market for pre-packed baby leaves is growing. It is worth more than £500 million per year in the UK alone and is forecast to continue expanding as the trend for convenient and healthy foods continues. Consumers are increasingly shunning the purchase of a single whole-head lettuce and are instead buying a bag of mixed baby leaves with a combination of different colours and flavours. But the fresh salad market is a competitive one with customers demanding the very highest quality, and the quality control task is not straightforward with leaves grown and harvested from locations across the world to provide the supermarkets with fresh, bagged, salad 365 days of the year. So what does ‘quality’ mean for Vitacress, and how does it go about delivering it?

Both taste and appearance are important to consumers, as they are to Vitacress’s retail customers, who also want the product to maintain its nutrient levels and healthy leaf appearance throughout its shelf life. The challenge facing Vitacress is that baby salad leaves are a highly delicate and perishable commodity, with damage to the leaves giving rise to an increased rate of breakdown and a reduced shelf life of the product. The leaves are subjected to potential damage at all stages in the processing chain from field to supermarket shelf, including washing, drying and packing. Understanding the science that underlies the growing, harvesting, packing, transportation and storage of their products is fundamental to maintaining quality levels. This is why Vitacress cooperates in university-based research projects that could impact final quality levels. For example, in one study supported by Vitacress, it

Short case Quality is vital at Vitacress8

was found that baby salad leaves had an increased post-harvest shelf life of from 1 to 6 days (depending on the variety), when harvested at the end of the day compared with leaves harvested at the start of the day. In another project, the effect of environmental legislation that reduced residual pesticide levels in fresh food posed a potential problem if insects or foreign bodies found their way into the final product. Mike Rushworth, Operations Director at Vitacress, contacted the University of Bath’s Faculty of Engineering Design for advice on improving their washing and inspection process. A new inspection system was designed that included devices to separate leaves without manual handling. The project also evaluated the leaf washing system, by introducing artificial ‘bugs’ to provide a continual check on the effectiveness of this system. The bugs contained transponders to transmit signals if they make it through the washing and inspection without being extracted.

Yet any natural food product will deteriorate over time, and no matter how inventive Vitacress are in processing their leaves, unless the product reaches the supermarket shelves soon after harvesting its effective ‘quality life’ will be truncated. This is why Vitacress takes only 24 hours to get salad goods from the field to the supermarket, chilling the crops within 60 minutes of harvesting to ensure maximum freshness. The company uses vacuum coolers on all its farms and speedy supply chains so products reach customers as fast as possible, an achievement recognized when Vitacress picked up the Zurich Best Factory of the Year award for good manufacturing practices at its Andover, UK plant.

S o

u rc

e :

A la

m y

Im a g

e s

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Where should the checks take place?

At the start of the process incoming resources may be inspected to make sure that they are to the correct specification. For example, a car manufacturer will check that components are of the right specification. A university will screen applicants to try to ensure that they have a high chance of getting through the programme. During the process checks may take place before a particularly costly process, prior to ‘difficult to check’, immediately after a process with a high defective rate, before potential damage or distress might be caused, and so on. Checks may also take place after the process itself to ensure that customers do not experience non-conformance.

Check every product and service or take a sample?

While it might seem ideal to check every single product or service, a sample may be more practical for a number of reasons.

● It might be dangerous to inspect everything. A doctor, for example, checks just a small sample of blood rather than taking all of a patient’s blood! The characteristics of this sample are taken to represent those of the rest of the patient’s blood.

● Checking everything might destroy the product or interfere with the service. Not every light bulb is checked for how long it lasts – it would destroy every bulb. Waiters do not check that customers are enjoying the meal every 30 seconds.

● Checking everything can be time-consuming and costly. It may not be feasible to check all output from a high-volume machine or to check the feelings of every bus commuter every day.

Also 100 per cent checking may not guarantee that all defects will be identified. Sometimes it is intrinsically difficult. For example, although a physician may undertake the correct testing procedure, he or she may not necessarily diagnose a (real) disease. Nor is it easy to notice everything. For example, try counting the number of ‘e’s on this page. Count them again and see if you get the same score.

Type I and type II errors

Although it reduces checking time, using a sample to make a decision about quality does have its own inherent problems. Like any decision activity, we may get the decision wrong. Take the example of a pedestrian waiting to cross a street. He or she has two main decisions: whether to continue waiting or to cross. If there is a satisfactory break in the traffic and the pedestrian crosses then a correct decision has been made. Similarly, if that person continues to wait because the traffic is too dense then he or she has again made a correct decision. There are two types of incorrect decisions or errors, however. One incorrect decision would be if he or she decides to cross when there is not an adequate break in the traffic, resulting in an accident – this is referred to as a type I error. Another incorrect decision would occur if he or she decides not to cross even though there was an adequate gap in the traffic – this is called a type II error. In crossing the road, therefore, there are four outcomes, which are summarized in Table 17.3.

Type I errors are those which occur when a decision was made to do something and the situation did not warrant it. Type II errors are those which occur when nothing was done, yet a decision to do something should have been taken as the situation did indeed warrant it. For example, if a school’s inspector checks the work of a sample of 20 out of 1,000 pupils and

Quality sampling

Part Three Planning and control506

Table 17.3 Type I and type II errors for a pedestrian crossing the road

Road conditions

Decision Unsafe Safe

Cross Type I error Correct decision

Wait Correct decision Type II error

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

all 20 of the pupils in the sample have failed, the inspector might draw the conclusion that all the pupils have failed. In fact, the sample just happened to contain 20 out of the 50 students who had failed the course. The inspector, by assuming a high fail rate would be making a type I error. Alternatively, if the inspector checked 20 pieces of work all of which were of a high standard, he or she might conclude that all the pupils’ work was good despite having been given, or having chosen, the only pieces of good work in the whole school. This would be a type II error. Although these situations are not likely, they are possible. Therefore any sampling procedure has to be aware of these risks (see the short case on, ‘Surgical statistics’).

How should the checks be performed?

In practice most operations will use some form of sampling to check the quality of their products or services. The decision then is what kind of sample procedure to adopt. There are two different methods in common use for checking the quality of a sample product or service so as to make inferences about all the output from an operation. Both methods take into account the statistical risks involved in sampling. The first, and by far the best known, is the procedure called statistical process control (SPC). SPC is concerned with sampling the process during the production of the goods or the delivery of service. Based on this sample, decisions are made as to whether the process is ‘in control’, that is, operating as it should be. The second method is called acceptance sampling and is more concerned with whether to regard an incoming or outgoing batch of materials or customers as acceptable or not. Both of these approaches are explained in the supplement to this chapter.

Steps 5 and 6 – Find and correct causes of poor quality and continue to make improvements

The final two steps in our list of quality management activities are, in some ways, the most important yet also the most difficult. They also blend into the general area of operations improvement. The material covered in Chapters 18, 19 and 20 all has contributions to make

Statistical process control

Acceptance sampling

Chapter 17 Quality management 507

Understanding the nature of type I and type II errors is an essential part of any surgeon’s quality planning. Take the well-known appendectomy operation, for example. This is the removal of the appendix when it becomes infected or inflamed. Removal is necessary because of the risk of the appendix bursting and causing peritonitis, a potentially fatal poisoning of the blood. The surgical procedure itself is a relatively simple operation with expected good results but there is always a small risk associated with any invasive surgery needing a general anaesthetic. In addition, like any surgical procedure, it is expensive. The cost of the USA’s approximately quarter-of-a-million appendectomies averages out to around $4,500 per operation. Unfortunately, appendicitis is difficult to diagnose accurately. Using standard X-ray procedures a definite diagnosis can only be obtained about 10 per cent of the time. But now a new technique, developed in the Massachusetts General Hospital in Boston, claims to be able to identify 100 per cent of true appendicitis cases before surgery is carried out. The new technique

Short case Surgical statistics9

(Focused Appendix Computed Tomography) uses spiral X-ray images together with a special dye. It scans only the relevant part of the body, so exposure to radiation is not as major an issue as with conventional X-ray techniques. The technique can also help in providing an alternative diagnosis when an appendectomy is not needed. Most significantly, the potential cost savings are very great. The test itself costs less than $250, which means that one single avoided surgery pays for around 20 tests.

S o

u rc

e :

C o

rb is

/R o

b e rt

L le

w e lly

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Part Three Planning and control508

to these two steps. Nevertheless, there is an aspect of quality management that has been par- ticularly important in shaping how quality is improved and the improvement activity made self-sustaining. This is total quality management (TQM). The remainder of the main body of this chapter is devoted to TQM.

Total quality management (TQM)

Total quality management (TQM) was one of the earliest of the current wave of management ‘fashions’. Its peak of popularity was in the late 1980s and early 1990s. As such it has suffered from something of a backlash in recent years and there is little doubt that many companies adopted TQM in the simplistic belief that it would transform their operations performance overnight. Yet the general precepts and principles that constitute TQM are still the dominant mode of organizing operations improvement. The approach we take here is to stress the import- ance of the ‘total’ in total quality management and how it can guide the agenda for improvement.

TQM as an extension of previous practice

TQM can be viewed as a logical extension of the way in which quality-related practice has pro- gressed (see Fig. 17.6). Originally quality was achieved by inspection – screening out defects before they were noticed by customers. The quality control (QC) concept developed a more systematic approach to not only detecting, but also treating quality problems. Quality assurance (QA) widened the responsibility for quality to include functions other than direct operations. It also made increasing use of more sophisticated statistical quality techniques. TQM included much of what went before but developed its own distinctive themes. We will use some of these themes to describe how TQM represents a clear shift from traditional approaches to quality.

What is TQM?

TQM is ‘an effective system for integrating the quality development, quality maintenance and quality improvement efforts of the various groups in an organization so as to enable production and service at the most economical levels which allow for full customer satisfac- tion’.10 However, it was the Japanese who first made the concept work on a wide scale and subsequently popularized the approach and the term ‘TQM’. It was then developed further by several, so-called, ‘quality gurus’. Each ‘guru’ stressed a different set of issues, from which emerged the TQM approach. It is best thought of as a philosophy of how to approach quality improvement. This philosophy, above everything, stresses the ‘total’ of TQM. It is an approach that puts quality at the heart of everything that is done by an operation and including all activities within an operation. This totality can be summarized by the way TQM lays particular stress on the following:

● meeting the needs and expectations of customers; ● covering all parts of the organization; ● including every person in the organization; ● examining all costs which are related to quality, especially failure costs and getting things

‘right first time’; ● developing the systems and procedures which support quality and improvement; ● developing a continuous process of improvement (this will be treated in the broader

context of improvement, in Chapter 18).

TQM means meeting the needs and expectations of customers

Earlier in this chapter we defined quality as ‘consistent conformance to customers’ expectations’. Therefore any approach to quality management must necessarily include the customer perspective. In TQM this customer perspective is particularly important. It may be referred

Total quality management

TQM is a philosophy of how to approach the organization of quality improvement

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

to as ‘customer-centricity’ (discussed briefly in Chapter 18) or the ‘voice of the customer’. However it is called, TQM stresses the importance of starting with an insight into customer needs, wants, perceptions and preferences. This can then be translated into quality objectives and used to drive quality improvement.

TQM means covering all parts of the organization

For an organization to be truly effective, every single part of it, each department, each activity, and each person and each level, must work properly together, because every person and every activity affects and in turn is affected by others. One of the most powerful con- cepts that has emerged from various improvement approaches is the concept of the internal customer or supplier. This is recognition that everyone is a customer within the organization and consumes goods or services provided by other internal suppliers, and everyone is also an internal supplier of goods and services for other internal customers. The implication of this is that errors in the service provided within an organization will eventually affect the product or service which reaches the external customer.

Service-level agreements

Some organizations bring a degree of formality to the internal customer concept by encour- aging (or requiring) different parts of the operation to agree service-level agreements (SLAs)

Chapter 17 Quality management 509

Figure 17.6 TQM as an extension

Service-level agreements

Internal customer or supplier

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

with each other. SLAs are formal definitions of the dimensions of service and the relationship between two parts of an organization. The type of issues which would be covered by such an agreement could include response times, the range of services, dependability of service supply, and so on. Boundaries of responsibility and appropriate performance measures could also be agreed. For example, an SLA between an information systems support unit and a research unit in the laboratories of a large company could define such performance measures as:

● the types of information network services which may be provided as ‘standard’; ● the range of special information services which may be available at different periods of

the day; ● the minimum ‘up-time’, i.e. the proportion of time the system will be available at different

periods of the day; ● the maximum response time and average response time to get the system fully operational

should it fail; ● the maximum response time to provide ‘special’ services, and so on.

TQM means including every person in the organization

Every person in the organization has the potential to contribute to quality. Although it may be necessary to develop some specialists to assist with maintaining quality levels, TQM was amongst the first approaches to stress the centrality of harnessing everyone’s impact on quality and therefore their potential contribution to quality. There is scope for creativity and innovation even in relatively routine activities, claim TQM proponents. The shift in attitude which is needed to view employees as the most valuable intellectual and creative resource which the organization possesses can still prove difficult for some organizations. When TQM practices first began to migrate from Japan in the late 1970s, the ideas seemed even more radical. Some Japanese industrialists even thought (mistakenly) that companies in Western economies would never manage to change. Take, for example, a statement by Konosuke Matsushito which attracted considerable publicity:

‘We are going to win and the industrial West is going to lose out – there is nothing much you can do about it, because the reasons for your failure are within yourselves. For you, the essence of management is getting the ideas out of the heads of bosses into the hands of labour. For us, the core of management is precisely the art of mobilizing and pulling together the intel- lectual resources of all employees in the service of the firm. Only by drawing on the combined brainpower of all its employees can a firm face up to the turbulence and constraints of today’s environment. That is why our large companies give their employees three to four times more training than yours. This is why they foster within the firm such intensive exchange and com- munication. This is why they seek constantly everybody’s suggestions and why they demand from the educational system increasing numbers of graduates as well as bright and well- educated generalists, because these people are the lifeblood of industry.’11

Part Three Planning and control510

While some see the strength of SLAs as the degree of formality they bring to customer– supplier relationships, there are also some clear drawbacks. The first is that the ‘pseudo- contractual’ nature of the formal relationship can work against building partnerships (see Chapter 13). This is especially true if the SLA includes penalties for deviation from service standards. Indeed, the effect can sometimes be to inhibit rather than encourage joint improvement. The second, and related, problem is that SLAs, again because of their formal documented nature, tend to emphasize the ‘hard’ and measurable aspects of per- formance rather than the ‘softer’ but often more important aspects. So a telephone may be answered within four rings, but how the caller is treated, in terms of ‘friendliness’, may be far more important.

Critical commentary

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

TQM means all costs of quality are considered

The costs of controlling quality may not be small, whether the responsibility lies with each individual or a dedicated quality control department. It is therefore necessary to examine all the costs and benefits associated with quality (in fact ‘cost of quality’ is usually taken to refer to both costs and benefits of quality). These costs of quality are usually categorized as prevention costs, appraisal costs, internal failure costs and external failure costs.

Prevention costs are those costs incurred in trying to prevent problems, failures and errors from occurring in the first place. They include such things as:

● identifying potential problems and putting the process right before poor quality occurs; ● designing and improving the design of products and services and processes to reduce

quality problems; ● training and development of personnel in the best way to perform their jobs; ● process control through SPC.

Appraisal costs are those costs associated with controlling quality to check to see if problems or errors have occurred during and after the creation of the product or service. They might include such things as:

● the setting up of statistical acceptance sampling plans; ● the time and effort required to inspect inputs, processes and outputs; ● obtaining processing inspection and test data; ● investigating quality problems and providing quality reports; ● conducting customer surveys and quality audits.

Internal failure costs are failure costs associated with errors which are dealt with inside the operation. These costs might include such things as:

● the cost of scrapped parts and material; ● reworked parts and materials; ● the lost production time as a result of coping with errors; ● lack of concentration due to time spent troubleshooting rather than improvement.

External failure costs are those which are associated with an error going out of the operation to a customer. These costs include such things as:

● loss of customer goodwill affecting future business; ● aggrieved customers who may take up time; ● litigation (or payments to avoid litigation); ● guarantee and warranty costs; ● the cost to the company of providing excessive capability (too much coffee in the pack or

too much information to a client).

The relationship between quality costs

In traditional quality management it was assumed that failure costs reduce as the money spent on appraisal and prevention increases. Furthermore, it was assumed that there is an optimum amount of quality effort to be applied in any situation, which minimizes the total costs of quality. The argument is that there must be a point beyond which diminishing returns set in – that is, the cost of improving quality gets larger than the benefits which it brings. Figure 17.7(a) sums up this idea. As quality effort is increased, the costs of providing the effort – through extra quality controllers, inspection procedures, and so on – increases proportionally. At the same time, however, the cost of errors, faulty products, and so on, decreases because there are fewer of them. However, TQM proponents believe that this logic is flawed. First, it implies that failure and poor quality are acceptable. Why, TQM proponents argue, should any operation accept the inevitability of errors? Some occupations seem to be able to accept a zero-defect standard. No one accepts that pilots are allowed to crash a certain proportion of their aircraft, or that nurses will drop a certain proportion of the babies

Prevention costs

Appraisal costs

Internal failure costs

External failure costs

Chapter 17 Quality management 511

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

they deliver. Second, it assumes that costs are known and measurable. In fact putting realistic figures to the cost of quality is not a straightforward matter. Third, it is argued that failure costs in the traditional model are greatly underestimated. In particular, all the management time wasted by failures and the loss of concentration it causes are rarely accounted for. Fourth, it implies that prevention costs are inevitably high because it involves expensive inspection. But why should not quality be an integral part of everyone’s work rather than employing extra people to inspect. Finally, the ‘optimum-quality level’ approach, by accepting com- promise, does little to challenge operations managers and staff to find ways of improving quality. Put these corrections into the optimum-quality effort calculation and the picture looks very different (see Fig. 17.7b). If there is an ‘optimum’, it is a lot further to the right, in the direction of putting more effort (but not necessarily cost) into quality.

The TQM quality cost model

TQM rejects the optimum-quality level concept and strives to reduce all known and unknown failure costs by preventing errors and failure taking place. Rather than looking for ‘optimum’ levels of quality effort, TQM stresses the relative balance between different types of quality cost. Of the four cost categories, two (costs of prevention and costs of appraisal) are open to managerial influence, while the other two (internal costs of failure and external costs of

Part Three Planning and control512

Figure 17.7 (a) The traditional cost of quality model, and (b) the traditional cost of quality model with adjustments to reflect TQM criticisms

A story which illustrates the difference in attitude between a TQM and a non-TQM company has become almost a legend among TQM proponents. It concerns a plant in Ontario, Canada, of IBM, the computer company. It ordered a batch of components from a Japanese manufacturer and specified that the batch should have

Short case Deliberate defectives

an acceptable quality level (AQL) of three defective parts per thousand. When the parts arrived in Ontario they were accompanied by a letter which expressed the supplier’s bewilderment at being asked to supply defective parts as well as good ones. The letter also explained that they had found it difficult to make parts which were defective, but had indeed managed it. These three defective parts per thousand had been included and were wrapped separately for the convenience of the customer.

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

failure) show the consequences of changes in the first two. So, rather than placing most emphasis on appraisal (so that ‘bad products and service don’t get through to the customer’) TQM emphasizes prevention (to stop errors happening in the first place). That is because the more effort that is put into error prevention, the more internal and external failure costs are reduced. Then, once confidence has been firmly established, appraisal costs can be reduced. Eventually even prevention costs can be stepped down in absolute terms, though prevention remains a significant cost in relative terms. Figure 17.8 illustrates this idea. Initially total quality costs may rise as investment in some aspects of prevention – mainly training – is increased. However, a reduction in total costs can quickly follow.

Getting things ‘right first time’

Accepting the relationships between categories of quality cost as illustrated in Figure 17.8 has a particularly important implication for how quality is managed. It shifts the emphasis from reactive (waiting for something to happen) to proactive (doing something before anything happens). This change in the view of quality costs has come about with a movement from an inspect-in (appraisal-driven) approach to a design-in (getting it right first time) approach.

Developing the systems and procedures which support quality and improvement

The emphasis on highly formalized systems and procedures to support TQM has declined in recent years, yet one aspect is still active for many companies. This is the adoption of the ISO 9000 standard. And although ISO 9000 can be regarded as a stand-alone issue, it is very closely associated with TQM.

The ISO 9000 approach

The ISO 9000 series is a set of worldwide standards that establishes requirements for com- panies’ quality management systems. ISO 9000 is being used worldwide to provide a frame- work for quality assurance. Registration requires a third-party assessment of a company’s quality standards and procedures and regular audits are made to ensure that the systems do

Chapter 17 Quality management 513

Figure 17.8 Increasing the effort spent on preventing errors occurring in the first place brings a more than equivalent reduction in other cost categories

Getting it right first time

ISO 9000

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

not deteriorate. Its purpose when it was first framed was to provide an assurance to the pur- chasers of products or services that they have been produced in such a way that they meet their requirements. The best way to do this, it was argued, was to define the procedures, standards and characteristics of the management control system which governs the operation. Such a system would help to ensure that quality was ‘built into’ the operation’s transformation processes.

In 2000 ISO 9000 was substantially revised. Rather than using different standards for different functions within a business it took a ‘process’ approach that focused on outputs from any operation’s process rather than the detailed procedures that had dominated the pre- vious version of ISO 9000. This process orientation requires operations to define and record core processes and sub-processes (in a manner very similar to the ‘hierarchy of processes’ principle that was outlined in Chapter 1). In addition, processes are documented using the process mapping approach that was described in Chapter 4. Also, ISO 9000 (2000) stresses four other principles.

● Quality management should be customer-focused. Customer satisfaction should be measured through surveys and focus groups and improvement against customer standards should be documented.

● Quality performance should be measured. In particular, measures should relate both to processes that create products and services and customer satisfaction with those products and services. Furthermore, measured data should be analysed in order to understand processes.

● Quality management should be improvement-driven. Improvement must be demonstrated in both process performance and customer satisfaction.

● Top management must demonstrate their commitment to maintaining and continually improving management systems. This commitment should include communicating the importance of meeting customer and other requirements, establishing a quality policy and quality objectives, conducting management reviews to ensure the adherence to qual- ity policies, and ensuring the availability of the necessary resources to maintain quality systems.

ISO 9000 is seen as providing benefits both to the organizations adopting it (because it gives them detailed guidance on how to design their control procedures) and especially to customers (who have the assurance of knowing that the products and services they purchase are produced by an operation working to a defined standard). It may also provide a useful discipline to stick to ‘sensible’ process-oriented procedures which lead to error reduction, reduced customer complaints and reduced costs of quality, and may even identify existing procedures which are not necessary and can be eliminated. Moreover, gaining the certificate demonstrates that the company takes quality seriously; it therefore has a marketing benefit.

Part Three Planning and control514

Notwithstanding its widespread adoption (and its revision to take into account some of its perceived failings), ISO 9000 is not seen as beneficial by all authorities, and is still subject to some specific criticisms. These include the following:

● The continued use of standards and procedures encourages ‘management by manual’ and over-systematized decision-making.

● The whole process of documenting processes, writing procedures, training staff and conducting internal audits is expensive and time-consuming.

● Similarly, the time and cost of achieving and maintaining ISO 9000 registration are excessive.

● It is too formulaic. It encourages operations to substitute a ‘recipe’ for a more customized and creative approach to managing operations improvement.

Critical commentary

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Chapter 17 Quality management 515

Summary answers to key questions

Check and improve your understanding of this chapter using self assessment questions and a personalised study plan, audio and video downloads, and an eBook – all at www.myomlab.com.

➤ What is quality and why is it so important?

■ The definition of quality used in this book defines quality as ‘consistent conformance to cus- tomers’ expectations’.

➤ How can quality problems be diagnosed?

■ At a broad level, quality is best modelled as the gap between customers’ expectations con- cerning the product or service and their perceptions concerning the product or service.

■ Modelling quality this way will allow the development of a diagnostic tool which is based around the perception–expectation gap. Such a gap may be explained by four other gaps: – the gap between a customer’s specification and the operation’s specification; – the gap between the product or service concept and the way the organization has specified it; – the gap between the way quality has been specified and the actual delivered quality; – the gap between the actual delivered quality and the way the product or service has been

described to the customer.

➤ What steps lead towards conformance to specification?

■ There are six steps: – define quality characteristics; – decide how to measure each of the quality characteristics; – set quality standards for each characteristic; – control quality against these standards; – find and correct the causes of poor quality; – continue to make improvements.

■ Most quality planning and control involves sampling the operations performance in some way. Sampling can give rise to erroneous judgements which are classed as either type I or type II errors. Type I errors involve making corrections where none are needed. Type II errors involve not making corrections where they are in fact needed.

➤ What is total quality management ( TQM)?

■ TQM is ‘an effective system for integrating the quality development, quality maintenance and quality improvement efforts of the various groups in an organization so as to enable produc- tion and service at the most economical levels which allow for full customer satisfaction’.

■ It is best thought of as a philosophy that stresses the ‘total’ of TQM and puts quality at the heart of everything that is done by an operation.

■ ‘Total’ in TQM means the following: – meeting the needs and expectations of customers; – covering all parts of the organization; – including every person in the organization; – examining all costs which are related to quality, and getting things ‘right first time’; – developing the systems and procedures which support quality and improvement; – developing a continuous process of improvement.

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Part Three Planning and control516

‘Before the crisis the quality department was just for looks, we certainly weren’t used much for problem solving, the most we did was inspection. Data from the quality department was brought to the production meeting and they would all look at it, but no one was looking behind it’ (Quality Manager, Preston Plant).

The Preston plant of Rendall Graphics was located in Preston, Vancouver, across the con- tinent from their headquarters in Massachusetts. The plant had been bought from the Georgetown Corporation by Rendall in March 2000. Precision coated papers for ink-jet printers accounted for the majority of the plant’s output, especially paper for specialist uses. The plant used coating machines that allowed precise coatings to be applied. After coating, the conversion department cut the coated rolls to the final size and packed the sheets in small cartons.

The curl problem In late 1998 Hewlett-Packard (HP), the plant’s main customer for ink-jet paper, informed the plant of some problems it had encountered with paper curling under conditions of low humidity. There had been no customer complaints to HP, but their own personnel had noticed the problem, and they wanted it fixed. Over the next seven or eight months a team at the plant tried to solve the problem. Finally, in October of 1999 the team made recommendations for a revised and considerably improved coating formulation. By January 2000 the process was producing acceptably. However, 1999 had not been a good year for the plant. Although sales were reasonably buoyant the plant was making a loss of around $2 million for the year. In October 1999, Tom Branton, previously accountant for the business, was appointed as Managing Director.

Slipping out of control In the spring of 2000, productivity, scrap and re-work levels continued to be poor. In response to this the opera- tions management team increased the speed of the line and made a number of changes to operating practice in order to raise productivity.

‘Looking back, changes were made without any proper discipline, and there was no real concept of control. We were always meeting specification, yet we didn’t fully understand how close we really were to not being able to make it. The culture here said, “If it’s within specification then it’s OK” and we were very diligent in making sure that the product which was shipped was in specification. However, Hewlett-Packard gets “process charts” that

Case study Turnround at the Preston plant

enable them to see more or less exactly what is happening right inside your operation. We were also getting all the reports but none of them were being internalized, we were using them just to satisfy the customer. By contrast, HP have a statistically-based analytical mentality that says to itself, “You might be capable of making this product but we are thinking two or three product generations forward and asking ourselves, will you have the capability then, and do we want to invest in this relationship for the future?” ’ ( Tom Branton)

The spring of 2000 also saw two significant events. First, Hewlett-Packard asked the plant to bid for the con- tract to supply a new ink-jet platform, known as the Vector project, a contract that would secure healthy orders for several years. The second event was that the plant was acquired by Rendall.

‘What did Rendall see when they bought us? They saw a small plant on the Pacific coast losing lots of money.’ (Finance Manager, Preston Plant)

Rendall were not impressed by what they found at the Preston plant. It was making a loss and had only just escaped from incurring a major customer’s disapproval over the curl issue. If the plant did not get the Vector con- tract, its future looked bleak. Meanwhile the chief concern continued to be productivity. But also, once again, there were occasional complaints about quality levels. However HP’s attitude caused some bewilderment to the operations management team.

‘When HP asked questions about our process the operations guys would say, “Look we’re making roll after roll of paper, it’s within specification. What’s the problem?” ’ (Quality Manager, Preston Plant)

But it was not until summer that the full extent of HP’s disquiet was made. ‘I will never forget June of 2000. I was at a meeting with HP in Chicago. It was not even about

S o

u rc

e :

G e tt

y Im

a g

e s /D

ig it a l V

is io

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

ing to show results. Process were coming under control, quality levels were improving and, most importantly, per- sonnel both on the shop floor and in the management team were beginning to get into the ‘quality mode’ of thinking. Paradoxically, in spite of stopping the line periodically, the efficiency of the plant was also improving.

Yet the Preston team did not have time to enjoy their emerging success. In September of 2000 the plant learned that it would not get the Vector project because of their recent quality problems. Then Rendall decided to close the plant. ‘We were losing millions, we had lost the Vector project, and it was really no surprise. I told the senior management team and said that we would announce it probably in April of 2001. The real irony was that we knew that we had actually already turned the corner.’ ( Tom Branton)

Notwithstanding the closure decision, the management team in Preston set about the task of convincing Rendall that the plant could be viable. They figured it would take three things. First, it was vital that they continue to improve quality. Progressing with their quality initiative involved establishing full statistical process control (SPC).

Second, costs had to be brought down. Working on cost reduction was inevitably going to be painful. The first task was to get an understanding of what should be an appropriate level of operating costs. ‘We went through a zero-based assessment to decide what an ideal plant would look like, and the minimum number of people needed to run it’ ( Tom Branton).

By December of 2000 there were 40 per cent fewer people in the plant than two months earlier. All departments were affected. The quality department shrank more than most, moving from 22 people down to 6. ‘When the plant was considering down-sizing they asked me, “How can we run a lab with six technicians?” I said, “Easy. We just make good paper in the first place, and then we don’t have to inspect all the garbage. That alone would save an immense amount of time.” (Quality Manager, Preston Plant)

Third, the plant had to create a portfolio of new product ideas which could establish a greater confidence in future sales. Several new ideas were under active investigation. The most important of these was ‘Protowrap’, a wrap for newsprint that could be repulped. It was a product that was technically difficult. However, the plant’s newly acquired capabilities allowed the product to be made economically.

Out of the crisis In spite of their trauma, the plant’s management team faced Christmas of 2000 with increasing optimism. They had just made a profit for the first time for over two years. By spring of 2001 even HP, at a corporate level, was start- ing to take notice. It was becoming obvious that the Preston plant really had made a major change. More significantly, HP had asked the plant to bid for a new product. April 2001 was a good month for the plant. It had chalked up three months of profitability and HP formally gave the new

quality. But during the meeting one of their engineers handed me a control chart, one that we supplied with every batch of product. He said “Here’s your latest control chart. We think you’re out of control and you don’t know that you’re out of control and we think that we are looking at this data more than you are.” He was absolutely right, and I fully understood how serious the position was. We had our most important customer telling us we couldn’t run our processes just at the time we were trying to persuade them to give us the Vector contract.’ ( Tom Branton)

The crisis Tom immediately set about the task of bringing the plant back under control. They first of all decided to go back to the conditions which prevailed in the January, when the curl team’s recommendations had been implemented. This was the state before productivity pressures had caused the process to be adjusted. At the same time the team worked on ways of implementing unambiguous ‘shut-down rules’ that would allow operators to decide under what condi- tions a line should be halted if they were in doubt about the quality of the product they were making.

‘At one point in May of 2000 we had to throw away 64 jumbo rolls of out-of-specification product. That’s over $100,000 of product scrapped in one run. Basically that was because they had been afraid to shut the line down. Either that or they had tried to tweak the line while it was running to get rid of the defect. The shut-down guidelines in effect say, “We are not going to operate when we are not in a state of control”. Until then our operators just couldn’t win. If they failed to keep the machines running we would say, “You’ve got to keep productivity up”. If they kept the machines running but had quality problems as a result, we criticized them for making garbage. Now you get into far more trouble for violating process procedures than you do for not meeting productivity targets.’ (Engineer, Preston Plant)

This new approach needed to be matched by changes in the way the communications were managed in the plant.

‘We did two things that we had never done before. First, each production team started holding daily reviews of control chart data. Second, one day a month we took people away from production and debated the control chart data. Several people got nervous because we were not producing anything. But it was necessary. For the first time you got operators from the three shifts meeting together and talking about the control chart data and other quality issues. Just as significantly we invited HP up to attend these meetings. Remember these weren’t staged meetings, it was the first time these guys had met together and there was plenty of heated discussion, all of which the Hewlett-Packard representatives witnessed.’ (Engineer, Preston Plant)

At last something positive was happening in the plant and morale on the shop floor was buoyant. By September 2000 the results of the plant’s teams efforts were start-

Chapter 17 Quality management 517

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

contract to Preston. Also in April, Rendall reversed their decision to close the plant.

Questions 1 What are the most significant events in the story of

how the plant survived because of its adoption of quality-based principles?

2 The plant’s processes eventually were brought under control. What were the main benefits of this?

3 SPC is an operational level technique of ensuring quality conformance. How many of the benefits of bringing the plant under control would you class as strategic?

Part Three Planning and control518

These problems and applications will help to improve your analysis of operations. You can find more practice problems as well as worked examples and guided solutions on MyOMLab at www.myomlab.com.

(Read the supplement on statistical process control, before attempting these problems.)

A call centre for a bank answers customers’ queries about their loan arrangements. All calls are automatically timed by the call centre’s information system and the mean and standard deviation of call lengths is monitored periodically. The bank has decided that only on very rare occasions should calls be less than 0.5 minute because customers would think this was impolite even if the query was so simple that it could be answered in this time. Also, the bank reckoned that it was unlikely that any query should ever take more than 7 minutes to answer satisfactorily. The figures for last week’s calls show that the mean of all call lengths was 3.02 minutes and the standard deviation was 1.58 minutes. Calculate the Cp and the Cpk for the call centre process.

In the above call centre, if the mean call length changes to 3.2 minutes and the standard deviation is 0.9 minute, how does this affect the Cp and Cpk? Do you think this is an appropriate way for the bank to monitor its call centre performance?

A vaccine production company has invested in an automatic tester to monitor the impurity levels in its vaccines. Previously all testing was done by hand on a sample of batches of serum. According to the company’s specifications, all vaccine must have impurity levels of less than 0.03 milligram per 1,000 litres. In order to test the effectiveness of its new automatic sampling equipment, the company runs a number of batches through the process with known levels of impurity. The table below shows the level of impurity of each batch and whether the new process accepted or rejected the batch. From these data, estimate the Type I and Type II error levels for the process.

A utility has a department that does nothing but change the addresses of customers on the company’s information systems when customers move house. The process is deemed to be in control at the moment and a random sample of 2,000 transactions shows that 2.5 per cent of these transactions had some type of error. If the company is to use statistical process control to monitor error levels, calculate the mean, upper control level (UCL) and lower control level (LCL) for their SPC chart.

0.035 0.028 0.031 0.029 0.028 0.034 0.031 (rejected) (accepted) (accepted) (accepted) (accepted) (accepted) (accepted)

0.040 0.011 0.028 0.025 0.019 0.018 0.033 (rejected) (accepted) (rejected) (accepted) (accepted) (accepted) (rejected)

0.022 0.029 0.012 0.034 0.027 0.017 0.021 (accepted) (rejected) (accepted) (accepted) (accepted) (accepted) (accepted)

0.031 0.015 0.037 0.030 0.025 0.034 0.020 (rejected) (accepted) (rejected) (accepted) (accepted) (rejected) (accepted)

Problems and applications

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Find two products, one a manufactured food item (for example, a pack of breakfast cereals, packet of biscuits, etc.) and the other a domestic electrical item (for example, electric toaster, coffee maker, etc.).

(a) Identify the important quality characteristics for these two products. (b) How could each of these quality characteristics be specified? (c) How could each of these quality characteristics be measured?

Many organizations check up on their own level of quality by using ‘mystery shoppers’. This involves an employee of the company acting the role of a customer and recording how they are treated by the operation. Choose two or three high-visibility operations (for example, a cinema, a department store, the branch of a retail bank, etc.) and discuss how you would put together a mystery shopper approach to testing their quality. This may involve you determining the types of characteristics you would wish to observe, the way in which you would measure these characteristics, an appropriate sampling rate, and so on. Try out your mystery shopper plan by visiting these operations.

Chapter 17 Quality management 519

Dale, B.G. (ed.) (2003) Managing Quality, Blackwell, Oxford. This latest version of a long-established, comprehensive and authoritative text.

Garvin, D.A. (1988) Managing Quality, The Free Press, New York. Somewhat dated now but relates to our discussion at the beginning of this chapter.

George, M.L., Rowlands, D. and Kastle, B. (2003) What Is Lean Six Sigma? McGraw-Hill. Very much a quick intro- duction on what Lean Six Sigma is and how to use it.

Pande, P.S., Neuman, R.P. and Kavanagh, R.R. (2000) The Six Sigma Way, McGraw-Hill, New York. There are many books written by consultants for practising managers on the now fashionable Six Sigma Approach. This is as read- able as any.

Selected further reading

www.quality-foundation.co.uk / The British Quality Founda- tion is a not-for-profit organization promoting business excellence.

www.juran.com The Juran Institute’s mission statement is to provide clients with the concepts, methods and guidance for attaining leadership in quality.

www.asq.org/ The American Society for Quality site. Good professional insights.

www.opsman.org Lots of useful stuff.

www.quality.nist.gov/ American Quality Assurance Institute. Well-established institution for all types of business quality assurance.

www.gslis.utexas.edu/~rpollock /tqm.html Non-commercial site on total quality management with some good links.

www.iso.org/iso/en/ISOOnline.frontpage Site of the Interna- tional Standards Organization that runs the ISO 9000 and ISO 14000 families of standards. ISO 9000 has become an international reference for quality management requirements.

Useful web sites

Now that you have finished reading this chapter, why not visit MyOMLab at www.myomlab.com where you’ll find more learning resources to help you make the most of your studies and get a better grade?

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Supplement to Chapter 17

Statistical process control (SPC)

Figure S17.1 Charting trends in quality measures

Introduction

Statistical process control (SPC) is concerned with checking a product or service during its creation. If there is reason to believe that there is a problem with the process, then it can be stopped and the problem can be identified and rectified. For example, an international airport may regularly ask a sample of customers if the cleanliness of its restaurants is satisfactory. If an unacceptable number of customers in one sample are found to be unhappy, airport managers may have to consider improving their procedures. Similarly, a car manufacturer periodically will check whether a sample of door panels conforms to its standards so it knows whether the machinery which produces them is performing correctly.

Control charts

The value of SPC is not just to make checks of a single sample but to monitor the quality over a period of time. It does this by using control charts to see if the process seems to be performing as it should, or alternatively if it is ‘out of control’. If the process does seem to be going out of control, then steps can be taken before there is a problem. Actually, most operations chart their quality performance in some way. Figure S17.1, or something like it,

Control charts

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Supplement to Chapter 17 Statistical process control (SPC) 521

Figure S17.2 The natural variation in the filling process can be described by a normal distribution

could be found in almost any operation. The chart could, for example, represent the percentage of customers in a sample of 1,000 who, each month, were dissatisfied with the restaurant’s cleanliness. While the amount of dissatisfaction may be acceptably small, management should be concerned that it has been steadily increasing over time and may wish to investigate why this is so. In this case, the control chart is plotting an attribute measure of quality (satisfied or not). Looking for trends is an important use of control charts. If the trend suggests the pro- cess is getting steadily worse, then it will be worth investigating the process. If the trend is steadily improving, it may still be worthy of investigation to try to identify what is happening that is making the process better. This information might then be shared with other parts of the organization, or, on the other hand, the process might be stopped as the cause could be adding unnecessary expense to the operation.

Variation in process quality

Common causes

The processes charted in Figure S17.1 showed an upwards trend. But the trend was neither steady nor smooth: it varied, sometimes up, sometimes down. All processes vary to some extent. No machine will give precisely the same result each time it is used. People perform tasks slightly differently each time. Given this, it is not surprising that the measure of qual- ity will also vary. Variations which derive from these common causes can never be entirely eliminated (although they can be reduced). For example, if a machine is filling boxes with rice, it will not place exactly the same weight of rice in every box it fills. When the filling machine is in a stable condition (that is, no exceptional factors are influencing its behaviour) each box could be weighed and a histogram of the weights could be built up. Figure S17.2 shows how the histogram might develop. The first boxes weighed could lie anywhere within the natural variation of the process but are more likely to be close to the average weight

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

(see Fig. S17.2a). As more boxes are weighed they clearly show the tendency to be close to the process average (see Fig. S17.2b and c). After many boxes have been weighed they form a smoother distribution (Fig. S17.2d), which can be drawn as a histogram (Fig. S17.2e), which will approximate to the underlying process variation distribution (Fig. S17.2f ).

Usually this type of variation can be described by a normal distribution with 99.7 per cent of the variation lying within ±3 standard deviations. In this case the weight of rice in the boxes is described by a distribution with a mean of 206 grams and a standard deviation of 2 grams. The obvious question for any operations manager would be: ‘Is this variation in the process performance acceptable?’ The answer will depend on the acceptable range of weights which can be tolerated by the operation. This range is called the specification range. If the weight of rice in the box is too small then the organization might infringe labelling regulations; if it is too large, the organization is ‘giving away’ too much of its product for free.

Process capability

Process capability is a measure of the acceptability of the variation of the process. The simplest measure of capability (Cp) is given by the ratio of the specification range to the ‘natural’ variation of the process (i.e. ±3 standard deviations):

Cp =

where

UTL = the upper tolerance limit LTL = the lower tolerance limit

s = the standard deviation of the process variability.

Generally, if the Cp of a process is greater than 1, it is taken to indicate that the process is ‘capable’, and a Cp of less than 1 indicates that the process is not ‘capable’, assuming that the distribution is normal (see Fig. S17.3a, b and c).

The simple Cp measure assumes that the average of the process variation is at the mid- point of the specification range. Often the process average is offset from the specification range, however (see Fig. S17.3d). In such cases, one-sided capability indices are required to understand the capability of the process:

Upper one-sided index Cpu =

Lower one-sided index Cpl =

where X = the process average.

Sometimes only the lower of the two one-sided indices for a process is used to indicate its capability (Cpk):

Cpk = min(Cpu, Cpl)

X − LTL 3s

UTL − X 3s

UTL − LTL 6s

Part Three Planning and control522

Specification range

Process capability

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Supplement to Chapter 17 Statistical process control (SPC) 523

Figure S17.3 Process capability compares the natural variation of the process with the specification range which is required

In the case of the process filling boxes of rice, described previously, process capability can be calculated as follows:

Specification range = 214 − 198 = 16 g Natural variation of process = 6 × standard deviation

= 6 × 2 = 12 g Cp = process capability

= =

= 1.333

If the natural variation of the filling process changed to have a process average of 210 grams but the standard deviation of the process remained at 2 grams:

Cpu = = = 0.666

Cpl = = = 2.0

Cpk = min(0.666, 2.0) = 0.666

210 − 198 3 × 2

214 − 210 3 × 2

214 − 198 6 × 2

UTL − LTL 6s

Worked example

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Assignable causes of variation

Not all variation in processes is the result of common causes. There may be something wrong with the process which is assignable to a particular and preventable cause. Machinery may have worn or been set up badly. An untrained person may not be following prescribed pro- cedures. The causes of such variation are called assignable causes. The question is whether the results from any particular sample, when plotted on the control chart, simply represent the variation due to common causes or due to some specific and correctable assignable cause. Figure S17.4, for example, shows the control chart for the average impact resistance of samples of door panels taken over time. Like any process the results vary, but the last three points seem to be lower than usual. So, is this natural (common cause) variation, or the symptom of some more serious (assignable) cause?

To help make this decision, control limits can be added to the control chart (the red dashed lines) which indicates the expected extent of ‘common-cause’ variation. If any points lie outside these control limits (the shaded zone) then the process can be deemed out of control in the sense that variation is likely to be due to assignable causes. These control limits could be set intuitively by examining past variation during a period when the process was thought to be free of any variation which could be due to assignable causes. But control limits can also be set in a more statistically revealing manner. For example, if the process which tests door panels had been measured to determine the normal distribution which represents its common-cause variation, then control limits can be based on this distribution. Figure S17.4 also shows how control limits can be added; here they are put at ±3 standard deviations (of the population of sample means) away from the mean of sample averages. It shows that the probability of the final point on the chart being influenced by an assignable cause is very high indeed. When the process is exhibiting behaviour which is outside its normal ‘common-cause’ range, it is said to be ‘out of control’. Yet there is a small but finite chance that the (seemingly out of limits) point is just one of the rare but natural results at the tail of the distribution which describes perfectly normal behaviour. Stopping the process under these circumstances would represent a type I error because the process is actually in control. Alternatively, ignoring a result which in reality is due to an assignable cause is a type II error (see Table S17.1).

Control limits are usually set at three standard deviations either side of the population mean. This would mean that there is only a 0.3 per cent chance of any sample mean falling outside these limits by chance causes (that is, a chance of a type I error of 0.3 per cent). The control limits may be set at any distance from the population mean, but the closer the limits

Part Three Planning and control524

Figure S17.4 Control chart for the impact resistance of door panels, together with control limits

Control limits

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

are to the population mean, the higher the likelihood of investigating and trying to rectify a process which is actually problem-free. If the control limits are set at two standard deviations, the chance of a type I error increases to about 5 per cent. If the limits are set at one standard deviation then the chance of a type I error increases to 32 per cent. When the control limits are placed at ±3 standard deviations away from the mean of the distribution which describes ‘normal’ variation in the process, they are called the upper control limit (UCL) and lower control limit (LCL).

The Taguchi loss function

Genichi Taguchi proposed a resolution of both the criticisms of SPC described in the critical commentary box.12 He suggested that the central issue was the first problem – namely that the consequences of being ‘off-target’ (that is, deviating from the required process average performance) were inadequately described by simple control limits. Instead, he proposed a quality loss function (QLF) – a mathematical function which includes all the costs of poor quality. These include wastage, repair, inspection, service, warranty and generally, what he termed, ‘loss to society’ costs. This loss function is expressed as follows:

L = D2C

where

L = total loss to society costs D = deviation from target performance C = a constant.

Figure S17.5 illustrates the difference between the conventional and Taguchi approaches to interpreting process variability. The more graduated approach of the QLF also answers the second problem raised in the critical commentary box. With losses increasing quadratically as performance deviates from target, there is a natural tendency to progressively reduce pro- cess variability. This is sometimes called a target-oriented quality philosophy.

Supplement to Chapter 17 Statistical process control (SPC) 525

Table S17.1 Type I and type II errors in SPC

Actual process state

Decision In control Out of control

Stop process Type I error Correct decision

Leave alone Correct decision Type II error

When its originators first described SPC more than half a century ago, the key issue was only to decide whether a process was ‘in control’ or not. Now, we expect SPC to reflect common sense as well as statistical elegance and promote continuous operations improvement. This is why two (related) criticisms have been levelled at the traditional approach to SPC. The first is that SPC seems to assume that any values of process performance which lie within the control limits are equally acceptable, while any values outside the limits are not. However, surely a value close to the process average or ‘target’ value will be more accept- able than one only just within the control limits. For example, a service engineer arriving only 1 minute late is a far better ‘performance’ than one arriving 59 minutes late, even if the control limits are ‘quoted time ± one hour’. Also, arriving 59 minutes late would be almost as bad as 61 minutes late! Second, a process always within its control limits may not be deteriorating, but is it improving. So rather than seeing control limits as fixed, it would be better to view them as a reflection of how the process is being improved. We should expect any improving process to have progressively narrowing control limits.

Critical commentary

Upper control limit

Lower control limit

Quality loss function

Target-oriented quality

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Why variability is a bad thing

Assignable variation is a signal that something has changed in the process which therefore must be investigated. But normal variation is itself a problem because it masks any changes in process behaviour. Figure S17.6 shows the performance of two processes both of which are subjected to a change in their process behaviour at the same time. The process on the left has such a wide natural variation that it is not immediately apparent that any change has taken place. Eventually it will become apparent because the likelihood of process performance violating the lower (in this case) control limit has increased, but this may take some time. By contrast, the process on the right has a far narrower band of natural variation. Because of this, the same change in average performance is more easily noticed (both visually and statistically). So, the narrower the natural variation of a process, the more obvious are changes in the behaviour of that process. And the more obvious are process changes, the easier it is to understand how and why the process is behaving in a particular way. Accepting any variation in any process is, to some degree, admitting to ignorance of how that process works.

Part Three Planning and control526

Figure S17.5 The conventional and Taguchi views of the cost of variability

Figure S17.6 Low process variation allows changes in process performance to be readily detected

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Supplement to Chapter 17 Statistical process control (SPC) 527

Control charts for attributes

Attributes have only two states – ‘right’ and ‘wrong’, for example – so the statistic calculated is the proportion of wrongs (p) in a sample. (This statistic follows a binomial distribution.) Control charts using p are called ‘p-charts’. In calculating the limits, the population mean (F) – the actual, normal or expected proportion of ‘defectives’ or wrongs to rights – may not be known. Who knows, for example, the actual number of city commuters who are dissatisfied with their journey time? In such cases the population mean can be estimated from the average of the proportion of ‘defectives’ (F), from m samples each of n items, where m should be at least 30 and n should be at least 100:

F =

One standard deviation can then be estimated from:

The upper and lower control limits can then be set as:

UCL = F + 3 standard deviations LCL = F − 3 standard deviations.

Of course, the LCL cannot be negative, so when it is calculated to be so it should be rounded up to zero.

F(1 − F) n

p1 + p2 + p3 + . . . + pn

A credit card company deals with many hundreds of thousands of transactions every week. One of its measures of the quality of service it gives its customers is the dependability with which it mails customers’ monthly accounts. The quality standard it sets itself is that accounts should be mailed within two days of the ‘nominal post date’ which is specified to the customer. Every week the company samples 1,000 customer accounts and records the percentage which was not mailed within the standard time. When the process is working normally, only 2 per cent of accounts are mailed outside the specified period, that is, 2 per cent are ‘defective’.

Control limits for the process can be calculated as follows:

Mean proportion defective, F = 0.02 Sample size n = 1,000

Standard deviation s =

= 0.0044

With the control limits at F ± 3s :

Upper control limit (UCL) = 0.02 + 3(0.0044) = 0.0332 = 3.32%

and

lower control limit (LCL) = 0.02 − 3(0.0044) = 0.0068 = 0.68%

0.02(0.98)

1,000

F(1 − F) n

Worked example

➔

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Part Three Planning and control528

Figure S17.7 shows the company’s control chart for this measure of quality over the last few weeks, together with the calculated control limits. It also shows that the process is in control. Sometimes it is more convenient to plot the actual number of defects (c) rather than the proportion (or percentage) of defectives, on what is known as a c-chart. This is very similar to the p-chart but the sample size must be constant and the process mean and control limits are calculated using the following formulae:

Process mean E =

Control limits = E ± 3 c

where

c = number of defects m = number of samples

c1 + c2 + c3 + . . . + cm m

Figure S17.7 Control chart for the percentage of customer accounts which are mailed outside their two-day period

Control chart for variables

The most commonly used type of control chart employed to control variables is the B – R chart. In fact this is really two charts in one. One chart is used to control the sample average or mean (B ). The other is used to control the variation within the sample by measuring the range (R). The range is used because it is simpler to calculate than the standard deviation of the sample.

The means (B ) chart can pick up changes in the average output from the process being charted. Changes in the means chart would suggest that the process is drifting generally away from its supposed process average, although the variability inherent in the process may not have changed (see Fig. S17.8).

The range (R) chart plots the range of each sample, that is the difference between the largest and the smallest measurement in the samples. Monitoring sample range gives an indication of whether the variability of the process is changing, even when the process average remains constant (see Fig. S17.8).

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Control limits for variables control chart

As with attributes control charts, a statistical description of how the process operates under normal conditions (when there are no assignable causes) can be used to calculate control limits. The first task in calculating the control limits is to estimate the grand average or population mean (D) and average range (A) using m samples each of sample size n.

The population mean is estimated from the average of a large number (m) of sample means:

D =

The average range is estimated from the ranges of the large number of samples:

A =

The control limits for the sample means chart are:

Upper control limit (UCL) = D + A2A Lower control limit (LCL) = D − A2A

The control limits for the range charts are:

Upper control limit (UCL) = D4A Lower control limit (LCL) = D3A

The factors A2, D3 and D4 vary with sample size and are shown in Table S17.2. The LCL for the means chart may be negative (for example, temperature or profit may

be less than zero) but it may not be negative for a range chart (or the smallest measurement in the sample would be larger than the largest). If the calculation indicates a negative LCL for a range chart then the LCL should be set to zero.

R1 + R2 + . . . + Rm m

B1 + B2 + . . . + Bm m

Supplement to Chapter 17 Statistical process control (SPC) 529

Figure S17.8 The process mean or the process range (or both) can change over time

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Part Three Planning and control530

Table S17.2 Factors for the calculation of control limits

Sample size n A2 D3 D4

2 1.880 0 3.267 3 1.023 0 2.575 4 0.729 0 2.282 5 0.577 0 2.115 6 0.483 0 2.004 7 0.419 0.076 1.924 8 0.373 0.136 1.864 9 0.337 0.184 1.816

10 0.308 0.223 1.777 12 0.266 0.284 1.716 14 0.235 0.329 1.671 16 0.212 0.364 1.636 18 0.194 0.392 1.608 20 0.180 0.414 1.586 22 0.167 0.434 1.566 24 0.157 0.452 1.548

GAM (Groupe As Maquillage) is a contract cosmetics company, based in France but with plants around Europe, which manufactures and packs cosmetics and perfumes for other companies. One of its plants, in Ireland, operates a filling line which automatically fills plastic bottles with skin cream and seals the bottles with a screw-top cap. The tightness with which the screw-top cap is fixed is an important part of the quality of the filling line process. If the cap is screwed on too tightly, there is a danger that it will crack; if screwed on too loosely it might come loose when packed. Either outcome could cause leakage of the product during its journey between the factory and the customer. The Irish plant had received some complaints of product leakage which it suspected was caused by inconsistent fixing of the screw-top caps on its filling line. The ‘tightness’ of the screw tops could be measured by a simple test device which recorded the amount of turning force (torque) that was required to unfasten the tops. The company decided to take samples of the bottles coming out of the filling-line process, test them for their unfasten- ing torque and plot the results on a control chart. Several samples of four bottles were taken during a period when the process was regarded as being in control. The following data were calculated from this exercise:

The grand average of all samples D = 812 g /cm3 The average range of the sample A = 6 g /cm3

Control limits for the means (B) chart were calculated as follows:

UCL = D + A2A = 812 + (A2 × 6)

From Table 17.6, we know, for a sample size of four, A2 = 0.729. Thus:

UCL = 812 + (0.729 × 6) = 816.37

LCL = D − (A2A) = 812 – (0.729 × 6) = 807.63

Worked example

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Supplement to Chapter 17 Statistical process control (SPC) 531

Control limits for the range chart (R) were calculated as follows:

UCL = D4 × A = 2.282 × 6 = 13.69

LCL = D3A = 0 × 6 = 0

After calculating these averages and limits for the control chart, the company regularly took samples of four bottles during production, recorded the measurements and plotted them as shown in Figure S17.9. The control chart revealed that only with difficulty could the process average be kept in control. Occasional operator interventions were required. Also the process range was moving towards (and once breaking) the upper control limit. The process seemed to be becoming more variable. After investigation it was discovered that, because of faulty maintenance of the line, skin cream was occasionally contaminat- ing the torque head (the part of the line which fitted the cap). This resulted in erratic tightening of the caps.

Figure S17.9 The completed control form for GAM’s torque machine showing the mean (X̄ ) and range (R̄) charts

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Part Three Planning and control532

Figure S17.10 In addition to points falling outside the control limits, other unlikely sequences of points should be investigated

Interpreting control charts

Plots on a control chart which fall outside control limits are an obvious reason for believing that the process might be out of control, and therefore for investigating the process. This is not the only clue which could be revealed by a control chart, however. Figure S17.10 shows some other patterns which could be interpreted as behaviour sufficiently unusual to warrant investigation.

Process control, learning and knowledge

In recent years the role of process control and SPC in particular, has changed. Increasingly, it is seen not just as a convenient method of keeping processes in control, but also as an activity which is fundamental to the acquisition of competitive advantage. This is a remarkable shift in the status of SPC. Traditionally it was seen as one of the most operational, immediate and ‘hands-on’ operations management techniques. Yet it is now being connected with an operation’s strategic capabilities. This is how the logic of the argument goes:

1 SPC is based on the idea that process variability indicates whether a process is in control or not.

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Supplement to Chapter 17 Statistical process control (SPC) 533

2 Processes are brought into control and improved by progressively reducing process vari- ability. This involves eliminating the assignable causes of variation.

3 One cannot eliminate assignable causes of variation without gaining a better understand- ing of how the process operates. This involves learning about the process, where its nature is revealed at an increasingly detailed level.

4 This learning means that process knowledge is enhanced, which in turn means that operations managers are able to predict how the process will perform under different circumstances. It also means that the process has a greater capability to carry out its tasks at a higher level of performance.

5 This increased process capability is particularly difficult for competitors to copy. It cannot be bought ‘off-the-shelf ’. It only comes from time and effort being invested in controlling operations processes. Therefore, process capability leads to strategic advantage.

In this way, process control leads to learning which enhances process knowledge and builds difficult-to-imitate process capability.

Acceptance sampling

Process control is usually the preferred method of controlling quality because quality is being ‘built in’ to the process rather than being inspected afterwards. However, sometimes it may be necessary to inspect batches of products or services either before or after a process. The purpose of acceptance sampling is to decide whether, on the basis of a sample, to accept or reject the whole batch, for example a batch of component parts from a supplier. Acceptance sampling uses the proportion of wrongs to rights (or defectives to acceptable) and is usually carried out on attributes rather than variables. Again, it is important to understand the risks inherent in using a sample to make a judgement about a far larger batch. Table S17.3 illustrates the risks of acceptance sampling in the form of type I and type II errors. In accept- ance sampling the type I risk is often referred to as the producer’s risk because it is the risk that the operation rejects a batch that is actually of good quality. The type II risk is usually called the consumer’s risk because it is the risk of accepting a batch that is actually poor and sending it to the consumer of the product or service.

Table S17.3 The risks inherent in acceptance sampling

Decision The batch actually is

OK Not OK

Reject batch Type I error Correct decision

Accept batch Correct decision Type II error

Sampling plans

Acceptance sampling involves a sample being taken from a batch and a decision to accept or reject the batch being made by comparing the number of ‘defects’ found in the sample to a predetermined acceptable number. The sampling plan which describes this procedure is defined by two factors, n and c, where:

n = the sample size c = the acceptance number of defects in the sample.

Process knowledge

Sampling plan

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

If x = number of defects actually found in the sample, a decision is made based on the fol- lowing simple decision rule:

If x ≤ c then accept the whole batch. If x > c then reject the whole batch.

A set of tables called the Dodge–Romig Sampling Inspection Tables provides values for n and c for a given set of risks (using these tables is beyond the scope of this book). The ability of this plan to discriminate between good batches and bad ones is based upon the binomial distribution and is described by an operating characteristic (OC) curve. The OC curve for a sampling plan shows the probability of accepting a batch as the actual percentage of defects varies. An ideal OC curve would look like the blue line in Figure S17.11.

In this example the level of defects which is regarded as acceptable is 0.4 per cent and the sampling plan is perfect at discriminating between acceptable and unacceptable batches. The probability of accepting a batch whose actual level of defects is less than 0.4 per cent is 100 per cent and there is no chance of ever accepting a batch whose actual level of defects is more than 0.4 per cent. However, in practice, no procedure based on sampling, and there- fore carrying risk, could ever deliver such an ideal curve. Only 100 per cent inspection using a perfect inspector could do so. Any use of sampling will have to accept the existence of type I and type II errors. Figure S17.11 also shows the blue line which represents a sampling plan for sampling 250 items (n = 250) and rejecting the batch if there is more than one defect (c = 1) in the sample. A batch is acceptable if it contains 0.4 per cent or fewer defects (1/250 = 0.04 per cent).

What is not known is the actual percentage of defective items in any one batch, and because the procedure relies on a sample, there will always be a probability of rejecting a good batch because the number of defects in the sample is two or more despite the batch in fact being acceptable (type I risk shown by the top shaded area). There is also a probability that in spite of accepting a batch (because the number of defects it contains is zero or one) the actual number of defects in the whole batch might be greater than 0.04 per cent (type II risk

Part Three Planning and control534

Figure S17.11 Ideal and real operating characteristics showing acceptable quality level (AQL), lot tolerance percentage defective (LTPD), producer’s risk and consumer’s risk

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

shown in the lower blue shaded area of Fig. S17.11). If the sizes of these risks are felt to be too great, the sample size can be increased, which will move the shape of the curve towards the ideal. However, this implies increased time and cost in inspecting the batch. To create an appropriate sampling plan four factors need to be specified (see Figure S17.11): type I error, type II error, acceptable quality level (AQL) and lot tolerance percentage defective (LTPD):

● Type I error. The usual value used for producer’s risk (type I error) is often set with a prob- ability of 0.05. This means that management is willing to take a 5 per cent chance that a batch of good quality will be rejected when it is actually acceptable.

● Type II error. The value for the consumer’s risk (type II error) is often set with a probability of 0.1. This means that management is willing to risk at most a 10 per cent chance that a poor-quality batch will be accepted.

● AQL. The acceptable quality level is the actual percentage of defects in a batch which the organization is willing to reject mistakenly (by chance) 5 per cent of the time (assuming a 0.05 type I error) when the batch is actually acceptable.

● LTPD. The lot tolerance percentage defective is the actual percentage of defects in a batch that management is willing to accept mistakenly 10 per cent of the time (assuming a 0.1 type II error).

Supplement to Chapter 17 Statistical process control (SPC) 535

A frequently made criticism of acceptance sampling is that it assumes that some amount of defects and failure is acceptable to the organization or its customers. By accepting the inevitability of failure and poor quality, it is argued, the operation will become ‘lazy’ at trying to eliminate the causes of bad quality. Rather than see quality as primarily some- thing to be improved, acceptance sampling views it as being almost ‘predetermined’ by the characteristics of the process. The main task is to measure output and understand the risks involved, not to get to the root causes of poor quality. More recent approaches to quality management (such as TQM) suggest that ‘right first time every time’ is the only acceptable approach and that organizations should strive to produce zero defective items rather than some ‘acceptable quality level’.

Critical commentary

Summary

■ Statistical process control (SPC) involves using control charts to track the performance of one or more quality characteristics in the operation. The power of control charting lies in its ability to set control limits derived from the statistics of the natural variation of processes. These control limits are often set at ± 3 standard deviations of the natural variation of the process samples.

■ Control charts can be used for either attributes or variables. An attribute is a quality charac- teristic which has two states (for example, right or wrong). A variable is one which can be measured on a continuously variable scale.

■ Process control charts allow operations managers to distinguish between the ‘normal’ vari- ation inherent in any process and the variations which could be caused by the process going out of control.

■ Acceptance sampling helps managers to understand the risks they are taking when they make decisions about a whole batch of products on the basis of a sample taken from that batch. The risks of any particular sampling plan are shown on its operating characteristic (OC) curve. However, some of its assumptions make acceptance sampling controversial.

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

Part Three Planning and control536

Woodall, W.H. (2000) Controversies and contradictions in statistical process control. Paper presented at the Journal of Quality Technology Session at the 44th Annual Fall Technical Conference of the Chemical and Process Industries Division and Statistics Division of the American Society for Quality and the Section on Physical & Engineering Sciences of the American Statistical Association in Minneapolis, Minnesota, October 12–13, 2000. Academic but interesting.

Selected further reading

www.asq.org/ The American Society for Quality site. Good professional insights.

Useful web sites

IS B

N : 0 -5

5 8 -8

7 7 4 7 -8

boom:

title: Operations Management 6e Ch17
subtitle:
abstract:
text:

instructions: This is a Protected PDF document. Please enter your user name and password to unlock the text.

form:

username_label: User Name:
response:

username:
docid: 1000-0001-6E2E-000DD271

pwd_label: Password:
neversend:

pwd:

unlocked:

rememberme: Yes
rememberme_label: Remember my user name and password.

help: If you are experiencing problems unlocking this document or you have questions regarding Protectedpdf files please contact a Technical Support representative: In the United States: 1-877-832-4867 In Canada: 1-800-859-3682 Outside the U.S. and Canada: 1-602-387-2222 Email: [email protected].
protectedpdf_version: 2012-4-13 (Version 2.6.0 Build 2.0.7.36)
version: 2012-09-09
logo:
info_publisher: This document is published by: Apollo Group, Inc.
info: