A framework for analysing manufacturing flexibility

A framework for analysing manufacturing flexibility - The journal of operation management

Purpose ~ The issue of manufacturing flexibility has been widely discussed in the literature One

major area of focus has been the development of taxonomies for flexibility This paper aims to review

the contributions m this area and to propose a new classification and a framework for analysing

flexibility in manufacturing companies

Design/methodology/approach ~ The study adopts a case study methodology approach The

framework proposed is used to analyse the implementation of flexibility in four UK manufacturing

plants in four major industria] sectors: electronics, process, household and general goods and food

Findings - From the empincal analysis, various enablers of flexibility are identified These are

classified into three broad sources of flexibility namely fundamental enablers, indirect enablers and

generic enablers as well as flexibility avoidance strategies referred to as flexibility evaders

Practical implications ~ The implication is that a mix of flexibility solutions rather than a single

solubon may be the most appropriate way for delivering flexibility in an organisation However, the

drivers of the need for flexibility have to be correctly identified in order to determine the best solutions

for delivering system flexibility

Originality/value - The development of a refined framework for analysing manufacturing

flexibility as well as the identification of various enablers of strategic flexibility are the major

contributions of this paper

Keywords Manufacturing industries, Flexibility, United Kingdom

Paper type General review

Introduction

Manufacturing Hexiihty has been heralded as a major competitive weapon for

manufacturing organisations operating in increasingly uncertain environments and

turbulent markets It has been argued that manufacturing flexibility has the capability

to provide organisations with the ability to change levels of production rapidly, to

develop new products more quickly and more frequently, and to respond more rapidly

to competitive threats The existing literature related to all aspects of flexibility is very

extensive and the survey reported by Sethi and Sethi (1990) lists more than 200

references Also, Beach ef af (2000a) carried out an extensive literature review on the

subject Manufacturing flexibility literature reveals that many of the new competitive

frontiers advocated by authors appear to depend on manufacturing flexibility For

instance, it Is argued (Bolwijn and Kumpe, 1990) that while, it is quite possible to be

flexible without being innovative, the reverse is not true: “you cannot be innovative

without being flexible” According to Roth (1996) strategic agility is achievable only

with competitive strength in a combined set of generic capabilities, of which flexibility

The author would hike to thank Professor Colin New (retired) of Cranfield School of Management

for his guidance during the course of carrying out this research project

www.emeraldinsight.com/01 44-3577 htm

Analysing manufacturing flexibility

973

Emerald

Intemational Journal of Operations & Production Management Vol 25 No 1G, 2005

pp 973-996

«© Emerald Group Publishing Limited

0141-3557 DOT 10.7 £08/014455705 1061 S482

on research in the area of manufacturing flexibility, the general understanding of the

Beach ef al (2000b, pp 11-12) report, “It is difficult to see a totally consistent view forming of the types or dimensions of flexibility from the literature Indeed, 1t might be that a more appropriate view is that a number of taxonomies is needed to describe flexibility for different purposes” Flexibility, it seems, still presents something of a conundrum, a paradoxical concept where authors cannot agree on answers to even the most basic questions, e.g What is flexibihty? When should a company strive for it?

How can it be measured? How can it be implemented? The general lack of clarity concerning the concept of flexibility is at least in part, due to the multi-faceted nature of the concept itself Because flexibility cuts across the entire orgamsation and academic literature, it has proved to be difficult to adequately conceptualise and understand, both for practitioners seeking guidance and for academics researching the topic The objectives of this paper are to propose an integrated framework for analysing flexibility and to investigate the applicability of this framework in manufacturing companies The first section critically reviews the literature on manufacturing flexibility taxonomies Next an integrated model or framework for analysing manufacturing flexibility is proposed The framework is then used to investigate the implementation of flexibility in four selected manufacturing plants

Literature review on taxonomies of flexibility Much of the existing literature on flexibility has concentrated on developing taxonomies of the concept However, these studies have produced little consensus over the years and seem to have introduced at least three levels of confusion into the theory framework

(1) Different names have been used to refer to the same type of flexibility

(2) Failure to differentiate between flexibility and change capability There is a real difference between the ability of a manufacturing system to flex across a range

of states such that it can return to its exact original state (flexibility) and the ability to alter state “permanently” from one configuration to another quickly and easily (change capability) This difference is directly linked to the ideas of range and response (Slack, 1987; Koste and Malhotra, 2000; D’Souza and Williams, 2000) and Upton’s (1995) operational, tactical and strategic time scales for flexibility

(3) Failure to differentiate between how flexibility is perceived externally by the market, how it is characterised at the manufacturing system level and the actual tools and techniques used to deliver it For example, one of Slack’s four system flexibility types, volume flexibility 1s perceived by the customers/market as the ability to deliver greatly varying levels of output without relaxing delivery requirements It can be achieved by manipulating the capacity of the manufacturing system (what Slack refers to as resource flexibility) for example,

customer demand levels manufacturin g

Manufacturing System” is often used (usually in its shortened form “FMS”) to

describe a very specific piece of technology used in the engineering industry to machine

technology are such that taxonomies developed in relation to such systems are not

easily transferable to manufacturing systems in general (Parker and Wirth, 1999)

Examples of studies on taxonomies of flexibility include Browne ef al (1984), Slack

(1983, 1987), Hyun and Ahn (1992), Suarez et al, (1996), Zhang et al (2003) and Koste

et al (2004), Except for Suarez et al, Zhang et al, and Koste et al.’s papers, these are all

discussed in detail in Beach et al (2000a), which offers a very useful integrating

framework

A number of theoretical studies have highlighted the multidimensional and

complex nature of the concept of flexibility However, one of the main weaknesses of

both reported empirical studies and analytical models of flexibility has been the

tendency to treat flexibility as an unidimensional concept Hence, as Beach ef al

(2000b) argued, useful and meaningful empirical research on the issue of flexibility

needs to treat the dimensions of flexibility separately within an integrative framework

The model that Beach ef al, (2000a) proposed is extremely useful in providing such a

framework However, in order to operationalise it for managers we would propose that

the model be refined to incorporate a more detailed treatment of the ideas of resource

and system flexibility first put forward by Slack (1983, 1987, 1991)

Slack (1991), which incorporates two highly influential earlier papers (1983 and

1987) offered a simple classification into resource flexibility and manufacturing system

flexibility Resource flexibility related to the actual tools and techniques used to deliver

flexible characteristics and included, for example, process technology, employment

terms, etc Manufacturing system flexibility included four identifiable types

(1) New product flexibility The ability to introduce and manufacture novel products

or to modify existing ones

(2) Mix flexibility The ability to change the range of products being made by the

manufacturing system within a given period

(3) Volume flexibility The ability to change the level of aggregated output

(4) Delivery flexibihty The ability to change planned or assumed delivery dates

The separation of the tools and techniques by which the flexibility is delivered

(resource flexibility) and the consequential system flexibility is an important

distinction, but it still does not link fully to the customer perceived benefits Slack

argued that resource flexibility contributes to or, in a sense, enables manufacturing

system flexibility In turn, this improves manufacturing performance and

subsequently increases competitiveness

The work of Suarez ef al (1996) supported and developed Slack’s model, although

their paper did not directly acknowledge the source of their four categories They refer

to Slack’s four proposed categories as first-order flexibility types, and argue that they

are the flexibility types that directly affect the competitive position of a firm These

authors point out that other types of flexibility proposed in the literature such as

as routeing flexibility, usually defined as the ability to route production through different production paths It is immaterial to the customer how much routeing flexibility a firm might have However, possessing such flexibility might enable the firm to deliver more quickly (Suarez et al, 1996) It 1s this ability to deliver more quickly that is visible to the market place, not the routeing flexibility that makes 1t possible Many more research studies have been carried out on lower-order flexibility types than on first-order flexibility despite the competitive importance of the latter

One obvious explanation for this is that research into lower-order flexibility can largely

be carried out entirely within the business as a technical issue Research into first-order flexibility necessarily requires an external perspective

Some researchers, for example, Cox (1989), New (1996), Oke (2003a, b) and Zhang

et al (2003) have argued that, in competitive terms, volume and mix flexibility are the two most important manufacturing flexibility types Delivery and new product flexibility are seen as subordinate in that they are simply consequences of a plant’s capability in terms of volume and mix flexibility, respectively We find merit in the argument relating to delivery flexibility but would argue that there is a specific need to account for new product flexibility as a fundamental type Consider Slack’s (1991) full definition of volume flexibility, which incorporates the response dimension of time to adjust production levels and the cost implications of changing aggregate production volumes The implication of this definition is surely that a plant that cannot deliver varying volumes of orders to its customers when the orders are required cannot be said

to possess volume flexibility at the operational level, even if it is able to meet the volume requirements at a later date (Oke, 1999) In Slack’s terms, the plant would have range (long term) capability but not response (short term) capability This definition of volume flexibility would, therefore, seem to automatically incorporate delivery time

flexibility types

flexibility In other words, a plant that possesses volume flexibility is, by definition,

capable of delivery time flexibility The critical issue would seem to be that volume

flexibility is not an end in itself but only a means of creating competitive advantage

and what the customer sees is not the volume flexibility itself but its consequences in

terms of delivery capability Indeed, Slack (1987, p 12) comments “ volume and

delivery flexibility seemed to be interchangeable to some extent Process and mass

producers were more concerned with volume flexibility whereas batch and especially

jobbing companies were more concerned with delivery flexibility” It is suggested that

this simply reflects the different perspectives of the businesses Make-for-stock

businesses see the issue as one of volume and make-to-order businesses see it as an

issue related to the delivery of specific orders

Slack’s original definition of product flexibility included reference to novel products

or modifications of existing products, but it is clearer to use the extended term, new

product flexibility since, in normal terminology, there is the potential for confusion

between product and mix flexibility Similar to the case argued above, it can be argued

that new product flexibility is an extreme form of mix flexibility in that manufacturing

a novel product simply expands the mix That is, new product flexibility exists when

there is extensive range capability related to mix flexibility However, from a practical

point of view it would seem to be important to maintain a clear distinction between the

capability of a manufacturing plant to chop and change between the existing products

easily and quickly and the capability to bring innovative products to market easily and

cheaply This is particularly true where this must be done without detriment to

production of the existing mix Thus, we would argue that there are in fact only three

forms of manufacturing system flexibility: new product flexibility, mix flexibility and

volume flexibility

An integrated classification of manufacturing flexibility

We would argue that all the flexibility mechanisms considered in the literature can in

fact be encompassed within the three categories of manufacturing system flexibility

defined above Furthermore, the lower-order flexibility types that contribute to system

flexibility can also be separated into three sub-categories: those that are fundamental to

a single element of system flexibility, those that contribute to two elements and those

that can contribute to all three This relationship is shown in Figure 2

Analysing manufacturing flexibility

977

Figure 2 Manufacturing system flexibility and contributing factors

JOPM Suarez et al (1996) argued that the factors that affect volume flexibility in

for internal mechanisms such as the use of overtime (volume flexibility) and rapid changeovers (mix flexibility), it would not seem to be the case for an external item such

as supply chain management This is clearly an essential factor in the ability to flex

technology may also be used to flex either mtx (through for example, software programming) or volume (through for example, unmanned operation)

Fundamental and shared factors The fundamental determining factor of new product system flexibility is the new product development (NPD) process itself By this, we mean the composite of the practices and procedures by which new products are developed and handed over to the manufacturing function For example, the way in which manufacturing process development is done alongside the product development so that appropriate production tooling or instructions are available at the time of product introduction

Similarly, the way in which new products may be put through pre-production processes prior to introduction facilitates this

For many manufacturing systems, the main cause of the loss of efficiency inherent

in achieving mix flexibility is set-up time Thus, New (1996) argues that a steel plant that is capable of producing, say, 10,000 tonnes per week if tt made the same product continuously, but only 5,000 tonnes per week if it had to switch between 10 and 20 different products — would be thought of as having very low mix flexibility He further states that the plant would, however, be thought to have comparatively high mix flexibility if its output could always be in the range 9,800-10,200 tonnes per week irrespective of the mix of products required The fundamental determining factor of volume flexibility is the type of employment term employed This relates to the type

of labour capacity solutions that are used to achieve volume flexibility The skill level

of workers is a factor for achieving both mix and volume flexibility For instance, a multiskilled workforce would be able to perform multiple tasks driven by changes in mix and volume requirements

Generic factors Generic factors will enable the achievement of all the flexibility types However, the level of the impact of each of these factors on system flexibility may vary and may be dependent on the industry or the plant concerned The generic factors identified in the model in Figure 2 are supply chains, process technology and information technology

Supply chain management extends the concept of functional integration beyond the firm to all firms in the supply chain (Eilram and Cooper, 1990) Supply chain management seeks to enhance competitive performance by closely integrating the internal functions within a company and effectively linking them with the external operations of suppliers and channe! members (Vickery ef a@/, 1999) Jack and Raturi (2002) suggest that effective supply chain management allows a firm to efficiently use its network of suppliers and vendors to respond to uncertainties in demand However, in their investigation of the sources of volume flexibility, they find no significant relationship between external sources of volume flexibility and a firm's volume flexibility capability

FMS, flexibility levels have, in fact, decreased when compared with the manual

systems that the FMS replaced due in part to a lack of understanding of the complex

(1986, p 69) concluded that “ they (US organisations) are buying the hardware of

flexible automation — but they are using it very poorly Rather than narrowing the

competitive gap with Japan, the technology of automation is widening it further With

few exceptions, the FMS installed in the United States show an astonishing lack of

flexibility In many cases, they perform worse than the conventional technology they

replace The technology itself is not to blame; it is management that makes the

difference Compared with Japanese systems, those in the US plants produce an

order-of-magnitude less variety of parts”

As the above studies suggest, there is no doubt about the capability of process

technology itself to provide flexibility but the management of this technology is

key if the objective of achieving flexibility is to be realised

Closs ef al, (1997) argue that information technology is one of the few productivity

tools that is both capable of increasing productivity through better planning and

scheduling and capable of decreasing cost simultaneously Ody and Newman (1991)

suggest that information technology can be seen as a way of reducing lead times for

selling, reordering and delivery cycle time A type of information technology that is

commonly being used by organisations is the electronic data interchange (EDI)

technology Jenkins (1994) defines EDI as “the direct computer-to-computer

communication of inter-company and intra-company business documents in a

machine readable standard format” Hill and Swenson (1994) note that there has been

a significant growth in the use of EDI in recent years According to Ricks (1997), this

is probably due to a general drive towards just-in-time manufacturing and the need

to be flexible With an information technology such as an EDI, it is possible for an

organisation to achieve high levels of flexibility through the acceleration of the

ordering process due to the speed and accuracy of data flow (Carter and Ragatz,

1991) McKinnon (1990) also suggests that an information technology leads to

shortening of order lead times and the improvement of the responsiveness of the

ordering system further back along the supply chain —- resulting in improved

flexibility

Having proposed an integrated classification of manufacturing flexibility (Figure 2),

the objective of the empirical phase of the study was to investigate the applicability of

the proposed framework in manufacturing plants

Research methodology

Because the research was exploratory, a field-based data collection method, which gave

the depth necessary to fully explore the sources of flexibility in manufacturing plants,

was preferred Yin (1994) argues that case studies are more suitable for answering

“how” or “why” questions compared with other research strategies Thus, the case

study method seemed appropriate to explore how manufacturing plants achieve

flexibility within the context of the theoretical framework developed in this study

beneficial for investigating new research areas (Eisenhardt, 1989) are most applicable

Croom (1996) argues that a case study is in effect a detailed description of an organisation, incident or phenomenon, a hybrid form of research method Contrary to

of its qualitative nature, Eisenhardt (1989) noted that a case study could actually contain both quantitative and qualitative data The use of case study methodology is not very common in operations management research Pannirselvan et ad (1999) reports that case study and field study research accounted for 4.94 and 3.84 per cent, respectively, of published papers Because of the level of “richness” that a case study-based research can provide, there have been calls for more research studies based on case study methodology in operations management, notably by Meredith and Vineyard (1993), Meredith (1998) and Lewis (1998) Voss ef a/ (2002) argue that in spite

of several challenges that researchers face in conducting case research, the results of case research can have very high impact

Sample selection Because different types of sources of flexibility were to be explored, a mutli-case study research method in four manufacturing plants was adopted These companies were selected from a sample of UK manufacturing plants that had responded to a previous survey research on manufacturing flexibility and had reported employing different flexibility solutions The companies belonged to different sectors, their perceived needs for flexibility as well as their perceived levels of flexibility differed (Table I) and they used different mechanisms to achieve manufacturing system flexibility Thus, a purposeful (Stake, 1995) and a theoretical sampling approach was adopted in which cases that differ as widely as possible from each other are chosen (Stuart e/ al, 2002)

Multiple cases’ evidence (especially if carried out in different sectors) 1s often more compelling and gives the overall study more robustness (Herriot and Firestone, 1983)

Also, Eisenhardt (1989) notes that multiple cases allow findings to be replicated, and that “while there is no ideal number of cases” below four it can prove difficult to generate theory with much complexity and with convincing empirical grounding (Beach ef al, 2001) The description of the manufacturing plants 1s given in Table I

Data collection

An interview protocol that was based on the literature review on manufacturing flexibility was developed before the commencement of the field research (Eisenhardt, 1989) The protocol comprised questions on the need for (or drivers of) flexibility in the manufacturing plants and what solutions were being employed to deliver different types of flexibility This protocol was updated and improved with each return visit

Data collection was done through interviews and collection of relevant data (both quantitative and qualitative), observations, plant tours in each manufacturing plant and a workshop in each of the plant with all the plants’ interviewees The format of the interview was semi-structured and each interview lasted between one and two hours, The primary researcher was accompanied on the visits by a second researcher who gathered data independently At least two key informants (typically manufacturing

and bagged systems: PCB for the food industry: materials for the

product/system test conversion markets

Production Production — 110 Production — 650 Production — 80

Design — 2

requirements

Perceived levels of New New product — high New product — high New product — low

Mix — low Volume — high

manager and/or planning manager) were interviewed face to face in each

manufacturing plant The tours of the facilities provided the opportunities to talk to

people on the shop floor and to observe some factors that had been discussed during

the interviews such as the use of multi-skilled workers and some other strategies

employed to achieve flexibility

The data collection strategy employed helped to control for the construct validity of

the research Collecting data from multiple sources, ie interviewing multiple

respondents and comparing respondents’ comments with copies of documentation

where available helped control for the construct validity For instance, as it will be

discussed later (Table II) the analysis of archival data helped to validate the interview

analysis by giving an indication of the levels of flexibility of the plants By confirming

and testing the findings of each case in subsequent cases, the external validity of the

research was controlled for but admittedly to a limited extent However, this is not a

major issue because of the exploratory nature of the study The case study protocol

helped to ensure the reliability of the research and a database was developed

comprising all the documents, extensive notes and transcripts of the interviews This

also helped to ensure reliability as it enables an easy auditable trail of events

Data analysis

The unit of analysis adopted for the research was “a manufacturing plant” as defined

by New and Szwejczewski (1995): “a relatively self-contained unit with its own

management staff which can be identified by separate facilities, by separate products,

or by a separate management structure”

researcher and the second researcher who accompanied the primary researcher on the

visits The resultant report was initially sent to the key informants for review to ensure

with the respondents to feedback the outcome of the field work and to address the

areas where there were disagreements in responses to reach a consensus Emerging

themes for exploration and explanation were identified from the interview and data

analyses and the relationships between the identified variables were explored and

defined in subsequent interviews (Miles and Huberman, 1994) Also, a cross-case

analysis was done by selecting the themes identified from the individual case analysis

and checking for similarities and differences across the cases Data reduction was done

through categorization Solutions that interviewees’ identified as sources of flexibility

in their plants were categorized based on the framework shown in Figure 2 It was

possible to identify which category of flexibility sources was being identified and

which flexibility types the sources were linked with

Summary of results and discussions

Table II shows the estimate of the levels of flexibility in the plants based on the

analysis of the archival and relevant documents collected

Note By our definitions, a plant that had a high average percentage of orders

delivered on time and minimum impacts on unit labour cost and quality defect levels

would be seen to have a high level of system flexibility Where the high delivery

performance is related to a high production volume fluctuation and a high average

SKU, this would be an indication of high volume and mix flexibility, respectively,

provided that these were associated with low unit labour cost and quality defect levels

(Suarez et al., 1996)

It is interesting to see from Table II that all the plants would have been perceived by

their customers (or externally) to be highly flexible given their respective high delivery

performance, This delivery performance is comparable to the performance of top 25 per

cent of UK plants in related sectors (Szwejczewski ef al, 2002) However, an

investigation into how this was achieved reveals a different story The analysis of the

archival data shows that Plant D was not truly internally volume flexible Although,

the plant appeared to have a high volume flexibility given by its high production

fluctuation ratio, it appeared to achieve this with a high overtime rate which by our

definition means that the plant did not have true volume flexibility The results of a

further investigation into how the plants achieved their perceived levels of flexibility

are discussed next

Achieving system flexibility - fundamental and shared factors

Table III shows the system flexibility drivers, solutions as well as the factors for

achieving them tn the plants studied

It will be recalled from Figure 2 that there are three corresponding fundamental

factors for achieving new product, mix and volume flexibility These are the NPD

process, changeover times and employment terms, respectively