Nigel S., Chambers S., Johnson R. Operations Management
Подождите немного. Документ загружается.
produced by the process, in terms of defects per million. The defects per million measure is
used within the Six Sigma approach to emphasize the drive towards a virtually zero defect
objective. Now the definition of Six Sigma has widened to well beyond this rather narrow
statistical perspective. General Electric (GE), who were probably the best known of the early
adopters of Six Sigma, defined it as, ‘A disciplined methodology of defining, measuring,
analysing, improving, and controlling the quality in every one of the company’s products,
processes, and transactions – with the ultimate goal of virtually eliminating all defects’.
So, now Six Sigma should be seen as a broad improvement concept rather than a simple
examination of process variation, even though this is still an important part of process
control, learning and improvement.
Measuring performance
The Six Sigma approach uses a number of related measures to assess the performance of
operations processes.
● A defect is a failure to meet customer-required performance (defining performance
measures from a customer’s perspective is an important part of the Six Sigma approach).
● A defect unit or item is any unit of output that contains a defect (i.e. only units of
output with no defects are not defective, defective units will have one or more than one
defect).
● A defect opportunity is the number of different ways a unit of output can fail to meet
customer requirements (simple products or services will have few defect opportunities,
but very complex products or services may have hundreds of different ways of being
defective).
● Proportion defective is the percentage or fraction of units that have one or more defect.
● Process yield is the percentage or fraction of total units produced by a process that are
defect-free (i.e. 1 − proportion defective).
● Defect per unit (DPU) is the average number of defects on a unit of output (the number
of defects divided by the number of items produced).
● Defects per opportunity is the proportion or percentage of defects divided by the total
number of defect opportunities (the number of defects divided by (the number items
produced × the number of opportunities per item)).
● Defects per million opportunities (DPMO) is exactly what it says, the number of defects
which the process will produce if there were one million opportunities to do so.
● The Sigma measurement is derived from the DPMO and is the number of standard
deviations of the process variability that will fit within the customer specification limits.
Part Four Improvement
554
Defects per million
Zero defect
An insurance process checks details of insurance claims and arranges for customers
to be paid. It samples 300 claims at random at the end of the process. They find that
51 claims had one or more defects and there were 74 defects in total. Four types of
error were observed, coding errors, policy conditions errors, liability errors and noti-
fication errors.
Proportion defective =
==0.17 (17% defective)
Yield = 1 − Proportion of defectives
= 1 − 0.17 = 0.83 or (83% yield)
51
300
Number of defects
Number of units processed
Worked example
M18_SLAC0460_06_SE_C18.QXD 10/20/09 15:22 Page 554
Although the scope of Six Sigma is disputed, among elements frequently associated with
Six Sigma include the following:
● Customer-driven objectives – Six Sigma is sometimes defined as ‘the process of comparing
process outputs against customer requirements’. It uses a number of measures to assess
the performance of operations processes. In particular it expresses performance in terms
of defects per million opportunities (DPMO). This is exactly what it says, the number of
defects which the process will produce if there were one million opportunities to do so.
This is then related to the ‘Sigma measurement’ of a process and is the number of standard
deviations of the process variability that will fit within the customer specification limits.
● Use of evidence – Although Six Sigma is not the first of the new approaches to operations
to use statistical methods it has done a lot to emphasize the use of quantitative evidence.
In fact much of the considerable training required by Six Sigma consultants is devoted to
mastering quantitative analytical techniques.
● Structured improvement cycle – The structured improvement cycle used in Six Sigma is the
DMAIC cycle.
● Process capability and control – Not surprisingly, given its origins, process capability and
control is important within the Six Sigma approach.
● Process design – Latterly Six Sigma proponents also include process design into the collec-
tion of elements that define the Six Sigma approach.
● Structured training and organization of improvement – The Six Sigma approach holds that
improvement initiatives can only be successful if significant resources and training are
devoted to their management. It recommends a specially trained cadre of practitioners
and internal consultants named after ‘martial arts’ grades, see below.
The ‘martial arts’ analogy
The terms that have become associated with Six Sigma experts (and denote their level of
expertise) are, Master Black Belt, Black Belt and Green Belt. Master Black Belts are experts
in the use of Six Sigma tools and techniques as well as how such techniques can be used and
implemented. Primarily Master Black Belts are seen as teachers who can not only guide
improvement projects, but also coach and mentor Black Belts and Green Belts who are closer
to the day-to-day improvement activity. They are expected to have the quantitative analytical
skills to help with Six Sigma techniques and also the organizational and interpersonal skills
to teach and mentor. Given their responsibilities, it is expected that Master Black Belts are
employed full-time on their improvement activities. Black Belts can take a direct hand in
organizing improvement teams. Like Master Black Belts, Black Belts are expected to develop
their quantitative analytical skills and also act as coaches for Green Belt. Black Belts are
dedicated full-time to improvement, and although opinions vary on how many Black Belts
should be employed in an operation, some organizations recommend one Black Belt for
Chapter 18 Operations improvement
555
Defects per unit =
==0.247 (or 24.7) DPU
Defects per opportunity =
==0.062 DPO
Defects per million opportunities = DPO × 10
6
= 62,000 DPMO
74
300 × 4
Number of defects
Number of units produced × Number of opportunities
74
300
Number of defects
Number of units processed
M18_SLAC0460_06_SE_C18.QXD 10/20/09 15:22 Page 555
every hundred employees. Green Belts work within improvement teams, possibly as team
leaders. They have significant amounts of training, although less than Black Belts. Green Belts
are not full-time positions; they have normal day-to-day process responsibilities but are
expected to spend at least twenty per cent of their time on improvement projects.
Part Four Improvement
556
‘I think Six Sigma is powerful because of its definition;
it is the process of comparing process outputs against
customer requirements. Processes operating at less than
3.4 defects per million opportunities means that you must
strive to get closer to perfection and it is the customer
that defines the goal. Measuring defects per opportunity
means that you can actually compare the process of, say,
a human resources process with a billing and collection
process.’ Paul Ruggier head of Process at Xchanging is a
powerful advocate of Six Sigma, and credits the success
of the company, at least partly, to the approach.
Xchanging, created in 1998, is one of a new breed
of companies, operating as an outsourcing business for
‘back-office’ functions for a range of companies, such
as Lloyds of London, the insurance centre. Xchanging’s
business proposition is for the client company to transfer
the running of the whole or part of their back office to
Xchanging, either for a fixed price or one determined by
cost savings achieved. The challenge Xchanging faces
is to run that back office in a more effective and efficient
manner than the client company had managed in the
past. So, the more effective Xchanging is at running
the processes, the greater its profit. To achieve these
efficiencies Xchanging offers larger scale, a higher level
of process expertise, focus and investment in technology.
But above all, they offer, a Six Sigma approach.
‘Everything we do can be broken down into a process’,
says Paul Ruggier. ‘It may be more straightforward in a
manufacturing business, frankly they’ve been using a lot
of Six Sigma tools and techniques for decades. But the
concept of process improvement is relatively new in many
service companies. Yet the concept is powerful. Through
the implementation of this approach we have achieved
30% productivity improvements in 6 months.’
The company also adopts the Six Sigma terminology
for its improvement practitioners – Master Black Belts,
Black Belts and Green Belts. Attaining the status of Black
Belt is very much sought after as well as being fulfilling,
says Rebecca Whittaker who is a Master Black Belt at
Xchanging. ‘At the end of a project it is about having a
process which is redesigned to such an extent, that is
simplified and consolidated and people come back and say,
“It’s so much better than it used to be”. It makes their lives
better and it makes the business results better and those
are the things that make being a Black Belt worthwhile.’
Rebecca was recruited by Xchanging along with a
number of other Master Black Belts as part of a strategic
Short case
Six Sigma at Xchanging
12
decision to kick-start Six Sigma in the company. It is seen
as a particularly responsible position by the company and
Master Black Belts are expected to be well versed in the
Six Sigma techniques and be able to provide the training
and knowhow to develop other staff within the company.
In Rebecca’s case she has been working as a Six Sigma
facilitator for five years, initially as a Green Belt, then as a
Black Belt.
Typically a person identified as having the right
analytical and interpersonal skills will be taken off their job
for at least a year, training and immersed in the concepts
of improvement and then sent to work with line staff as
project manager/facilitator. Their role as Black Belt will be
to guide the line staff to make improvements in the way
they do the job. One of the new Black Belts at Xchanging,
Sarah Frost, is keen to stress the responsibility she owes
to the people who will have to work in the improvement
process. ‘Being a Black Belt is about being a project
manager. It is about working with the staff and combining
our skills in facilitation and our knowledge of the Six
Sigma process with their knowledge of the business. You
always have to remember that you will go onto another
project but they [process staff] will have to live with the
new process. It is about building solutions that they can
believe in.’
Source: Rex Features
M18_SLAC0460_06_SE_C18.QXD 10/20/09 15:22 Page 556
Differences and similarities
In this chapter we have chosen to very briefly explain four improvement approaches. It
could have been more. Enterprise resource planning (ERP, see Chapter 14), total preventive
maintenance (TPM, see Chapter 19), lean Sigma (a combination of lean and Six Sigma),
and others could have been added. But these four constitute a representative sample of
the most commonly used approaches. Nor do we have the space to describe them fully. Each
of the approaches is the subject of several books that describe them in great detail. There
is no shortage of advice from consultants and academics as to how they should be used.
And there are clearly some common elements between some of these approaches that we
have described. Yet there are also differences between them in that each approach includes
a different set of elements and therefore a different emphasis and these differences need
to be understood. For example, one important difference relates to whether the approaches
emphasize a gradual, continuous approach to change, or whether they recommend a more
radical ‘breakthrough’ change. Another difference concerns the aim of the approach. What
is the balance between whether the approach emphasizes what changes should be made or
how changes should be made? Some approaches have a firm view of what is the best way
to organize the operation’s processes and resources. Other approaches hold no particular
view on what an operation should do but rather concentrate on how the management of
an operation should decide what to do. Indeed we can position each of the elements and the
approaches that include them. This is illustrated in Figure 18.7. The approaches differ in the
extent that they prescribe appropriate operations practice. BPR for example is very clear in
what it is recommending. Namely, that all processes should be organized on an end-to-end
basis. Its focus is what should happen rather than how it should happen. To a slightly lesser
extent lean is the same. It has a definite list of things that processes should or should not be
– waste should be eliminated, inventory should be reduced, technology should be flexible,
and so on. Contrast this with both Six Sigma and TQM which focus to a far greater extent
Chapter 18 Operations improvement
557
One common criticism of Six Sigma is that it does not offer anything that was not available
before. Its emphasis on improvement cycles comes from TQM, its emphasis on reducing
variability comes from statistical process control, its use of experimentation and data
analysis is simply good quantitative analysis. The only contribution that Six Sigma has
made, argue its critics, is using the rather gimmicky martial arts analogy of Black Belt etc.
to indicate a level of expertise in Six Sigma methods. All Six Sigma has done is package
pre-existing elements together in order for consultants to be able to sell them to gullible
chief executives. In fact it’s difficult to deny some of these points. Maybe the real issue is
whether it is really a criticism. If bringing these elements together really does form an
effective problem-solving approach, why is this is a problem? Six Sigma is also accused
of being too hierarchical in the way it structures its various levels of involvement in the
improvement activity (as well as the dubious use of martial-arts-derived names such as
Black Belt). It is also expensive. Devoting such large amount of training and time to
improvement is a significant investment, especially for small companies. Nevertheless,
Six Sigma proponents argue that the improvement activity is generally neglected in most
operations and if it is to be taken seriously, it deserves the significant investment implied
by the Six Sigma approach. Furthermore, they argue, if operated well, Six Sigma improve-
ment projects run by experienced practitioners can save far more than their cost. There
are also technical criticisms of Six Sigma. Most notably that in purely statistical terms the
normal distribution which is used extensively in Six Sigma analysis does not actually
represent most process behaviour. Other technical criticisms (that are not really the subject
of this book) imply that aiming for the very low levels of defects per million opportunities,
as recommended by Six Sigma proponents, is far too onerous.
Critical commentary
M18_SLAC0460_06_SE_C18.QXD 10/20/09 15:22 Page 557
on how operations should be improved. Six Sigma in particular has relatively little to say
about what is good or bad in the way operations resources are organized (with the possible
exception of its emphasizing the negative effects of process variation). Its concern is largely
the way improvements should be made: using evidence, using quantitative analysis, using the
DMAIC cycle, and so on. They also differ in terms of whether they emphasize gradual or rapid
change. BPR is explicit in its radical nature. By contrast TQM and lean both incorporate
ideas of continuous improvement. Six Sigma is relatively neutral on this issue and can be
used for small or very large changes.
Improvement techniques
All the techniques described in this book and its supplements can be regarded as ‘improvement’
techniques. However, some techniques are particularly useful for improving operations and
processes generally. Here we select some techniques which either have not been described
elsewhere or need to be reintroduced in their role of helping operations improvement
particularly.
Scatter diagrams
Scatter diagrams provide a quick and simple method of identifying whether there is evidence
of a connection between two sets of data: for example, the time at which you set off for work
every morning and how long the journey to work takes. Plotting each journey on a graph
which has departure time on one axis and journey time on the other could give an indication
of whether departure time and journey time are related, and if so, how. Scatter diagrams
can be treated in a far more sophisticated manner by quantifying how strong is the relation-
ship between the sets of data. But, however sophisticated the approach, this type of graph
Scatter diagrams
Part Four Improvement
558
Figure 18.7 The four approaches on the two dimensions of improvement
M18_SLAC0460_06_SE_C18.QXD 10/20/09 15:22 Page 558
only identifies the existence of a relationship, not necessarily the existence of a cause–effect
relationship. If the scatter diagram shows a very strong connection between the sets of data,
it is important evidence of a cause–effect relationship, but not proof positive. It could be
coincidence!
Process maps (flow charts)
Process maps (sometimes called ‘flow charts’ in this context) can be used to give a detailed
understanding prior to improvement. They were described in Chapter 4 and are widely used
in improvement activities. The act of recording each stage in the process quickly shows up
poorly organized flows. Process maps can also clarify improvement opportunities and shed
further light on the internal mechanics or workings of an operation. Finally, and probably
most importantly, they highlight problem areas where no procedure exists to cope with a
particular set of circumstances.
Process maps
Chapter 18 Operations improvement
559
Kaston Pyral Services Ltd (A)
Kaston Pyral Services Ltd (KPS) installs and maintains environmental control, heating
and air conditioning systems. It has set up an improvement team to suggest ways in
which it might improve its levels of customer service. The improvement team has
completed its first customer satisfaction survey. The survey asked customers to score
the service they received from KPS in several ways. For example, it asked customers to
score services on a scale of one to ten on promptness, friendliness, level of advice, etc.
Scores were then summed to give a ‘total satisfaction score’ for each customer – the
higher the score, the greater the satisfaction. The spread of satisfaction scores puzzled
the team and they considered what factors might be causing such differences in the way
their customers viewed them. Two factors were put forward to explain the differences.
(a) the number of times in the past year the customer had received a preventive main-
tenance visit;
(b) the number of times the customer had called for emergency service.
All these data were collected and plotted on scatter diagrams as shown in Figure 18.8(a).
It shows that there seems to be a clear relationship between a customer’s satisfaction
score and the number of times the customer was visited for regular servicing. The
scatter diagram in Figure 18.8(b) is less clear. Although all customers who had very
high satisfaction scores had made very few emergency calls, so had some customers
with low satisfaction scores. As a result of this analysis, the team decided to survey
customers’ views on its emergency service.
Worked example
Figure 18.8 Scatter diagrams for customer satisfaction versus (a) number of preventive
maintenance calls and (b) number of emergency service calls
M18_SLAC0460_06_SE_C18.QXD 10/20/09 15:22 Page 559
Part Four Improvement
560
Kaston Pyral Services Ltd (B)
As part of its improvement programme the team at KPS is concerned that customers are
not being served well when they phone in with minor queries over the operation of their
heating systems. These queries are not usually concerned with serious problems, but often
concern minor irritations which can be equally damaging to the customers’ perception of
KPS’s service. Figure 18.9 shows the process map for this type of customer query. The
team found the map illuminating. The procedure had never been formally laid out in
this way before, and it showed up three areas where information was not being recorded.
These are the three points marked with question marks on the process map in Figure 18.9.
As a result of this investigation, it was decided to log all customer queries so that analysis
could reveal further information on the nature of customer problems.
Worked example
Figure 18.9 Process map for customer query
M18_SLAC0460_06_SE_C18.QXD 10/20/09 15:22 Page 560
Kaston Pyral Services Ltd (C)
The improvement team at KPS was working on a particular area which was proving a
problem. Whenever service engineers were called out to perform emergency servicing
for a customer, they took with them the spares and equipment which they thought
would be necessary to repair the system. Although engineers could never be sure exactly
what materials and equipment they would need for a job, they could guess what was
likely to be needed and take a range of spares and equipment which would cover most
eventualities. Too often, however, the engineers would find that they needed a spare that
they had not brought with them. The cause–effect diagram for this particular problem,
as drawn by the team, is shown in Figure 18.10.
Worked example
Pareto diagrams
In any improvement process, it is worthwhile distinguishing what is important and what
is less so. The purpose of the Pareto diagram (which was first introduced in Chapter 12) is
to distinguish between the ‘vital few’ issues and the ‘trivial many’. It is a relatively straight-
forward technique which involves arranging items of information on the types of problem
or causes of problem into their order of importance (usually measured by ‘frequency of
occurrence’). This can be used to highlight areas where further decision-making will be
Cause–effect diagrams
Cause–effect diagrams
Cause–effect diagrams are a particularly effective method of helping to search for the root
causes of problems. They do this by asking what, when, where, how and why questions, but
also add some possible ‘answers’ in an explicit way. They can also be used to identify areas where
further data are needed. Cause–effect diagrams (which are also known as ‘Ishikawa diagrams’)
have become extensively used in improvement programmes. This is because they provide a
way of structuring group brainstorming sessions. Often the structure involves identifying
possible causes under the (rather old-fashioned) headings of: machinery, manpower, materials,
methods and money. Yet in practice, any categorization that comprehensively covers all
relevant possible causes could be used.
Chapter 18 Operations improvement
561
Figure 18.10 Cause–effect diagram of unscheduled returns at KPS
M18_SLAC0460_06_SE_C18.QXD 10/20/09 15:22 Page 561
useful. Pareto analysis is based on the phenomenon of relatively few causes explaining the
majority of effects. For example, most revenue for any company is likely to come from
relatively few of the company’s customers. Similarly, relatively few of a doctor’s patients will
probably occupy most of his or her time.
Pareto analysis
Part Four Improvement
562
Kaston Pyral Services Ltd (D)
The KPS improvement team which was investigating unscheduled returns from emer-
gency servicing (the issue which was described in the cause–effect diagram in Fig. 18.11)
examined all occasions over the previous 12 months on which an unscheduled return
had been made. They categorized the reasons for unscheduled returns as follows:
1 The wrong part had been taken to a job because, although the information which the
engineer received was sound, he or she had incorrectly predicted the nature of the
fault.
2 The wrong part had been taken to the job because there was insufficient information
given when the call was taken.
3 The wrong part had been taken to the job because the system had been modified in
some way not recorded on KPS’s records.
4 The wrong part had been taken to the job because the part had been incorrectly issued
to the engineer by stores.
5 No part had been taken because the relevant part was out of stock.
6 The wrong equipment had been taken for whatever reason.
7 Any other reason.
The relative frequency of occurrence of these causes is shown in Figure 18.11. About a
third of all unscheduled returns were due to the first category, and more than half the
returns were accounted for by the first and second categories together. It was decided
that the problem could best be tackled by concentrating on how to get more information
to the engineers which would enable them to predict the causes of failure accurately.
Worked example
Figure 18.11 Pareto diagram for causes of unscheduled returns
M18_SLAC0460_06_SE_C18.QXD 10/20/09 15:22 Page 562
Why–why analysis
Why–why analysis starts by stating the problem and asking why that problem has occurred.
Once the reasons for the problem occurring have been identified, each of the reasons is taken
in turn and again the question is asked why those reasons have occurred, and so on. This
procedure is continued until either a cause seems sufficiently self-contained to be addressed
by itself or no more answers to the question ‘Why?’ can be generated.
Why–why analysis
Chapter 18 Operations improvement
563
Kaston Pyral Services Ltd (E)
The major cause of unscheduled returns at KPS was the incorrect prediction of reasons
for the customer’s system failure. This is stated as the ‘problem’ in the why–why analysis
in Figure 18.12. The question is then asked, why was the failure wrongly predicted? Three
answers are proposed: first, that the engineers were not trained correctly; second, that
they had insufficient knowledge of the particular product installed in the customer’s
location; and third, that they had insufficient knowledge of the customer’s particular
system with its modifications. Each of these three reasons is taken in turn, and the ques-
tions are asked, why is there a lack of training, why is there a lack of product knowledge,
and why is there a lack of customer knowledge? And so on.
Worked example
Figure 18.12 Why–why analysis for ‘failure wrongly predicted’
M18_SLAC0460_06_SE_C18.QXD 10/20/09 15:22 Page 563