CHAPTER

Analysing service operations: service delivery, queuing and shift scheduling

INTRODUCTION

Chapter S1 guided you through the analysis of a manufacturing operation. In this chapter, we will apply a similar approach to analysing some common situations that occur in service operations. A typical case, the Tampopo Noodle Bar, is presented for you to analyse, using a step-by-step model to analyse the service management system, the service delivery system and the queuing system.

Aims and objectives

This chapter will provide you with experience in analysing service operations in a structured fashion, and provide practice with some commonly used tools and techniques.

After reading this chapter you will be able to:

Use Normann's model to analyse the elements of an organization's service management system

Blueprint the service delivery system and identify actual or potential failure points

Apply mathematical techniques for analysing queuing, shift scheduling, and Monte Carlo simulation.

DESIGNING AND ANALYSING SERVICE SYSTEMS

Suppose that you have been hired as a consultant by a recent lottery winner who is considering investing his winnings in setting up a local Japanese noodle bar. Your job is to analyse the operations of Wagamama and its closest competitors, and to see if the noodle bar concept could be successfully replicated in your local area. After visiting the restaurant and its competitors, and doing research over the Internet, you have gathered a lot of data on Japanese restaurants.

Your first step has been to check out the competition. With the aid of a Time Out guide to London restaurants, a visit to London's Japanese shopping mall Yaohan Plaza and its Japanese supermarket, and lots of back issues of Caterer and Hotelkeeper magazine, you have identified the main competitors: Wagamama, Wok-Wok, and Yo! Sushi.

Competitive analysis summary

Wagamama

Wagamama's philosophy is ‘to serve great, fresh and nutritious food in an elegant, yet simple environment, and to provide helpful, friendly service and value for money’. According to the company's web site, Wagamama means ‘wilfulness or selfishness: selfishness in terms of looking after oneself, looking after oneself in terms of positive eating and positive living’. The restaurant's slogan is ‘positive eating’, which is ‘consciously feeding the body the nourishment it needs to build and maintain a peak physiological state, selecting foods that cleanse and nurture; controlled, balanced consumption’.

The menu specifically points out that ‘destinational eating’ is not part of Wagamama's policy – a policy supported by the décor, which is a brightly lit, minimalist's dream.

Because of Wagamama's popularity, only those customers who arrive when the restaurant opens don't have to queue for a seat. As the evening progresses, a queue of people stretches along the stairs or corridor just inside the exit. The no-booking policy means that couples are usually seated quickly, but it is difficult to find space for larger groups. Customers queuing to be seated are offered a fascinating view of the food being prepared in a high-tech open kitchen (where dozens of cooks are preparing noodle dishes of all kinds), and the bustling dining area (where the wait staff are constantly busy inside taking customer orders). In the kitchen, a courtesy cloak check is provided, as there is no room for hats and coats downstairs.

Customers enter the huge dining area, where they are seated refectory style, at long rows of tables with bench-style seating. As soon as diners are seated they're handed menus, which describe not only the food and drinks on offer, but also Wagamama's philosophy of balanced eating and healthy living. The menu extols a world view: ‘positive meal suggestions for positive value’; ‘to cleanse and nurture – the excellent natural synergy of nutrients’; ‘helps to cleanse the body of toxins’. Both freshly-squeezed juices and Chinese tea are on offer to support this, but beer and wine are available for the hungry hedonist. If this all seems a bit too serious, then the menu's advice is less so: ‘the way of the noodle is to make slurping noises while eating – the extra oxygen adds to the taste!’.

Once customers have made their choice, waiting staff take the order and punch it into hand-held, electronic order pads, which then transmit the order to the appropriate station in the kitchen. (Customers too shy to attempt the Japanese pronunciation can order by number.) To make sure that each customer receives the right order, the wait staff also write the numbers of each dish directly onto each person's paper mat. Each dish is cooked to order and then served immediately, so even though main courses, side orders and drinks are ordered together, each is delivered to the table as soon as it has been prepared. Service is furiously paced – Wagamama is no place to linger! – but customers aren't overly pressured to finish up and leave.

Wagamama offers a variety of drinks, noodle dishes (superb bowls of ramen, soba and udon) and specialty meals such as yakitori (grilled skewers of chicken), teriyaki, and tempura-based dishes. Most dishes cost just over £5, with many of the side dishes considerably less. The minimal decor is not reflected in the generosity of the food – most people have trouble finishing one of the giant bowls of soup or generous helpings of pan-fried noodles. A surprise hit is the edamame, lightly salted fresh green soya beans which you pop from their pods, a side dish which is incredibly addictive – the Japanese equivalent of crisps or popcorn. Each restaurant serves more than 125 kg of noodles daily to over 1000 customers.

Wok-Wok

Like Wagamama, at Wok Wok the ambience is minimalist and comfortable. Wok-Wok is bright and attractive, with simple wooden tables and large windows looking out on to the main road. Frosted glass and gleaming metal are offset by earthy tones and textures to create a warm, inviting and stylish atmosphere. Central to the Wok Wok concept are the restaurants’ open kitchens, which create a focal point.

Wok-Wok is laid out and operated much like any other restaurant. Waiting staff are young and funky. Customers defy categorization: family groups with small children, lone diners, twenty- and thirty-somethings and smartly dressed middle-aged couples. Dining can be inexpensive, quick and functional, or you can hang around for a chat.

Brand Director Tania Webb developed the Wok-Wok concept based on a childhood spent in South East Asia and time spent as a restaurateur in Hong Kong, both of which are reflected in Wok-Wok's eclectic menu. Wok-Wok offers a choice of noodle- or rice-based dishes with Chinese, Japanese, Korean and Thai influences, the independent but complementary cuisines of South East Asia. Wok-Wok offers soup noodles and pan-fried noodles as well as rice dishes. Using only the freshest ingredients, which are prepared and cooked to order in the restaurant, the style of food is simple, healthy and bursting with natural flavours.

Yo! Sushi

Not surprisingly, Yo! Sushi concentrates on sushi. The flagship Yo! Sushi opened in Soho in January 1998, featuring the world's longest sushi conveyor belt. Founder (and rock-concert promoter) Simon Woodroffe borrowed the idea of a self-service, conveyor-belt sushi bar (known in Japan as kaiten) and situated it in a groovy, upbeat, hi-tech environment combining the modern world of high tech robotics and animated theatre. At Yo! Sushi, customers sit around a long U-shaped counter bar. Colour-coded plates travel around the conveyor belt, each with two pieces of sushi, with a different price assigned to each of the five colours.

Drinks, including cold Japanese beers, hot and cold sake, wine and sodas, are provided by talking robot trolleys that make their way around the restaurant, avoiding collision by the use of highly tuned sensors, and talking to the customers when they get in their way! Still and fizzy mineral water are available in unlimited quantities from pumps at each seating unit for a one-off price of £1.

At the end of the meal, staff add up the value of the plates and glasses to determine the price of the meal. Help is always available from the restaurant staff, who can be called by use of help buttons at every seating station.

There is also a list with specials behind the counter, where you can order from the chefs. Diners enjoy both gourmet and vegetarian sushi while watching sushi being made in the raw, plus live footage from Japan on Sony widescreen televisions – giant TV screens show sumo wrestling matches; and the funky music of Prince comes thumping out of loudspeakers. If you like the food (or just the concept), you can get Yo! Sushi clothing and even Yo! Sushi delivery scooters!

The Service management system

Your employer has visited Wagamama, and was impressed by the speed and throughput of the restaurant. However, he feels that the refectorystyle seating and industrial feel of Wagamama is starting to get a bit dated, and has asked you to take a look at alternate ways of setting up and running the restaurant, based on the ones that you have visited. Your first task is to recommend either Wagamama's service delivery system or an entirely new one for the new restaurant chain.

A first step that you might take in analysing a service operation is to analyse the organization's service management system using the model developed by Professor Richard Normann (1991). This model ties together five important aspects of the service management system, as shown in Figure S2.1. These five aspects – culture and philosophy, market segment, service image, service concept and service delivery system – are described further below.

Figure S2.1 The service management system (based on Normann, 1991).

Culture and philosophy

The organization's culture and philosophy are central to the service management system. This describes the overall values and principles guiding the organization, including values about human dignity and worth. It has also been called the organization's service vision.

Market segment

The market segment describes the particular types of clients for whom the service management system was designed. Some different ways of deciding on a market segment are:

Customer-orientation – a wide range of services to a limited range of customers, using a customer-centred database and developing new offerings to existing customers (e.g. Rentokil)

Service-orientation – a focused, ‘limited menu’ of services to a wide range of customers, usually through specialization in a narrow range of services (e.g. Kwik-Fit, Supasnaps)

Customer- and service-orientation – providing a limited range of services to a highly targeted set of customers (e.g. McDonald's).

Service image

The service image is an information system for influencing clients and customers. An important part of the service image is the physical environment in which the service is produced, because customers are physically present during the production of a service as well as its consumption. The physical environment comprises:

The external environment, including location, premises, ease of access, and ambiance

The internal environment, including the atmosphere and structure within which the service personnel operate.

Service concept

The service concept describes the benefits offered by the service. The service concept embodies a complex set of values – physical, psychological and emotional – and thus affects both what the company does and how it is perceived by its customers and clients. In other words, the service concept describes the way the organization would like its employees and stakeholders to perceive its service (Heskett, 1986).

The structural elements of the service concept are the delivery system, facility design, location, and capacity planning. The service design supports the service concept and strategy to provide a service with features that differentiate it from the competition. Major considerations in the design process are:

Degree of complexity and degree of divergence

Identifying customer requirements

Designing supporting facilities and facilitating goods

Queuing.

The managerial elements of the service concept are the service encounter, quality, managing capacity and demand, and information.

The service delivery system

The service delivery system is the way in which the service concept and service package are provided to the consumer. It is the process in which consumers participate and through which the product is created and delivered to customers, including personnel, clients, technology and physical support. The service delivery system is dictated by and defined by the service concept. Some of the key aspects you might want to consider are the core service, the supporting goods and services, the facilitating goods and services, the role of staff, and the entertainment provided.

The service package is the embodiment of the service concept, and includes both the physical and tangible elements of the service offering and its intellectual/intangible elements. The total service package – the bundle of goods and services (Sasser et al., 1978) – includes:

Physical items – the physical goods that are changing hands, if any (often called facilitating goods in services)

Sensual benefits – aspects that can be experienced through the sensory system (explicit intangibles)

Psychological benefits – emotional or other aspects (implicit intangibles).

Task 1: Analyse the service management system

Using Normann's model as described in this section, analyse the three service management systems described above. You might find it helpful to use a grid similar to the one below to organize your analysis, or to fill in your thoughts about each element on a chart like Figure S2.1.

Service element	Traditional restaurant	Wagamama
Culture and philosophy
Market segment
Service image
Service concept
Service delivery system

Once you have identified the five service management system elements, a good question to ask is whether the different elements fit together.

Service blueprinting

Your employer is pleased with the quality of your first consulting report, and decides that he will go ahead with the investigation into starting up his own noodle bar. As part of your report, your employer has asked you to prepare a service blueprint for the Tampopo Noodle Bar, based on your first consulting report.

Service blueprinting, developed by G. Lynn Shostack (1984), is a useful way of mapping the service process. A service blueprint can identify all the points where a customer is in contact with the service provider or the organization, and thus those points where things are likely to go wrong. A service blueprint can also be used to identify areas for process improvement. Finally, it can be a useful tool for organizations who want to replicate their service, since it identifies the critical resources and processes in use.

To draw a service blueprint, it is useful to begin by drawing a flowchart for how a customer interacts with the service operation. Three elements included in a service blueprint that make it different from the flow charts commonly used in manufacturing are:

1   Line of visibility

2   Line of internal interaction

3   Failure points.

The example shown in Figure S2.2 develops a service blueprint for a generic restaurant. After you have been through the example, you might find it useful to develop a service blueprint for the service management systems and service delivery system that you developed as Task 1.

Step 1. Identify the main processes that a customer goes through, from his or her initial contact with the restaurant until he or she leaves the restaurant. Figure S2.2 shows the most important steps in this process, although each step could be broken down into more detail.

Figure S2.2 Customer interactions with the service system.

Step 2. Identify the interactions between the customer and the frontline staff in the restaurant – the activities that go on within the line of visibility. This identifies the interactions between the customer and the front-office staff, as shown in Figure S2.3.

Figure S2.3 Customer interactions with front-line staff.

Figure S2.4 Interaction between front-office and back-office staff.

Step 3. Identify the interactions between the front-line staff and other staff that work beyond the line of visibility (Figure S2.4).

Step 4. Identify the failure points in the system, and wherever possible redesign the service system to minimize or eliminate the causes of failures (Figure S2.5).

Figure S2.5 The failure points.

In particular, you may wish to focus on how Wagamama and Yo! Sushi have each integrated service pokayokes, fail-safe devices, into their service systems.

MATHEMATICAL TOOLS AND TECHNIQUES FOR ANALYSING SERVICES

Your service blueprint has been warmly received by your employer, who has decided to go ahead with the Tampopo Noodle Bar. You have been retained to perform a more detailed service design, specifically to look at issues of queuing and shift scheduling, so that an architect can be hired to start on the physical design of the noodle bar.

Queuing

The importance of analysing queuing systems

Authentic noodle bars don't take reservations, so customers must queue once the available seats have been filled. Although customers often find slow service or long waiting times a reason to complain, organizations want to trade off customer queuing time against the cost of providing additional resources. Customers would like never to wait, and organizations would like facilities and personnel to be utilized 100 per cent of the time.

This section will help you answer the following questions:

1   How should the Clifton Noodle Bar organize its queuing system to maximize efficiency?

2   What is the best priority system to use for allocating customers to tables?

3   What effect does increasing the speed with which customers are serviced have on queues?

Mathematically analysing queuing

The five essential features of queuing systems are the calling population, the arrival process, the queue configuation, the queue discipline and the service process. Any queuing system can be represented using these five elements, as shown in Figure S2.6.

Figure S2.6 A basic model of a service system with queuing.

Let's examine these in more detail.

Calling population

This is the source of customers, which consists of the potential customers for the service. This population may be finite or infinite. A finite calling population consists of a countable number of customers, such as a company's mobile phone customers – normally only Orange's customers will call the Orange call centre, only Vodaphone customers will call its centre, etc. An infinite calling population, on the other hand, consists of an uncountable number of customers, such as the number of people who might drop into a news stand on the high street. Generally, finite and infinite calling populations can be treated similarly, unless a finite population consists of only a few people, where the behaviour of one person can conceivably affect the others.

The calling population may also be homogenous or heterogeneous. A homogenous customer set consists of customers with the same requirements; a heterogeneous customer set consists of customers who have different requirements.

Arrival process

This is how customers arrive at the service in time and space, and can be described as a probability distribution of either the number of arrivals per unit of time or the time between successive arrivals. The time between customer arrivals (inter-arrival time) generally follows an exponential distribution (Figure S2.7). The exponential distribution is characterized by most observations falling near the origin, and a long tail of decreasing numbers of observations at higher inter-arrival times. If we know the mean time between arrivals, then we can compute the probability that the time between arrivals will be time t or less using the exponential distribution.

The number of customer arrivals per unit of time can be described using a related distribution, the Poisson distribution, which gives the probability that n customers will arrive during time period t.

Figure S2.7 The exponential distribution.

Both the exponential distribution and the Poisson distribution are discrete distributions; that is, the probabilities are calculated for a specific inter-arrival time or number of arrivals rather than being continuous.

Queue configuration

This describes how many queues there are, how they are arranged, and how customers behave in them.

A service system can be designed to have either a single or multiple queues, as shown in Figure S2.8. If there are multiple servers, a service can have either one (single) or many (multiple) queues feeding these servers. On the other hand, services such as supermarkets may use single queues, but create special queues for customers with only a few items, paying cash and so on.

Figure S2.8 Different possible queuing arrangements.

As mentioned in Chapter 7, queues are not always physical lines, but we can think of these situations in the same way.

Another characteristic of interest is customer behavioural tendencies. A patient customer is one who enters the system and remains there until served. This is characteristic of queues where service is mandatory or important. Customers may also be impatient, which leads them to behave in one or more of the following ways:

Balking – customers may refuse to join a queue if it looks as though they will have to wait too long before being served

Reneging – customers may join the queue, but get tired of queuing and leave before they are served

Jockeying – customers may switch between queues if there are multiple queues moving at different speeds.

Queues can also be affected by size constraints. If the physical space for the queue is large enough to hold all customers desiring service, then the queue can be described as an infinite queue; if the space is limited, it is called a finite queue. Size constraints are not always physical, of course. If you have ever tried to reach a busy call centre or a busy web site, you will have experienced a finite queue.

Queue discipline

This describes how the next customer to be served is selected. Many queues use a first-come, first-served discipline, especially when people arrive randomly and have similar service needs. Emergency services, for example, assign priority based on the critical nature of patients’ injuries or illnesses, rather than the order of their arrival.

Service process

This includes the facilities (people, machines or both) that will service the customers. Service facilities are commonly arranged in one of the four ways shown in Figure S2.8.

The time that it takes to serve customers can also be described using a probability distribution. In this case, the exponential distribution has been found to be useful in describing the probability that the time that it takes a particular server to service a particular customer is no more than t time periods.

Mathematical analysis of queuing systems

The mathematical analysis of queues, known as queuing theory, is a core topic in operations research. Queuing theory provides managers with a way of analysing both customer waiting time and server utilization, in order to minimize the total costs of customer waiting time and idle server capacity.

This section introduces you to the basic elements of mathematical queuing analysis, which you can use to analyse the relationship between customer queuing systems and waiting times. To analyse a queuing system, you will need to estimate or observe the following information:

The rate of customer arrivals

The number of service facilities

The number of phases

The number of servers per facility

The efficiency of servers

The discipline used

The queue arrangement.

Once you know this information, you can study the following variables:

Queue length

Number of customers in the system

Waiting time in the queue

Total time in the system

Service facility utilization.

Example

Suppose that for the Tampopo Noodle Bar you have been asked to analyse the queue of people waiting to be seated, and that you have decided to use a single-server queue for such customers. The assumptions for the single-server model are:

Infinite input source

No balking or reneging

Arrival distribution – Poisson

Service distribution – exponential

Queue – unlimited length

Priority discipline – first-come, first-served.

Based on obervations of similar restaurants, we have estimated that 75 customers will arrive per hour. This figure gives us 1.25 customers arriving per minute, which is an average of 0.8 minutes between customer arrivals. If you observed a queuing system for 200 customer arrivals, you might expect to observe arrival times something like those seen in Figure S2.7.

We also estimate that the customers can be seated at a rate of 80 per hour.

We use the following notation for queuing models:

S = mean service rate per server

A = mean arrival rate

Q = average number in the queue

U = utilization of the service facility

N = number of customers

W = average time spend in the system

The average utilization of the restaurant is:

U = A/S = 75/80 = 93.75%.

The average number of customers in the system is:

N = A/(S – A) = 75/(80 – 75) = 75/5 = 15 people.

The average number of customers in the queue is:

Q = U(S – A) = 93.75*5 = 14.06.

The average time spent in the system is:

W = 1/S – A = 1/(80 – 75) = 1/5 = 12 minutes.

Given these initial calculations, you can then explore the implications of different queuing parameters for the queuing system. For example, if you decide that the average time spent in the queue can be as much as 15 minutes, then you can work back through the calculations to see what changes to the service time or number of customers you could make. Similarly, if you decided that the utilization was too high for staff to maintain, you could see the effects of lowering it to 80 per cent.

Monte Carlo simulation

When you solved the queuing model above, you used equations that you could solve exactly. In queuing, as we depart from the singleserver, single-phase model, or from looking at averages, the equations get more and more difficult.

Simulation is the process of reproducing the behaviour of a system using a model and manipulating certain variables to see their effect on the operation. Simulation models, unlike the models in Chapter S1, are descriptive rather than analytical.

Monte Carlo simulation allows us to simulate probabilistic events, such as the arrival of customers at the restaurant. The simplest form of Monte Carlo simulation is to determine what value a probabilistic variable will take by flipping a coin or other chance process. For example, the number of people arriving together in a party could be determined by rolling a die, and using that outcome to assign a value from 1 to 6.

A more common method of selecting values is using random numbers generated by either a random number table or a computer program. A cumulative probability distribution is used to convert the random number to a value.

Example

The cumulative probability distribution for a variable that is normally distributed with a mean of 5 and a standard deviation of 2 is shown in the second column of Table S2.1. For each of 10 trials, the random number in the fourth column is used to look up the number based on the cumulative probability distribution. Thus in the first trial the value of the random number is 0.862, which falls in the range for the cumulative probability distribution for at least 7 but less than 8, from 0.841 to 0.933.

Table S2.1 Monte Carlo simulation for a normal distribution with mean 5 and standard

Value	Cumulative probability	Trial	Random number	Value
1	0.022750	1	0.861997	7
2	0.066807	2	0.183346	3
3	0.158655	3	0.441703	4
4	0.308538	4	0.562898	5
5	0.500000	5	0.072431	2
6	0.691462	6	0.416745	4
7	0.841345	7	0.217615	3
8	0.933193	8	0.933714	8
9	0.977250	9	0.043755	1
10	0.993790	10	0.069527	2

Using a Monte Carlo approach, we can simulate the arrivals and service times for the noodle bar in a much more realistic fashion than in the queuing example before.

Example

Suppose that your employer would like to look at a more detailed picture of the proposed restaurant operation, down at the individual table level. The following assumptions have been made:

Each table can seat 20 people

The average number of people in a party is four, and is normally distributed with a standard deviation (SD) of 2

The inter-arrival time is exponentially distributed, with a mean of 6 minutes between arrivals

On average, each party is at the table for 35 minutes, which is normally distributed with a standard deviation of 10 minutes.

Table S2.2 shows the cumulative probabilities for the distributions above.

Your employer has asked you to determine whether these parameters are realistic, using Monte Carlo simulation. Using the random numbers in Table S2.2, simulate the arrival and departure times for the first 25 parties arriving in an evening.

Table S2.2. Cumulative probability distributions for example

Shift scheduling

Pleased with your calculations, your employer has now asked you to use shift scheduling to determine how many wait staff the noodle bar will need to hire for its initial opening. The number of people per day has been calculated, but given the tight job market and full employment in the local area, the restaurant will only get staff by offering positions with a 5-day week, including two consecutive days off.

The following technique was described by Chase et al. (2000) for scheduling workers in this way.

Step 1. Determine the number of workers needed per day.

Step 2. Copy the table in Step 1, and mark the two consecutive days with the lowest worker requirement (italicized in the table). This will give the first worker's schedule of days on and off. (Note that this might fall across Sunday and Monday, as in this example!). Then subtract 1 from each of the days that is not shaded.

Again, choose the two consecutive days off for the second worker by identifying the two adjacent days with the lowest worker requirement. If there are two identical pairs, as in this example, choose the pair with the lowest adjacent requirement. If there is still a tie, choose the first pair.

Keep repeating this procedure until all requirements have been filled.

Thus, even though the maximum number of workers on any one day is 12, it will take 14 workers to cover the week under the 5 days on and 2 days off requirement. Note that the thirteenth worker scheduled is required for only Monday to Thursday, although the worker is hired to work 5 days, and the fourteenth worker is only required for 1 day! Perhaps worker 13 could be contracted to work on Saturday instead of Friday, to minimize the total workforce required.

SUMMARY

This chapter has introduced you to a structured process for analysing service operations, and to some common models and tools that are useful for analysing them. Normann's service management system model provides a high-level model for analysing service operations in terms of five service elements and the relationship between them. You can focus further on the service delivery element, through applying the transformation model, to understand how operations can help satisfy customers and clients more effectively. Finally, because you will find queues being used to match capacity and demand in so many types of service operations, the mathematical approach to analysing queuing was introduced.

Key terms

Arrival rate

Back office

Customer contact

Core services

Front office

Monte Carlo simulation

Multiple channel

Multiple phase

Queuing theory

Service blueprinting

Service guarantee

Service operations

Service package

Service pokayoke

Service rate

Service time

Shift scheduling

Simulation

Single channel

Single phase

Utilization

References

Chase, R. B., Acquilano, N. J. and Jacobs, F. R.(2000). Operations Management for Competitive Advantage,9^th edn. McGraw-Hill.

Heskett, J. L. (ed.) (1986). The multinational development of service industries. In Managing in the Service Economy, pp. 135–52. Harvard Business School Press.

Normann, R. (1991). Service Management: Strategy and Leadership in Business, 2nd edn. John Wiley & Sons.

Sasser, W. E., Olsen, R. P. and Wyckoff, D. D. (1978). Management of Service Operations: Text, Cases and Readings. Allyn & Bacon. Shostack, G. L. (1984). Designing services that deliver. Har. Bus. Rev., Jan–Feb, 133–9.

Further reading

Fitzsimmons, J. A. and Fitzsimmons, M. J. (2001).Service Management: Operations, Strategy, and Information Technology, 3rd edn (see especially Chapter 11). McGraw Hill.

Voss, C., Blackmon, K., Chase, R. et al. (1997). Achieving World-Class Service. Severn–Trent/London Business School.

Zeithaml, V. A. and Bitner, M. J. (2000). Services Marketing: Integrating Customer Focus Across the Firm, 2nd edn. McGraw Hill.

Sites of interest

http://www.wagamama.com/

http://www.yosushi.co.uk/