KEY QUESTIONS
6.1 What is process design?
6.2 What should be the objectives of process design?
6.3 How do volume and variety affect process design?
6.4 How are processes designed in detail?
INTRODUCTION
In Chapter 1 we described how all operations consist of a collection of processes that interconnect with
each other to form an internal network. Each process acts as a smaller version of the whole operation of
which they form a part, and transformed resources flow between them. We also defined a process as ‘an
arrangement of resources and activities that transform inputs into outputs that satisfy (internal or external)
customer needs’. They are the ‘building blocks’ of all operations, and as such they play a vital role in how
well operations operate. This is why process design is so important. Unless its individual processes are
well designed, an operation as a whole will not perform as well as it could. And operations managers are at
the forefront of how processes are designed. In fact, all operations managers are designers. When they
purchase or rearrange the position of a piece of equipment, or when they change the way of working within
a process, it is a design decision because it affects the physical shape and nature of their process, as well
as its performance. This chapter examines the design of processes. Figure 6.1 shows where this topic fits
within the overall model of operations management.
Figure 6.1 This chapter examines process design
Figure 6.1: Long Description
6.1 What is process design?
To ‘design’ is to conceive the looks, arrangement and workings of something before it
is created. In that sense, it is a conceptual exercise. Yet it is one that must deliver a
solution that will work in practice. Design is also an activity that can be approached at
different levels of detail. One may envisage the general shape and intention of
something before getting down to defining its details. This is certainly true for process
design . At the start of the process design activity it is important to understand the
design objectives, especially when the overall shape and nature of the process is being
decided. The most common way of doing this is by positioning it according to its
volume and variety characteristics. Eventually the details of the process must be
analysed to ensure that it fulfils its objectives effectively.
Process design and product/service design are interrelated
Often we will treat the design of services and products, on the one hand, and the design
of the processes that make them, on the other, as though they were separate activities.
Yet they are clearly interrelated. It would be foolish to commit to the detailed design of
any product or service without some consideration of how it is to be produced. Small
changes in the design of products and services can have profound implications for the
way the operation eventually has to produce them. Similarly, the design of a process can
constrain the freedom of product and service designers to operate as they would wish
(see Figure 6.2). This holds good for all operations. However, the overlap between the
two design activities is generally greater in operations that produce services. Because
many services involve the customer in being part of the transformation process, the
service, as far as the customer sees it, cannot be separated from the process to which
the customer is subjected. Certainly, when product designers also have to make or use
the things that they design, it can concentrate their minds on what is important. For
example, in the early days of flight, the engineers who designed the aircraft were also
the test pilots who took them out on their first flight.
Operations principle
The design of processes cannot be done independently of the services and/or products that are being
created.
Figure 6.2 The design of products/services and the design of processes are interrelated and should be
treated together
Figure 6.2: Long Description
Airports are complex operations – really complex. Their processes handle passengers, aircraft, crew,
baggage, commercial cargo, food, security, restaurants and numerous customer services. Their
operations managers must cope with aviation administration rules and regulations, a huge number of
airport service contracts, usually thousands of staff with a wide variety of specialisms, airlines with
competing claims to service priority, and customers, some of whom are experienced, others less so.
Their processes are also vulnerable to disruptions from late arrivals, aircraft malfunction, weather, the
industrial action of workers two continents away, conflicts and terrorism. Designing the processes
that can operate under these conditions must be one of the most challenging operations tasks. So, to
win prizes for ‘Best Airport’ customer service and operating efficiency year after year has to be
something of an achievement, which is what the sixth-busiest international airport, Changi airport in
Singapore, has done. As a major air hub in Asia, Changi serves more than 100 international airlines
flying to some 300 cities in about 70 countries and territories worldwide. It handles almost 60 million
passengers (that’s roughly 10 times the size of Singapore’s population). A flight takes off or lands at
Changi roughly once every 90 seconds.
When Changi opened its new Terminal 4, it increased the airport’s annual passenger handling
capacity to around 82 million. Every stage of the customers’ journey through the terminal was
designed to be as smooth as possible. The aim of all the processes within and around the terminal
was to provide fast, smooth and seamless flow for passengers. Each stage in the customer journey
was provided with enough capacity to cope with anticipated demand. Once passengers arrive at the
two-storey terminal building they pass through kiosks and automated options for self-check-in, self-
bag tagging and self-bag-drops. Their bags are transported to the aircraft via an advanced and
automated baggage handling system. Similarly, automated options, including face recognition
technology, are used at immigration counters and departure gates. Biometric technology and ‘fast
and seamless travel’ (FAST) services help to speed passenger throughput and increase efficiency.
2
After security checks, passengers find themselves in 15,000 m of shopping, dining and other retail
spaces. The feelings of passengers were an important part of the design of T4. Architecturally, it
aimed to be functional, and yet have its own aesthetic character, while ensuring that the design was
passenger-centric and user-friendly. And with so many different companies involved in the day-to-day
operation of the airport it was vital to include as many stakeholders as possible during the design.
Workshops were conducted with various stakeholders, including airlines, ground handlers,
immigration and security agencies, retail, and food and beverage operators as well as other users to
ensure that the T4 design met the needs of each party.
Process networks
In Chapter 1 we used the ‘hierarchy of operations’ to illustrate how any operation is
both made up of networks (of processes) and a part of networks (of other operations).
This idea is essential in making all networks, including process networks, operate
effectively. Figure 6.3 shows a simplified internal process network for one business. It
has many processes that transform items and transfer them to other internal processes.
Through this network there are many ‘process chains’, that is, threads of processes
within the network. And thinking about processes as part of a network has a number of
advantages. First, understanding how and where a process fits into the internal network
helps to establish appropriate objectives for the process. Second, one can check to
make sure that everyone in a process has a clear ‘line of sight’ forward through to end
customers, so that the people working in each process have a better chance of seeing
how they contribute to satisfying the operation’s customers. Even more important, one
can ask the question, ‘how can each process help the intermediate processes that lie
between them and the customer, to operate effectively?’ Third, a clear ‘line of sight’
backwards through to the operation’s suppliers makes the role and importance of
suppliers easier to understand.
Figure 6.3 A process network within an operation showing an internal ‘process chain’
Figure 6.3: Long Description
6.2 What should be the objectives of process design?
The whole point of process design is to make sure that the performance of the process is
appropriate for whatever it is trying to achieve. For example, if an operation competes
primarily on its ability to respond quickly to customer requests, many of its processes
will need to be designed to give fast throughput times. Similarly, if an operation
competes on low price, cost-related objectives are likely to dominate its process
design. In other words, some kind of logic should link what the operation as a whole is
attempting to achieve, and the performance objectives of its individual processes. As
when we examined product and service design innovation in Chapter 4, we will include
‘sustainability’ as an operational objective of process design, even though it is really a
far broader societal issue that is part of the organisation’s ‘triple bottom line’ (see
Chapter 2). This is illustrated in Table 6.1.
Operations principle
The design of any process should be judged on its quality, speed, dependability, flexibility, cost and
sustainability performance.
Table 6.1 The impact of strategic performance objectives on process design objectives and performance
‘Micro’ process objectives
Because processes are managed at a very operational level, process design also needs
to consider a more ‘micro’and detailed set of objectives. These are largely concerned
with flow through the process. When whatever is being ‘processed’ enters a process, it
will progress through a series of activities where it is ‘transformed’ in some way.
Between these activities it may dwell for some time in inventories, waiting to be
transformed by the next activity. This means that the time that a unit spends in the
process (its throughput time) will be longer than the sum of all the transforming
activities that it passes through. Also, the resources that perform the process’s activities
may not be used all the time because not all items will necessarily require the same
activities and the capacity of each resource may not match the demand placed upon it.
So, neither the items moving through the process, nor the resources performing the
activities may be fully utilised. Because of this, the way that items leave the process is
unlikely to be exactly the same as the way they arrive at the process. It is common for
more ‘micro’ performance flow objectives to be used that describe process flow
performance. For example:
▶ Throughput rate (or flow rate) is the rate at which items emerge from the process,
i.e. the number of items passing through the process per unit of time.
▶ Cycle time is the reciprocal of throughput rate: it is the time between items
emerging from the process. The term ‘takt time ’ is the same, but is normally
applied to ‘paced’ processes like moving-belt assembly lines. It is the ‘beat’ or
2
tempo of working required to meet demand.
▶ Throughput time is the average elapsed time taken for inputs to move through the
process and become outputs.
▶ ‘Work-in-progress’ or process inventory is the number of items in the process, as an
average over a period of time.
▶ The utilisation of process resources is the proportion of available time that the
resources within the process are performing useful work.
Operations principle
Process flow objectives should include throughput rate, throughput time, work-in-progress and resource
utilisation, all of which are interrelated.
Some claim the first drive-through (or drive-thru, if you prefer) was the In-N-Out in California. Other
claimants include the Pig Stand restaurant in Los Angeles, that allowed customers to drive by the
back door where the chef would come out and deliver the restaurant’s famous ‘Barbequed Pig’
sandwiches. What became apparent though, as the drive-through idea began to spread (and included
other services such as banks), was that their design could have a huge impact on their efficiency and
profitability. Today, drive-through processes are slicker, and far, far, faster, although most stick to a
proven formula with orders generally placed by the customer using a microphone and picked up at a
window. It is a system that allows drive-throughs to provide fast and dependable service. In fact, there
is strong competition between drive-throughs to design the fastest and most reliable process. For
example, some Starbucks drive-throughs have strategically placed cameras at the order boards so
that servers can recognise regular customers and start making their order – even before it’s placed.
Other drive-throughs have experimented with simpler menu boards and see-through food bags to
ensure greater accuracy. There is no point in being fast if you don’t deliver what the customer
ordered. These details matter. It has been estimated that sales increase 1 per cent for every six
seconds saved at a drive-through. One experiment in making drive-through process times slicker
was carried out by a group of McDonald’s restaurants. On California’s central coast, 150 miles from
Los Angeles, a call centre took orders remotely from 40 McDonald’s outlets around the country. The
orders were then sent back to the restaurants and the food was assembled only a few metres from
where the order was placed. Although saving only a few seconds on each order, it could add up to
extra sales at busy times of the day. Another innovation is express lines for customers who place
digital orders ahead of time. A good drive-through process should also help customers to contribute
to speeding things up. So, for example, menu items must be easy to read and understand.
This is why what are often called ‘combo meals’ (burger, fries and a cola) can save time at the
ordering stage. By contrast, complex individual items or meals that require customisation can slow
down the process. This can become an issue for drive-through operators when fashion moves
towards customised salads and sandwiches. Yet there are signs that above a certain speed of
service, other aspects of process performance become more important. As one drive-through chief
operations manager points out, ‘you can get really fast but ruin the overall experience, because now
you’re not friendly’.
Standardisation of processes
One of the most important process design objectives, especially in large organisations,
concerns the extent to which process designs should be standardised. By standardisation
in this context we mean ‘doing things in the same way’, or ‘adopting a common
sequence of activities, methods and use of equipment’. It is a significant issue in large
organisations because, very often, different ways of carrying out similar or identical
tasks emerge over time in the various parts of the organisation. But, why not allow many
different ways of doing the same thing? That would give a degree of autonomy and
freedom for individuals and teams to exercise their discretion. The problem is that
allowing numerous ways of doing things causes confusion, misunderstandings and,
eventually, inefficiency. In healthcare processes, it can even cause preventable deaths.
For example, the Royal College of Physicians in the United Kingdom revealed that more
than 100 different charts were used for monitoring patients’ vital signs in UK
4
hospitals. Clinicians have to learn how to read new ones whenever they move, leading
to confusion. Potentially, many hospital deaths could be prevented if clinicians used a
standardised bed chart and process. The practical dilemma for most organisations is
how to draw the line between processes that are required to be standardised, and those
that are allowed to be different.
Operations principle
Standardising processes can give some significant advantages, but not every process can be
standardised.
Legal & General (L&G) is not the type of company one would expect to be building houses. One of the
United Kingdom’s leading financial services groups, which has invested over £19 billion in projects
including homebuilding, urban regeneration and clean energy, it became involved in building modular
homes. Modular construction of housing is more like the way you would expect a vehicle to be made.
Modules are made ‘off-site’ in a factory then transported to the building site. As some modular
construction proponents pointed out, once all cars were hand-built, but now cars are assembled in a
factory. Rosie Toogood, the Chief Executive, Modular Homes, and champion of the modern
manufacturing and construction approach, came from aerospace company Rolls-Royce, and Stuart
Lord, the Manufacturing Operations Director, spent his prior career in the automotive industry.
Starting a modular homes business from scratch required the considerable investment that a
major financial group like L&G could provide. By retaining quality construction but using modern
processes L&G called its approach ‘everything new, but nothing new’. What this meant was an
assembly line that had four giant computer-operated cutting and milling machines, and four smaller
ones, all of which were capable of cutting timber panels to higher levels of precision than could
normally be achieved on a conventional building site. The finished modules, fitted with wiring and
plumbing, decorated, carpeted and fitted out with kitchens and bathrooms, can then be loaded on to a
lorry and delivered to sites. A home can consist of a single module or several combined. This
approach meant digitally modelling every millimetre of every home before production started,
standardising and simplifying processes for efficient, high-quality production. Just as important, it
involved embracing continuous improvement – an approach that almost halved the time to deliver a
completed house. By standardising and simplifying processes, it could drive up quality and
productivity – driving down costs. Also, the adoption of mass processing was influenced by
increasingly tough energy targets. The process used less water than conventional building methods
and reduced site-generated waste. This, in turn, required fewer skips and provided a tidier, safer site.
A report by the Ellen MacArthur Foundation highlighted several potential environmental benefits from
off-site construction, including more energy-efficient homes.
Environmentally sensitive process design
With the issues of environmental protection becoming more important, process
designers have to take account of ‘green’ (sustainability) issues. In many countries in the
Global North, legislation has already provided some basic standards. Interest has
focused on some fundamental issues:
▶ The sources of inputs to a product or service. (Will they damage rainforests? Will
they use up scarce minerals? Will they exploit the poor or use child labour?)
▶ Quantities and sources of energy consumed in the process. (Do plastic beverage
bottles use more energy than glass ones? Should waste heat be recovered and used
in fish farming?)
▶ The amounts and type of waste material that are created in the manufacturing
processes. (Can this waste be recycled efficiently, or must it be burnt or buried in
landfill sites?)
▶ The life of the product itself. (If a product has a long useful life, will it consume
fewer resources than a product with a shorter life?)
▶ The end-of-life of the product. (Will the redundant product be difficult to dispose of
in an environmentally friendly way?)
Designers are faced with complex trade-offs between these factors, although it is
not always easy to obtain all the information that is needed to make the ‘best’ choices.
To help make more rational decisions in the design activity, some industries are
experimenting with life cycle analysis . This technique analyses all the production
inputs, the life cycle use of the product and its final disposal, in terms of total energy
used and all emitted wastes. The inputs and wastes are evaluated at every stage of a
service or product’s creation, beginning with the extraction or farming of the basic raw
materials. The ‘Operations in practice’ example ‘Ecover’s ethical operation design’
demonstrates that it is possible to include ecological considerations in all aspects of
product and process design.
Operations principle
The design of any process should include consideration of ethical and environmental issues.
Ecover cleaning products, such as washing liquid, are famously ecological. In fact, it is the
company’s whole rationale. ‘We clean with care’, say Ecover, ‘whether you’re washing your sheets,
your floors, your hands or your dishes, our products don’t contain those man-made chemicals that
can irritate your skin’. But it isn’t just the company’s products that are based on an ecologically
sustainable foundation. Ecover’s ecological factories in France and Belgium also embody the
company’s commitment to sustainability. Whether it is the factory roof, the use of energy or the way it
treats the water used in the production processes, Ecover points out that it does its best to limit
environmental impact. For example, the Ecover factory operates entirely on green electricity – the
type produced by wind generators, tidal generators and other natural sources. What is more, it
makes the most of the energy it does use by choosing energy-efficient lighting, and then using it only
when needed. And, although the machinery it uses in the factories is standard for the industry, it
keeps its energy and water consumption down by choosing low-speed appliances that can multi-task
and don’t require water to clean them. For example, the motors on its mixing machines can mix 25
tonnes of Ecover liquid while ‘consuming no more electricity than a few flat irons’. And it has a
‘squeezy gadget that’s so efficient at getting every last drop of product out of the pipes, they don’t
need to be rinsed through’. Ecover say that they ‘hate waste, so we’re big on recycling. We keep the
amount of packaging used in our products to a minimum, and make sure whatever cardboard or
plastic we do use can be recycled, reused or re-filled. It’s an ongoing process of improvement; in fact,
we’ve recently developed a new kind of green plastic we like to call “Plant-astic” that’s 100%
renewable, reusable and recyclable – and made from sugarcane’.
Even the building is ecological. It is cleverly designed to follow the movement of the sun from east
to west, so that production takes place with the maximum amount of natural daylight (good for saving
power and good for working conditions). The factory’s frame is built from pine rather than more
precious timbers and the walls are constructed using bricks that are made from clay, wood pulp and
mineral waste. They require less energy to bake, yet they’re light, porous and insulate well. The
factories’ roofs are covered in thick, spongy Sedum (a flowering plant, often used for natural roofing)
that gives insulation all year round. In fact, it’s so effective, that they don’t need heating or air
conditioning – the temperature never drops too low or rises too high.
6.3 How do volume and variety affect process design?
In Chapter 1 we saw how processes range from those producing at high volume (for
example, credit card transaction processing) to a low volume (for example, funding a
large complex takeover deal). Processes can also range from producing a very low
variety of products or services (for example, in an electricity utility) to a very high
variety (for example, in an architects’ practice). Usually, the two dimensions of volume
and variety go together in a reversed way. Low-volume processes often produce a high
variety of products and services, and high-volume processes, a narrow variety. Thus,
there is a continuum from low-volume–high variety through to high-volume–low variety,
on which we can position processes. And within a single operation there could be
processes with very different positions on this volume–variety spectrum. For example,
compare the approach taken in a medical service during mass medical treatments, such
as large-scale immunisation programmes, with that taken in transplant surgery where the
treatment is designed specifically to meet the needs of one person. In other words, no
one type of process design is best for all types of requirements in all circumstances –
different products or services with different volume–variety positions require different
processes.
Operations principle
The design of any process should be governed by the volume and variety it is required to produce.
Figure 6.4 Different process types imply different volume–variety characteristics for the process
Figure 6.4: Long Description
Process types
The position of a process on the volume–variety continuum shapes its overall design
and the general approach to managing its activities. These ‘general approaches’ are
called process types. Different terms are used to identify process types depending on
whether they are predominantly manufacturing or service processes, and there is some
variation in the terms used. For example, it is not uncommon to find the ‘manufacturing’
terms used in service industries. Figure 6.4 illustrates how these ‘process types’ are
used to describe different positions on the volume–variety spectrum.
Project processes
Project processes deal with discrete, usually highly customised products, often with
a relatively long timescale between the completion of each item, where each job has a
well-defined start and finish. Project processes have low volume and high variety.
Activities involved in the process can be ill-defined and uncertain. Transforming
resources may have to be organised especially for each item (because each item is
different). The process may be complex, partly because the activities in such processes
often involve significant discretion to act according to professional judgement.
Examples of project processes include software design, movie production, most
construction work and large fabrication operations such as those manufacturing turbo
generators.
The major construction site shown in the picture is a project process. Each ‘item’
(building) is different and poses different challenges to those running the process
(civil engineers).
Jobbing processes
Jobbing processes also deal with high variety and low volumes. However, while in
project processes each item has resources devoted more or less exclusively to it, in
jobbing processes each product has to share the operation’s resources with many others.
Resources will process a series of items but, although each one will require similar
attention, they may differ in their exact needs. Many jobs will probably be ‘one-offs’
that are never repeated. Again, jobbing processes could be relatively complex, however
they usually produce physically smaller products and, although sometimes involving
considerable skill, such processes often involve fewer unpredictable circumstances.
Examples of jobbing processes include the work of made-to-measure tailors, many
precision engineers such as specialist toolmakers, furniture restorers and the printer that
produces tickets for the local social event.
This craftsperson is using general purpose wood-cutting technology to make a
product for an individual customer. The next product made will be different (although
maybe similar) for a different customer.
Batch processes
Batch processes may look like jobbing processes, but do not have the same degree
of variety. As the name implies, batch processes produce more than one item at a time,
so each part of the process has periods when it is repeating itself, at least while the
‘batch’ is being processed. If the size of the batch is just two or three items, it is little
different to jobbing. Conversely, if the batches are large, and especially if the products
are familiar to the operation, batch processes can be fairly repetitive. Because of this,
the batch type of process can be found over a wide range of volume and variety levels.
Examples of batch processes include machine tool manufacturing, the production of
some special gourmet frozen foods and the manufacture of most of the component parts
that go into mass-produced assemblies such as vehicles.
In this kitchen, food is being prepared in batches. All batches go through the same
sequence (preparation, cooking and storage) but each batch is of a different dish.
Mass processes
Mass processes are those that produce items in high volume and relatively narrow
variety (narrow in terms of its fundamentals – a vehicle assembly process might
produce thousands of variants, yet essentially the variants do not affect the basic
process of production). The activities of mass processes are usually repetitive and
largely predictable. Examples of mass processes include frozen food production,
automatic packing lines, vehicle production plants and television factories.
The vehicle production plant is everyone’s idea of a mass process. Each product is
almost (but not quite) the same, and made in large quantities.
Continuous processes
Continuous processes
have even higher volume and usually lower variety than mass
processes. They also usually operate for longer periods of time. Sometimes they are
literally continuous in that their products are inseparable, being produced in an endless
flow. They often have relatively inflexible, capital-intensive technologies with highly
predictable flow, and although products may be stored during the process, their
predominant characteristic is of smooth flow from one part of the process to another.
Examples of continuous processes include water processing, petrochemical refineries,
electricity utilities, steel making and some paper making.
This continuous water treatment plant almost never stops (it only stops for
maintenance) and performs only one task (filtering impurities). Often, we only notice
the process if it goes wrong.
Professional services
Professional services
are high-contact processes where customers spend a
considerable time in the service process. They can provide high levels of customisation
(the process being highly adaptable in order to meet individual customer needs).
Professional services tend to be people-based rather than equipment-based, and usually
staff are given considerable discretion in servicing customers. Professional services
include management consultants, lawyers’ practices, architects, doctors’ surgeries,
auditors, health and safety inspectors, and some computer field service operations.
Here consultants are preparing to start a consultancy assignment. They are
discussing how they might approach the various stages of the assignment, from
understanding the real nature of the problem through to the implementation of their
recommended solutions. This is a process map, although a very high-level one. It
guides the nature and sequence of the consultants’ activities.
Service shops
Service shops
have levels of volume and variety (and customer contact,
customisation and staff discretion) between the extremes of professional and mass
services (see next paragraph). Service is provided via mixes of front- and back-office
activities. Service shops include banks, high-street shops, holiday tour operators, car
rental companies, schools, most restaurants, hotels and travel agents.
The health club shown in the picture has front-office staff who can give advice on
exercise programmes and other treatments. Although every client has a unique fitness
programme, certain activities (for example, safety issues) have to follow defined
processes.
Mass services
Mass services have many customer transactions, involving limited contact time and
little customisation. Staff are likely to have a relatively defined division of labour
and have to follow set procedures. Mass services include supermarkets, a national rail
network, an airport, telecommunications service, library, television station, the police
service and the enquiry desk at a utility. For example, one of the most common types of
mass service are the call centres used by almost all companies that deal directly with
consumers. Coping with a very high volume of enquiries requires some kind of
structuring of the process of communicating with customers. This is often achieved by
using a carefully designed enquiry process (sometimes known as a script).
Operations principle
Process types indicate the position of processes on the volume–variety spectrum.
This is the ‘back office’ of part of a retail bank (the type that we all use). It is a call
centre that deals with many thousands of calls every day. Staff are required to follow
defined processes (scripts) to make sure customers receive a standard service.
The product–process matrix
The most common method of illustrating the relationship between a process’s volume–
variety position and its design characteristics is shown in Figure 6.5. Often called the
‘product–process ’ matrix, it can in fact be used for any type of process, whether
7
producing products or services. The underlying idea of the product–process matrix is
that many of the more important elements of process design are strongly related to the
volume–variety position of the process. So, for any process, the tasks that it undertakes,
the flow of items through the process, the layout of its resources, the technology it uses
and the design of jobs, are all strongly influenced by its volume–variety position. This
means that most processes should lie close to the diagonal of the matrix that represents
the ‘fit’ between the process and its volume–variety position. This is called the
‘natural’ diagonal, or the ‘line of fit’.
Critical commentary
Although the idea of process types can be useful, it is in many ways simplistic. In reality, there is no clear
boundary between process types. For example, many processed foods are manufactured using massproduction processes but in batches. So, a ‘batch’ of one type of cake (say) can be followed by a ‘batch’
of a marginally different cake (perhaps with different packaging), followed by yet another, etc. Essentially
this is still a mass process, but not quite as pure a version of mass processing as a manufacturing
process that only makes one type of cake. Similarly, the categories of service processes are likewise
blurred. For example, a specialist camera retailer would normally be categorised as a service shop, yet it
also will give, sometimes very specialised, technical advice to customers. It is not a professional service
like a consultancy, of course, but it does have elements of a professional service process within its
design. This is why the volume and variety characteristics of a process are sometimes seen as being a
more realistic way of describing processes. The product–process matrix adopts this approach.
Moving off the natural diagonal
A process lying on the natural diagonal of the matrix shown in Figure 6.5 will normally
have lower operating costs than one with the same volume–variety position that lies off
the diagonal. This is because the diagonal represents the most appropriate process
design for any volume–variety position. Processes that are on the right of the ‘natural’
diagonal would normally be associated with lower volumes and higher variety. This
means that they are more flexible than is warranted by their actual volume–variety
position. They are not taking advantage of their ability to standardise their activities, so
their costs are likely to be higher than if they were closer to the diagonal. Conversely,
processes that are on the left of the diagonal have adopted a position that would
normally be used for higher-volume and lower-variety processes. These processes are
‘over-standardised’ and probably too inflexible for their volume–variety position. This
lack of flexibility can also lead to high costs because the process will not be able to
8
change from one activity to another as readily as a more flexible process. So, a first
step in examining the design of an existing process is to check if it is on the natural
diagonal of the product–process matrix.
Figure 6.5 Deviating from the ‘natural’ diagonal on the product–process matrix has
consequences for cost and flexibility
Source: Based on Hayes and Wheelright (1984)
Figure 6.5: Long Description
Operations principle
Moving off the ‘natural diagonal’ of the product–process matrix will incur excess cost.
Making clothes must have been one of the very first ‘production’ tasks carried out by early humans,
and it is still an important industrial sector. With the world garment market worth around € 1.3 trillion, it
10
is estimated as employing up to 75 million people. Of all garment producing countries, China has
long been seen as the master of clothing manufacturing, and its largest producer and exporter of
apparel and textiles is the Dishang Group. Founded in 1993, Dishang has an annual turnover of $1.5
billion, producing garments for such well-known brands as Zara, Matalan and Adidas. But, although
China is still a leader in terms of technical expertise and production efficiency, when its labour costs
increased, Dishang, like other producers, expanded to set up operations in Cambodia, Myanmar and
Bangladesh. It is a huge and sophisticated enterprise, manufacturing over 73 million garments per
year from 80 wholly owned factories across 12 global locations. Its modern operations include
automated technology, full online and end-of-line quality control systems and the use of acceptable
quality limit (AQL) inspection levels according to the requirements of each customer. At Dishang’s
headquarters, customers can peruse 50,000 fabrics in the company’s large digital library. Once
customers have chosen a pattern they like, they can upload a picture into an internal system and be
shown similar styles. Dishang’s Chairman, Lihua Zhu believes that the Group’s success is down to
three factors. ‘Firstly our volume strength is very important. Due to the size of the group we can
secure the more competitive prices for our customers (when purchasing raw materials and
components). Secondly we offer in-house design expertise and have factories and internal teams
that specialise in different products, meaning brands and retailers can come to us for everything. And
thirdly, we approach international markets differently with our own offices, which saves costs by
11
cutting out the middle man’.
One certainly would not see such high volume at Sands Film’s costume-making workshop. Every
film or television programme that is set in any period, other than the present day, needs costumes for
its actors. And most films have a lot of characters, so that means a lot of costumes. Sands Films in
London has a well-established and permanent garment workshop. It is what we described earlier in
the chapter as a typical ‘jobbing’ process. Sands Films provides a wide range of wardrobe and
costume services. Its customers are the film, stage and TV production companies, each of which
have different requirements and time constraints. And because each project is different and has
different requirements, the workshop’s jobs go from making a single simple outfit to providing a wide
variety of specially designed costumes and facilities over an extended production period. The facilities
include most normal tailoring processes such as cutting, dyeing and printing, and varied specialist
services such as corset and crinoline making as well as millinery (hats). During the design and
making process actors often visit the workshop, which has been called an ‘Aladdin’s cave’ of
theatrical costumes. The workshop is where actors come face to face with their character for the first
time. Making a costume can only start once a project has been approved and a costume designer
appointed, although discussions with the workshop may have started prior to that. When the budget
and the timing have been agreed, the designer can start to present ideas and finished design to the
workshop. And although the processes in the workshop are well established, each costume requires
different skills and so has different routes through the stages.
Worked example
The ‘meter installation’ unit
The ‘meter installation’ unit of a water utility company installed and repaired water meters. Each
installation job could vary significantly because meters had to be fitted into different water pipe
systems. When a customer requested an installation, a supervisor would survey the customer’s
water system. An appointment would then be made for an installer to visit the customer and install the
meter. Then the company decided to install a new ‘standard’ remote-reading meter to replace the
wide range of existing meters. This new meter was designed to make installation easier by including
universal quick-fit joints that reduced installation times. As a pilot, it was also decided to prioritise
those customers with the oldest meters and conduct trials of how the new meter worked in practice.
All other aspects of the installation process were left as they were. However, after the new meters
were introduced the costs of installation were far higher than forecast, so the company decided to cut
out the survey stage of the process because, using the new meter, 98 per cent of installations could
be fitted in one visit. Just as significantly, fully qualified installers were often not needed, so installation
could be performed by less-expensive staff.
This example is illustrated in Figure 6.6. The initial position of the installation process is at point A.
The installation unit was required to install a wide variety of meters into a very wide variety of water
systems. This needed a survey stage to assess the nature of the job and the use of skilled staff to
cope with the complex tasks. The installation of the new type of meter changed the volume–variety
position for the process, reducing the variety and increasing volume. However, the process was not
changed so the design of the process was appropriate for its old volume–variety position, but not the
new one. In effect, it had moved to point B in Figure 6.6. It was off the diagonal, with unnecessary
flexibility and high operating costs. Redesigning the process to take advantage of the reduced variety
and complexity of the job (position C on Figure 6.6) allowed installation to be performed more
efficiently.
Figure 6.6 A product–process matrix with process positions from the water
meter example
Figure 6.6: Long Description
6.4 How are processes designed in detail?
After the overall design of a process has been determined, its individual activities must be
configured. At its simplest this detailed design of a process involves identifying all the
individual activities that are needed to meet the objectives of the process, and deciding on the
sequence in which these activities are to be performed and who is going to do them. There will,
of course, be some constraints to this. Some activities must be carried out before others and
certain people or equipment can only do some activities. Nevertheless, for a process of any
reasonable size, the number of alternative process designs is usually large. Because of this,
process design is often done using some simple visual approach such as process mapping.
Process mapping
Process mapping simply involves describing processes in terms of how the activities within the
process relate to each other. There are many techniques that can be used for process mapping
(or process blueprinting, or process analysis, as it is sometimes called). They all identify the
different types of activity and show the flow of materials or people or information through the
process. They use process mapping symbols to classify different types of activity. And although
there is no universal set of symbols, there are some that are commonly used. Most of these
derive either from the early days of ‘scientific’ management around a century ago (see Chapter
9) or, more recently, from information system flowcharting. These symbols can be arranged in
order, and in series or in parallel, to describe any process. For example, Figure 6.7 shows one
of the processes used in a theatre lighting operation that hires out lighting and stage effects
equipment to theatrical companies and event organisers. It deals with how customers’ calls are
processed by the technicians.
Operations principle
Process mapping is needed to expose the reality of process behaviour.
Figure 6.7 Process map for ‘enquire to delivery’ process at stage lighting operation
Figure 6.7: Long Description
Figure 6.8 The ‘supply and install’ operations process mapped at three levels
Figure 6.8: Long Description
Different levels of process mapping
For a large process, drawing process maps at this level of detail can be complex. This is why
processes are often mapped at a more aggregated level, called high-level process mapping,
before more detailed maps are drawn. Figure 6.8 illustrates this for the total ‘supply and install
lighting’ process in the stage lighting operation. At the highest level the process can be drawn
simply as an input–transformation–output process with materials and customers as its input
resources and lighting services as outputs. No details of how inputs are transformed into outputs
are included. At a slightly lower or more detailed level, what is sometimes called an outline
process map (or chart) identifies the sequence of activities but only in a general way. So, the
process of ‘enquire to delivery’ that is shown in detail in Figure 6.7 is here reduced to a single
activity. At the more detailed level, all the activities are shown in a ‘detailed process map’ (the
activities within the process ‘install and test’ are shown).
Although not shown in Figure 6.8, an even more micro set of process activities could be
mapped within each of the detailed process activities. Such a micro detailed process map could
specify every single motion involved in each activity. Some quick-service restaurants, for
example, do exactly that. In the lighting hire company example, most activities would not be
mapped in any more detail than that shown in Figure 6.8. Some activities, such as ‘return to
base’, are probably too straightforward to be worth mapping any further. Other activities, such
as ‘rectify faulty equipment’, may rely on the technician’s skills and discretion to the extent that
the activity has too much variation and is too complex to map in detail. Some activities,
however, may need mapping in more detail to ensure quality or to protect the company’s
interests. For example, the activity of safety checking the customer’s site to ensure that it is
compliant with safety regulations will need specifying in some detail to ensure that the company
can prove it exercised its legal responsibilities.
Mapping visibility in process design
‘Processing’ people is different. Processes with a high level of customer ‘visibility’ cannot be
designed in the same way as processes that deal with inanimate materials or information. As we
discussed in Chapter 1, operations and processes that primarily ‘transform’, people experience
the process. When customers ‘see’ part of the process, it is useful to map them in a way that
makes the degree of visibility of each part of the process obvious. Figure 6.9 shows yet another
part of the lighting equipment company’s operation: ‘the collect and check’ process. The
process is mapped to show the visibility of each activity to the customer. Here four levels of
visibility are used. There is no hard and fast rule about this; many processes simply distinguish
between those activities that the customer could see and those that they couldn’t. The boundary
between these two categories is often called the ‘line of visibility’. In Figure 6.9 three
categories of visibility are shown. At the very highest level of visibility, above the ‘line of
interaction’, are those activities that involve direct interaction between the lighting company’s
staff and the customer. Other activities take place at the customer’s site or in the presence of the
customer but involve less or no direct interaction. Yet further activities (the two transport
activities in this case) have some degree of visibility because they take place away from the
company’s base and are visible to potential customers, but are not visible to the immediate
customer.
Visibility, customer experience and emotional mapping
When customers experience a process, it results in the customer feeling emotions, not all of
which are necessarily rational. Most of us have been made happy, angry, frustrated, surprised,
reassured or furious as customers in a process. Nor is the idea of considering how processes
affect customer emotions confined to those processes that are intended to engage the emotions:
for example, entertainment-type organisations such as theme parks. Any high customer contact
product (or more likely, service) always creates an experience for the customer. Moreover,
customer experience will affect customer satisfaction, and therefore has the potential to produce
customer loyalty, influence expectations and create emotional bonds with customers. This is
why many service organisations are seeing how customers experience their processes (the socalled ‘customer journey’) at the core of their process design.
Figure 6.9 The ‘collect and check’ process mapped to show different levels of process visibility
Figure 6.9: Long Description
Designing the customer experience
Designing processes with a significant experience content requires the systematic consideration
of how customers may react to the experiences that the process exposes them to. This will
include the sights, sounds, smells, atmosphere and general ‘feeling’ of the service. The concept
of a ‘service scape ’, which is discussed in the next chapter, is strongly related to
consideration of engaging customers so that they connect with the process in a personal way.
One of the most common methods of designing such processes is to consider what are commonly
called ‘touchpoints’. These have been described as ‘Everything the consumer uses to verify
12
their service’s effectiveness’.
They are the points of contact between a process and
customers, and there might be many different touchpoints during the customer journey. It is the
accumulation of all the experiences from every touchpoint interaction that shapes customers’
judgement of the process. The features of a process at the touchpoints are sometimes called
‘clues’ (or ‘cues’): these are the messages that customers receive or experience as they progress
through the process. The emotions that result from these clues contain the messages that the
customer will receive and therefore influence how a customer will judge the process.
When designing processes, managers need to ensure that all the messages coming from the
clues at each stage of the process are consistent with the emotions they want the customers to
experience and do not give them wrong or misleading messages about the process. In the same
way as process mapping indicates the sequence and relationship between activities, so
emotional mapping can indicate the type of emotions engendered in the customer’s mind as they
experienced the process. Figure 6.10 shows how this might work for a visit to a clinic for a
computerised tomography (CT) scan. There are many ways that emotions can be mapped and
different diagrammatic representations can be used. In this case experiences are captured by
asking what the patient is intended to, and actually, thinks, feels, says and acts. From this, a
simple scoring system has been used ranging from +3 (very positive) to –3 (very negative).
Operations principle
The design of processes that deal with customers should consider the emotions engendered at each stage of the
process.
Throughput time, cycle time and work-in-progress
So far, we have looked at the more conceptual (process types) and descriptive (process
mapping) aspects of process design. We now move on to the equally important analytical
perspective. And the first stage is to understand the nature of, and relationship between,
throughput time, cycle time and work-in-progress. As a reminder, throughput time is the elapsed
time between an item entering the process and leaving it, cycle time is the average time between
items being processed and work-in-progress is the number of items within the process at any
point in time. In addition, the work content for each item will also be important for some
analysis. It is the total amount of work required to produce a unit of output. For example,
suppose that in an assemble-to-order sandwich shop, the time to assemble and sell a sandwich
(the work content) is two minutes and that two people are staffing the process. Each member of
staff will serve a customer every two minutes; therefore, every two minutes, two customers are
being served and so on average a customer is emerging from the process every minute (the cycle
time of the process). When customers join the queue in the process they become work-inprogress (sometimes written as WIP). If the queue is ten people long (including that customer)
when the customer joins it, they will have to wait ten minutes to emerge from the process. Or put
more succinctly:
Throughput time = Work-in-progress × Cycle time
In this case:
10 minutes wait = 10 people in the system × 1 minute per person
Figure 6.10 Customer experience map of a visit for a computerised tomography (CT) scan
Figure 6.10: Long Description
Little’s law
This mathematical relationship (throughput time = work-in-progress × cycle time) is called
Little’s law. It is simple but very useful, and it works for any stable process. Little’s law states
that the average number of things in the system is the product of the average rate at which things
leave the system and the average time each one spends in the system. Or, put another way, the
average number of objects in a queue is the product of the entry rate and the average holding
time. For example, suppose it is decided that in a new sandwich assembly and sales process, the
average number of customers in the process should be limited to around ten and the maximum
time a customer is in the process should be on average four minutes. If the time to assemble and
sell a sandwich (from customer request to the customer leaving the process) in the new process
has been reduced to 1.2 minutes, how many staff should be serving?
Operations principle
Process analysis derives from an understanding of the required process cycle time.
Putting this into Little’s law:
Throughput time =
4 minutes
Work in progress, WIP =10
So, since:
Throughput time =WIP × cycle time
Cycle time =Throughput time
WIP
4
The cycle time f or the process =
= 0.4 minutes
10
That is, a customer should emerge from the process every 0.4 minutes, on average.
Given that an individual can be served in 1.2 minutes:
1.2
The number of servers required =
= 3
0.4
In other words, three servers would serve three customers in 1.2 minutes: that is, one
customer in 0.4 minutes.
Operations principle
Little’s law states that throughput time = work-in-progress × cycle time.
Worked example
You’ll never get them back in time
Mike was totally confident in his judgement: ‘You’ll never get them back in time’, he said. ‘They aren’t just
wasting time, the process won’t allow them all to have their coffee and get back for 11 o’clock’. Looking outside
the lecture theatre, Mike and his colleague Dina were watching the 20 businesspeople who were attending the
seminar queuing to be served coffee and biscuits. The time was 10.45 and Dina knew that unless they were all
back in the lecture theatre at 11 o’clock there was no hope of finishing his presentation before lunch. ‘I’m not
sure why you’re so pessimistic’, said Dina. ‘They seem to be interested in what I have to say and I think they
will want to get back to hear how operations management will change their lives’. Mike shook his head. ‘I’m not
questioning their motivation’, he said, ‘I’m questioning the ability of the process out there to get through them all
in time. I have been timing how long it takes to serve the coffee and biscuits. Each coffee is being made fresh
and the time between the server asking each customer what they want and them walking away with their coffee
and biscuits is 48 seconds. Remember that, according to Little’s law, throughput equals work-in-progress
multiplied by cycle time. If the work-in-progress is the 20 managers in the queue and cycle time is 48 seconds,
the total throughput time is going to be 20 multiplied by 0.8 minutes which equals 16 minutes. Add to that
sufficient time for the last person to drink their coffee and you must expect a total throughput time of a bit over
20 minutes. You just haven’t allowed long enough for the process’. Dina was impressed. ‘Err . . . what did you
say that law was called again?’ ‘Little’s law’, said Mike.
Worked example
Workstation renovation
Every year it was the same. All the workstations in the building had to be renovated (tested, new software
installed, etc.) and there was only one week in which to do it. The one week fell in the middle of the August
vacation period when the renovation process would cause minimum disruption to normal working. Last year the
company’s 500 workstations had all been renovated within one working week (40 hours). Each renovation last
year took on average 2 hours and 25 technicians had completed the process within the week. This year there
would be 530 workstations to renovate but the company’s IT support unit had devised a faster testing and
renovation routine that would only take on average 1.5 hours instead of 2 hours. How many technicians will be
needed this year to complete the renovation processes within the week?
Last year:
500 workstations
Work-in-progress (WIP) =
40 hours
Time available (T t ) =
Average time to renovate =
Theref ore throughput rate (T r ) =
2 hours
0.5 hours per technician
=0.5N
where N
=number of technicians
Little′s law WIP =T
t
500 =
N
40 × 0.5N
=
=
× Tr
500
40 × 0.5
25 technicians
This year:
530 workstations
Work in progress (WIP) =
40 hours
Time available =
Average time to renovate =
Throughput rate (T r ) =
1.5 hours
1/1.5 per technician
=0.67N
where N
=number of technicians
Little′s law WIP =T
530 =
N
=
=
t
× Tr
40 × 0.67N
530
40 × 0.67
19.88 (say 20) technicians
Throughput efficiency
This idea that the throughput time of a process is different from the work content of whatever it
is processing has important implications. What it means is that for significant amounts of time no
useful work is being done to the materials, information or customers that are progressing through
the process. In the case of the simple example of the sandwich process described earlier,
customer throughput time is restricted to 4 minutes, but the work content of the task (serving the
customer) is only 1.2 minutes. So, the item being processed (the customer) is only being
‘worked on’ for 1.2/4 = 30 per cent of its time. This is called the throughput efficiency of the
process:
Work content
Percentage throughput ef f iciency =
× 100
Throughput time
In this case the throughput efficiency is very high, relative to most processes, perhaps
because the ‘items’ being processed are customers who react badly to waiting. In most material
and information transforming processes, throughput efficiency is far lower, usually in single
percentage figures.
Worked example
The vehicle licensing centre
A vehicle licensing centre receives application documents, keys in details, checks the information provided on
the application, classifies the application according to the type of licence required, confirms payment and then
issues and mails the licence. It is currently processing an average of 5,000 licences every eight-hour day. A
recent spot check found 15,000 applications that were ‘in progress’ or waiting to be processed. The sum of all
activities that are required to process an application is 25 minutes. What is the throughput efficiency of the
process?
Work in progress =15,000 applications
Cycle time =time producing
Time producing
8 hours
=
Number produced
480 minutes
=
5, 000
= 0.096 minutes
5,000
From Little's law, throughput time = WIP × cycle time
Throughput time = 15,000 × 0.096
= 1,440 minutes = 24 hours = 3 days of working
Work content
Throughput ef f iciency =
25
=
Throughput time
= 1.74 per cent
1, 440
Although the process is achieving a throughput time of three days (which seems reasonable for this kind of
process) the applications are only being worked on for 1.7 per cent of the time they are in the process.
Value-added throughput efficiency
The approach to calculating throughput efficiency that is described above assumes that all the
‘work content’ is actually needed. Yet we have already seen from ‘The vehicle licensing centre’
worked example that changing a process can significantly reduce the time that is needed to
complete the task. Therefore, work content is actually dependent upon the methods and
technology used to perform the task. It may be also that individual elements of a task may not be
considered ‘value-added’. In ‘The vehicle licensing centre’ worked example, the new method
eliminated some steps because they were ‘not worth it’: that is, they were not seen as adding
value. So, value-added throughput efficiency restricts the concept of work content to only
those tasks that are literally adding value to whatever is being processed. This often eliminates
activities such as movement, delays and some inspections. For example, if in the worked
example ‘The vehicle licensing centre’, of the 25 minutes of work content, only 20 minutes was
actually adding value, then:
20
Value-added throughput ef f iciency =
= 1.39 per cent
1, 440
Workflow
When the transformed resource in a process is information (or documents containing
information), and when information technology is used to move, store and manage the
information, process design is sometimes called ‘workflow’ or ‘workflow management’. It is
defined as ‘the automation of procedures where documents, information or tasks are passed
between participants according to a defined set of rules to achieve, or contribute to, an overall
business goal’. Although workflow may be managed manually, it is almost always managed
using an IT system. The term is also often associated with business process reengineering (see
Chapter 15). More specifically, workflow is concerned with the following:
▶ Analysis, modelling, definition and subsequent operational implementation of business
processes.
▶ The technology that supports the processes.
▶ The procedural (decision) rules that move information/documents through processes.
▶ Defining the process in terms of the sequence of work activities, the human skills needed to
perform each activity and the appropriate IT resources.
Process bottlenecks
A bottleneck in a process is the activity or stage where congestion occurs because the workload
placed is greater than the capacity to cope with it. In other words, it is the most overloaded part
of a process. And as such it will dictate the rate at which the whole process can operate. For
example, look at the simple process illustrated in Figure 6.11. It has four stages and the total
time to complete the work required for each item passing through the process is 10 minutes. In
this simple case, each of the four stages has the same capacity. In the first case (a) the 10
minutes of work is equally allocated between the four stages, each having 2.5 minutes of work.
This means that items will progress smoothly through the process without any stage holding up
the flow, and the cycle time of the process is 2.5 minutes. In the second case (b) the work has
not been allocated evenly. In fact, this is usually the case because usually it is difficult (actually
close to impossible) to allocate work absolutely equally. In this case, stage 4 of the process has
the greatest load (3 minutes). It is the bottleneck, and will constrain the cycle time of the process
to 3 minutes. Bottlenecks reduce the efficiency of a process because, although the bottleneck
stage will be fully occupied, the other stages will be underloaded. The activity of trying to
allocate work equally between stages is called ‘balancing’.
Operations principle
Allocating work equally to each stage in a process (balancing) smooths flow and avoids bottlenecks.
Figure 6.11 The bottleneck is that part of the process that is the most overloaded relative to its capacity
Figure 6.11: Long Description
Anyone who has travelled on a busy mass transport system like London Underground know how busy it can be,
often with queues of passengers building up at various points as they move to or from their trains. One point
that can become a bottleneck for passengers on London Underground is the escalators. Traditionally, in
London, passengers stand on the right-hand side of the escalator, leaving the left-hand side free for those who
want to walk up or down. But in an attempt to reduce the bottleneck at the escalators, Transport for London,
who run the system, trialled a new arrangement that they believed would increase the capacity of its escalator
at the Holborn station. Building new stations is expensive, so any way of increasing the capacity of existing
ones is going to be attractive, and Holborn is a particularly busy station. The new (and radical, for Londoners)
arrangement was to instruct passengers at peak times not to walk, but to stand on both sides of the escalator.
The decision was also based on the fact that the escalators at Holborn are over 24 metres high. Apparently,
height makes a big difference to the willingness of passengers to walk up escalators. When they are only a few
metres high, most people will walk up them. At 30 metres, only the very energetic will. As shown in Figure 6.12,
the trial was technically successful in that capacity increased significantly. However, the experiment was not
made permanent. Why? Apparently it offended two aspects of human behaviour. First, it slowed the (vocal)
minority of people who want to race up the escalator as their gym workout for the day. Second, it took away the
feeling that travellers had at least some degree of choice (even if most chose not to exercise it).
Figure 6.12 Requiring passengers to stand on both sides of the escalator makes it less of a bottleneck
Figure 6.12: Long Description
Balancing work time allocation
Allocating work to process stages must respect the ‘precedence’ of the individual tasks that
make up the total work content of the job that the process is performing. The most common way
of showing task precedence is by using a ‘precedence diagram’. This is a representation of the
ordering of the elements, where individual tasks are represented by circles connected by
arrows, which signify the ordering of the tasks. Figure 6.13 in the worked example ‘Karlstad
Kakes’ illustrates how precedence diagrams can be used.
Operations principle
Process design must respect task precedence.
Worked example
Karlstad Kakes
Karlstad Kakes (KK) is a manufacturer of speciality cakes, which has recently obtained a contract to supply a
major supermarket chain with a speciality cake in the shape of a space rocket. It has been decided that the
volumes required by the supermarket warrant a special production process to perform the finishing, decorating
and packing of the cake. This line would have to carry out the elements shown in Table 6.2.
Table 6.2 The individual tasks that make up the total job of the finishing, decorating and packing of the cake
Figure 6.13 shows the precedence diagram for the total job. The initial order from the supermarket is for
5,000 cakes a week and the number of hours worked by the factory is 40 per week. From this:
40 hrs × 60 mins
The required cycle time =
5,000
= 0.48 mins
1.68 mins (the total work content)
The required number of stages =
0.48 mins (the required cycle time)
= 3.5 stages
This means four stages.
Working from the left on the precedence diagram, tasks a and b can be allocated to stage 1. Allocating task c
to stage 1 would exceed the cycle time. In fact, only task c can be allocated to stage 2 because including task d
would again exceed the cycle time. Task d can be allocated to stage 3. Either task e or task f can also be
allocated to stage 3, but not both or the cycle time would be exceeded. In this case, task e is chosen. The
remaining tasks then are allocated to stage 4. The dotted lines in Figure 6.13 show the final allocation of tasks
to each of the four stages.
Figure 6.13 Precedence diagram for Karlstad Kakes with allocation of tasks to each stage
Figure 6.13: Long Description
Arranging the stages
All the stages necessary to fulfil the requirements of the process may not be arranged in a
sequential ‘single line’. For example, suppose a mortgage application process requires four
stages working on the task to maintain a cycle time of one application processed every 15
minutes. One possible arrangement of the four stages would be to arrange them sequentially,
each stage having 15 minutes’ worth of work. However, (theoretically) the same output rate
could also be achieved by arranging the four stages as two shorter lines, each of two stages with
30 minutes’ worth of work. Alternatively, following this logic to its ultimate conclusion, the
stages could be arranged as four parallel stages, each responsible for the whole work content.
Figure 6.14 shows these options.
This is a simplified example, but it represents a genuine issue. Should the process be
organised as a single ‘long-thin ’, sequential arrangement, or as several ‘short-fat ’,
parallel arrangements, or somewhere in-between? (Note that ‘long’ means the number of stages
and ‘fat’ means the amount of work allocated to each stage.) In any particular situation, there are
usually technical constraints, which limit either how ‘long and thin’ or how ‘short and fat’ the
process can be, but there is usually a range of possible options within which a choice needs to
be made. The advantages of each extreme of the long thin to short fat spectrum are very different
and help to explain why different arrangements are adopted.
The advantages of the long-thin arrangement include:
▶ Controlled flow of items – This is easy to manage.
▶ Simple handling – This is especially true if the items being processed are heavy, large or
difficult to move.
▶ Lower capital requirements – If a specialist piece of equipment is needed for one task in the
job, only one piece of equipment would need to be purchased; on short-fat arrangements
every stage would need one.
▶ More efficient operation – If each stage is performing only a small part of the total job, the
person at the stage will have a higher proportion of direct productive work as opposed to
the non-productive parts of the job, such as picking up tools and materials.
(This latter point is particularly important and is fully explained in Chapter 9 when we discuss
job design.)
Figure 6.14 The arrangement of stages in a process can be described on a spectrum from ‘long thin’ to ‘short fat’
Figure 6.14: Long Description
The advantages of the short-fat arrangement include:
▶ Higher mix flexibility – If the process needs to work on several types of item, each stage or
whole process could specialise in different types.
▶ Higher volume flexibility – As volume varies, stages can simply be closed down or started
up as required; long-thin arrangements would need rebalancing each time the cycle time
changed.
▶ Higher robustness – If one stage breaks down or ceases operation in some way, the other
parallel stages are unaffected; a long-thin arrangement would cease operating completely.
▶ Less monotonous work – In the mortgage example, the staff in the short-fat arrangement are
repeating their tasks only every hour; in the long-thin arrangement, it is every 15 minutes.
Responsible operations
In every chapter, under the heading of ‘Responsible operations’, we summarise how the particular topic covered
in the chapter touches upon important social, ethical and environmental issues.
When we presented the advantages and disadvantages of designing processes to have a short cycle time
(as in the ‘long-thin’ arrangement of process stages) we made no comment, or judgement, on the ethics of
such highly repetitive work. Yet there clearly are issues of how repeating the same tasks over long periods, with
little or no variety, affects both the mental and physical health of those asked to perform such work. That
relatively few people actively prefer highly repetitive work with little variety is itself a telling indication of how
unattractive most of us would find it. The general assumption is that task repetition leads to monotony,
boredom, a lack of a feeling of meaningfulness and an increase in stress. In turn, this increases the chance of
job dissatisfaction, absenteeism, burnout and labour turnover. In fact, the relationship between repetitive jobs
and negative psychological effects is more nuanced than this, but nevertheless it would be difficult to argue that
repetitive jobs are more fulfilling than those with more variety. Moreover, it is not only the potential psychological
damage of repetitive work that needs to be considered. When the repeated task is physical, as in assembly-line
work, there is also the possibility of damage such as hand pain and tendonitis. Again, studies are not always
totally unambiguous, but repeated movements that involve awkward positions, and especially high levels of
14
force and repetition have been shown to be associated with physical pain and disorders.
Confronting the negative effect of repetitive jobs often involves one of two ‘solutions’ – redesign the job so that
it is less repetitive (job enrichment) or automate it so that a human does not have to do it. (We will deal with
the idea of job enrichment in Chapter 9 (People in operations).) The automation of repetitive jobs was, for many
years, confined to some manufacturing tasks. However, the increase in the use of robotic process
automation (RPA, treated later) and even artificial intelligence (AI)
is allowing the automation of routine
tasks, such as those found in the back-office processes found in many professional services.
Low-volume, high-variety processes
Many of the ideas and analytical approaches described in this chapter derive largely from highvolume, low-variety processes. This does not mean that they cannot be used in low-volume,
high-variety process design, but they often have to be modified or adapted in some way. For
example, splitting activities into very small increments so that work can be balanced between
stages (see earlier) is often neither possible nor desirable when the variety of activities is very
wide. This does not mean that trying to allocate work equally between work groups is not
important, just that it will need to be done in a more approximate way. Even process mapping
can be problematic. Some low-volume, high-variety processes are intrinsically difficult to
describe as simple step-by-step sequential activities. There may be many alternative routes
through a process that can be taken by whatever is being processed. Decisions about how to
treat whatever is being processed may be a matter of judgement. The exact circumstances
associated with processing something may not have occurred before. If it is information that is
being processed, the information may be partial, uncertain or ambiguous.
Automating processes
The majority of processes used in this chapter to illustrate various aspects of process design are
essentially manual in nature. That is, they involve a person or persons performing some kind of
tasks. Go back far enough and almost all processes (although they would not have been called
that) would have been manual. The history of operations management could be told as one of the
progressive substitution of technology for people-based effort. First it was manufacturing
processes that were automated in some way; and although there are still plenty of manual
manufacturing processes, there are also factories that operate virtually ‘dark’, with very few
humans involved. The equivalent automation in many service operations, especially those that
primarily process information are specifically designed information technology (IT) systems.
Anyone with a bank account is the recipient of the service provided by these IT systems. They
are automated mass processes, usually designed from first principles, that may be sometimes
rigid and impersonal but are remarkably efficient compared with performing such tasks
manually.
Robotic process automation
Yet, although what earlier we called ‘core’ operations processes, especially high-volume ones,
have increasingly become automated, there are many processes outside the operations function
that could be automated. These processes are often lower volume than routine core operations
processes, yet still follow a logical set of rules. They have been called ‘swivel chair’ processes
– meaning that people take inputs of information from one set of systems (emails for example),
process the information using rules, and then record the processed outputs into another system.
Examples might include the ‘onboarding’ process for new employees, recording and entering
invoices onto internal payment authorisation systems, entering details of client information onto
customer relationship management systems, and so on. Typically, these processes are routine,
predictable, rules-based and performed by professional employees whose time could be more
profitably employed.
This is the area of application for what has become known (rather tautologically) as ‘robotic
process automation’, or RPA. It is a general term for tools that function on the human interface of
other computer systems. It does not, of course, use actual physical robots, rather it deploys
software routines (often just called ‘bots’) to perform the most mundane and repetitive tasks
previously done by people. Its aim is to enhance efficiency by automating the everyday
processes that would otherwise require human effort. Admittedly, the same aim could be
attributed to almost any IT-based automation. The difference between RPA and traditional IT
systems is:
▶ RPA is best used away from the extremes of the volume–variety spectrum. High-volume,
low-variety processes can be automated using conventional specifically designed IT
systems. At the other extreme, very-high-variety, low-volume tasks are likely to need the
flexible thought processes and decision-making of humans. RPA can be used in-between
these extremes.
▶ RPA is relatively easy to develop compared with specifically designed IT systems. The
latter require significant systems analysis and coding skills. RPA often uses simple ‘drag
and drop’ instructions that can be used by people who understand the purpose of the
processes being automated.
▶ RPA works around existing processes rather than trying to reengineer them. It is sometimes
referred to as ‘lightweight’ IT because it tries not to disturb underlying computer systems.
The effects of process variability
So far in our treatment of process design we have assumed that there is no significant variability
either in the demand to which the process is expected to respond, or in the time taken for the
process to perform its various activities. Clearly, this is not the case in reality. So, it is
important to look at the variability that can affect processes and take account of it.
There are many reasons why variability occurs in processes. These can include the late (or
early) arrival of material, information or customers, a temporary malfunction or breakdown of
process technology within a stage of the process, the recycling of ‘mis-processed’ materials,
information or customers to an earlier stage in the process, variation in the requirements of
items being processed, etc. All these sources of variation interact with each other, but result in
two fundamental types of variability:
▶ Variability in the demand for processing at an individual stage within the process, usually
expressed in terms of variation in the inter-arrival times of items to be processed.
▶ Variation in the time taken to perform the activities (i.e. process a unit) at each stage.
Critical commentary
Some commentators are critics of the very idea of thinking in terms of ‘processes’. They claim that defining jobs
as processes incites managers to look on all activities as a machine-like set of routine activities, verging on the
mindless. At best, it encourages going through the stages in a process without thinking about what is really
involved (what is known as ‘box ticking’). At worst, defining all activities into the straitjacket of ‘process’ kills the
essential humanity of working life. The counterargument is that this is a misunderstanding of what is (or should be)
meant by a ‘process’. A process is simply a framework, around which you can think about who should do what,
and when. It simply means that one has thought about, and described, how to do something. Processes need not
necessarily be formal, highly constrained or detailed – though they might be. When a process is seen as being too
rigid, it is usually because it has been designed inappropriately for its volume–variety position.
Figure 6.15 The relationship between process utilisation and number of items waiting to be processed for constant
and variable arrival and process times
Figure 6.15: Long Description
To understand the effect of arrival variability on process performance it is first useful to
examine what happens to process performance in a very simple process as arrival time changes
under conditions of no variability. For example, the simple process shown in Figure 6.15
comprises one stage that performs exactly 10 minutes of work. Items arrive at the process at a
constant and predictable rate. If the arrival rate is one unit every 30 minutes, then the process
will be utilised for only 33.33 percent of the time, and the items will never have to wait to be
processed. This is shown as point A on Figure 6.15. If the arrival rate increases to one arrival
every 20 minutes, the utilisation increases to 50 per cent, and again the items will not have to
wait to be processed. This is point B on Figure 6.15. If the arrival rate increases to one arrival
every 10 minutes, the process is now fully utilised but, because a unit arrives just as the
previous one has finished being processed, no unit has to wait. This is point C on Figure 6.15.
However, if the arrival rate ever exceeded one unit every 10 minutes, the waiting line in front of
the process activity would build up indefinitely, as is shown as point D in Figure 6.15. So, in a
perfectly constant and predictable world, the relationship between process waiting time and
utilisation is a rectangular function as shown by the red line in Figure 6.15.
Operations principle
Variability in a process acts to reduce its efficiency.
However, when arrival and process times are variable, then sometimes the process will
have items waiting to be processed, while at other times the process will be idle, waiting for
items to arrive. Therefore, the process will have both a ‘non-zero’ average queue and be
underutilised in the same period. So, a more realistic point is that shown as point X in Figure
6.15. If the average arrival time were to be changed with the same variability, the blue line in
Figure 6.15 would show the relationship between average waiting time and process utilisation.
As the process moves closer to 100 per cent utilisation, the average waiting time will become
longer. Or, to put it another way, the only way to guarantee very low waiting times for the items
is to suffer low process utilisation.
The greater the variability in the process, the more the waiting time–utilisation relationship
deviates from the simple rectangular function of the ‘no variability’ conditions that was shown
in Figure 6.15. A set of curves for a typical process is shown in Figure 6.16(a). This
phenomenon has important implications for the design of processes. In effect, it presents three
options to process designers wishing to improve the waiting time or utilisation performance of
their processes, as shown in Figure 6.16(b):
▶ accept long average waiting times and achieve high utilisation (point X);
▶ accept low utilisation and achieve short average waiting times (point Y); or
▶ reduce the variability in arrival times, activity times, or both, and achieve higher utilisation
and short waiting times (point Z).
Figure 6.16 The relationship between process utilisation and number of items waiting to be processed for variable
arrival and activity times
Figure 6.16: Long Description
To analyse processes with both inter-arrival and activity time variability, queuing or
‘waiting line’ analysis can be used. This is treated in the supplement to Chapter 11. But do not
dismiss the relationship shown in Figures 6.15 and 6.16 as some minor technical phenomenon. It
is far more than this. It identifies an important choice in process design that could have strategic
implications. Which is more important to a business, fast throughput time or high utilisation of
its resources? The only way to have both of these simultaneously is to reduce variability in its
processes, which may itself require strategic decisions such as limiting the degree of
customisation of products or services, or imposing stricter limits on how products or services
can be delivered to customers, and so on. It also demonstrates an important point concerned
with the day-to-day management of process – the only way to absolutely guarantee 100 per cent
utilisation of resources is to accept an infinite amount of work-in-progress and/or waiting time.
Operations principle
Process variability results in simultaneous waiting and resource underutilisation.
Summary answers to key questions
6.1 What is process design?
▶ Design is the activity that shapes the physical form and purpose of both products and services and the
processes that produce them.
▶ The design activity is more likely to be successful if the complementary activities of product or service
design and process design are coordinated.
6.2 What should be the objectives of process design?
▶ The overall purpose of process design is to meet the needs of customers through achieving appropriate
levels of quality, speed, dependability, flexibility and cost.
▶ The design activity must also take account of environmental issues. These include examination of the
source and suitability of materials, the sources and quantities of energy consumed, the amount and type
of waste material, the life of the product itself and the end-of-life state of the product.
6.3 How do volume and variety affect process design?
▶ The overall nature of any process is strongly influenced by the volume and variety of what it has to
process.
▶ The concept of process types summarises how volume and variety affect overall process design.
▶ In manufacturing, these process types are (in order of increasing volume and decreasing variety)
project, jobbing, batch, mass and continuous processes. In service operations, although there is less
consensus on the terminology, the terms often used (again in order of increasing volume and
decreasing variety) are professional services, service shops and mass services.
6.4 How are processes designed in detail?
▶ Processes are designed initially by breaking them down into their individual activities. Often common
symbols are used to represent types of activity. The sequence of activities in a process is then indicated
by the sequence of symbols representing activities. This is called ‘process mapping’. Alternative
process designs can be compared using process maps and improved processes considered in terms
of their operations performance objectives.
▶ The throughput time, work-in-progress and cycle time aspects of process performance are related by a
formula known as Little’s law: throughput time equals work-in-progress multiplied by cycle time.
▶ Variability has a significant effect on the performance of processes, particularly the relationship between
waiting time and utilisation.