Manufacturing strategy – understanding the fitness landscape


Published on

This theoretical paper presents, extends and integrates a number of systems and
evolutionary concepts, to demonstrate their relevance to manufacturing strategy formulation.
Specifically it concentrates on fitness landscape theory as an approach for visually mapping the
strategic options a manufacturing firm could pursue. It examines how this theory relates to
manufacturing competitiveness and strategy and proposes a definition and model of
manufacturing fitness. In accordance with fitness landscape theory, a complex systems
perspective is adopted to view manufacturing firms. It is argued that manufacturing firms are
a specific type of complex system ± a complex adaptive system ± and that by developing and
applying fitness landscape theory it is possible to create models to better understand and visualise
how to search and select various combinations of capabilities.

Published in: Business, Technology
  • Be the first to comment

Manufacturing strategy – understanding the fitness landscape

  1. 1. The Emerald Research Register for this journal is available at The current issue and full text archive of this journal is available at,2 Manufacturing strategy: understanding the tness124 landscape Ian P. McCarthy SFU Business, Simon Fraser University, Vancouver, Canada Keywords Dynamics, Manufacturing systems, Systems theory Abstract This theoretical paper presents, extends and integrates a number of systems and evolutionary concepts, to demonstrate their relevance to manufacturing strategy formulation. Speci cally it concentrates on tness landscape theory as an approach for visually mapping the strategic options a manufacturing rm could pursue. It examines how this theory relates to manufacturing competitiveness and strategy and proposes a de nition and model of manufacturing tness. In accordance with tness landscape theory, a complex systems perspective is adopted to view manufacturing rms. It is argued that manufacturing rms are a speci c type of complex system ± a complex adaptive system ± and that by developing and applying tness landscape theory it is possible to create models to better understand and visualise how to search and select various combinations of capabilities. Introduction In 1997, the UK Confederation of British Industry (CBI) produced a report titled Fit for the Future (Confederation of British Industry, 1997). This report was targeted at UK manufacturing and it claimed that an additional £60 billion a year could be added to the UK gross domestic product if manufacturing performance was raised to US manufacturing performance levels. The report stressed the need for UK improvement programmes based on cost-cutting, ef ciency gains and better product lead-times and quality. As a consequence of the popularity of this report, the concept of being ª tº was synonymous with being world-class. However, what do the terms ª tº and ª tnessº mean in the context of manufacturing and operations management? This question is the motivation and starting point of this paper. It is addressed by viewing manufacturing rms as complex adaptive systems and by developing and relating tness landscape theory to the process of manufacturing strategy formulation. These concepts provide an interpretation of tness that is consistent with the CBI view of t, but with a number of subtle and important differences that revolve around rm survival and imitation. This paper also develops and extends existing work that has judged tnessInternational Journal of Operations &Production Management landscape theory appropriate for understanding organisational developmentVol. 24 No. 2, 2004pp. 124-150 and rm dynamics (Kauffman and MacReady, 1995; Levinthal, 1996; Reuf,q Emerald Group Publishing Limited0144-3577 1997; Beinhocker, 1999; Barnett and Sorenson, 2002). These articles explain theDOI 10.1108/01443570410514858 value and relevance of tness landscape theory, but avoid de ning tness and
  2. 2. relating it to the performance and behaviour of rms. There is an implicit Manufacturingassumption that tness simply relates to competitiveness and effectiveness. strategy With this introduction, the contribution that this paper makes is to: . explore how a complex systems view can be used to understand manufacturing strategy and competitiveness; . create a de nition and conceptual model of manufacturing tness that 125 provides a basis for better understanding the exploration of strategic options; and . to consider the relevance of this model and tness landscape theory to the practice of manufacturing strategy.Manufacturing and complex systems theoryFitness landscape theory has its origins in complex systems research and, inparticular, the study of evolutionary properties in biological systems. Thus,before introducing tness landscape theory and its relevance to manufacturingstrategy, it is necessary to understand the term ªcomplex systemº. According to the Shorter Oxford Dictionary (Brown, 1993), the wordªsystemº rst appeared in 1619 and is now de ned as ªAn organised orconnected group of objects: a set or assemblage of things connected, associated,or interdependent so as to form a complex unityº. It is a ubiquitous term that isused to describe and consider many entities in our social, physical andbiological world. Kuhn (1962), Capra (1986) and McCarthy et al. (2000a) discuss and reviewseveral eras and movements of systems thinking, including: . the Aristotelian view (organic, living and spiritual); . the Cartesian view (mechanistic and reductionism); . the Newtonian view (principles of mechanics); . the romantic view (self-organizing wholes); . the general systems science view (elements and their relationship to the whole, and open systems versus closed systems); . the cybernetic view (feedback, self-balancing, self-regulating and self-organisation); . the soft systems view (mental constructs); and . the complex systems view (non-linearity, self-organisation and emergence).Despite the different stance each view has, a common and binding theme is thatthey are trying to understand complicated entities by: . determining the system boundary, components, inputs and outputs, relationships and attributes; and
  3. 3. IJOPM . supporting the integration of views and knowledge to study the total24,2 system and how it interacts with its environment. The complex systems view, also known as complex systems theory (Stacey, 1995; Anderson, 1999; Choi et al., 2001; Dooley and Van de Ven, 1999; Morel and Ramanujam, 1999), seeks to understand the interactions between the system126 elements and between the system whole and its environment. These interactions generate non-linearity, self-organisation and emergence, which are dif cult to represent and understand using a mechanistic and reductionism view. For example, the mechanistic and reductionism view would typically attempt to understand systems by reducing the whole system (e.g. the whole manufacturing rm) into manageable individual elements (e.g. manufacturing departments or other sub-units). By separating and studying these individual elements of the system, this view seeks to understand and formulate theories about the behaviour of the whole system, while the complex systems view asserts that the whole system cannot be truly understood by reducing it into smaller manageable units. This is because non-linearity, emergence and self-organisation are a product of the individual system element rules and behaviours, which are often independent of any rules that may have been imposed on the system as a whole. Thus, a key reason why manufacturing rms change is because they are complex and evolving systems in uenced by alterations in their environmental (internal or external) conditions. It is this ability to evolve that makes manufacturing strategy formulation necessary and dif cult (Rakotobe-Joel et al., 2002). To further understand complex systems behaviour and its relevance to manufacturing rms, a discussion on these three related characteristics is provided. Non-linearity This is a system characteristic in which an input or change in the system is not proportional to the output or effect. Thus, effect is rarely relative to cause, and what happens locally in one part of a system often does not apply to other parts of a system (Sterman, 2002). For instance, if a manager decides to add additional resource, (e.g. workers and machines) to a production plant, the result is not always a corresponding and linear increase in the number of products manufactured. As most managers know, if one system parameter is changed there are interactions between the system elements (workers, equipment, departments, etc.) that can produce an aggregate behaviour which could not be derived by adding up the individual element behaviours or interactions. Emergence This attribute of a complex system results from the system’s evolution and non-linearity. Literally, emergence means ªto dive outº or to come out of the
  4. 4. depths. Thus, emergence is the manifestation of new system performance due Manufacturingto the collective behaviour of the elements, as opposed to the individual strategybehaviour of each element. Efforts to understand organisations in termsformalization, differentiation and social adhesion cannot solely focus onindividual members of the rm (Lazarsfeld and Menzel, 1961). Emergentbehaviours are typically unanticipated and sometimes novel. For example, if amanager decides to discipline or dismiss an employee, then the unexpected and 127emergent result could be that the workforce goes on strike in protest and bringsthe business to a standstill. The phenomenon of system emergence is consistentwith Mintzberg’s view of emergent strategies (Mintzberg, 1978), where anunplanned and unpredicted event can materialise regardless of the plannedintention.Self-organisationVon Foerster (1960) de nes a self-organising system as the rate of increase oforder or regularity in a system. This de nition is also dependent on theobserver’s frame of reference. For example, as a manufacturing rm changesover time (e.g. more products, new technology, and new working practices) amanager should accordingly update and enlarge his understanding of thesystem and its possible states and behaviours. Self-organisation is also aproduct of the interactions, dependency and circularity of organisationalsystems and how they address and engage with the domains in which theyoperate. This leads to a range of dependent systems processes such asself-creation, self-production, self-maintenance and self-con guration, all ofwhich are consistent with the complex systems and cybernetic view of rmsand are known as autopoiesis ± the process by whereby a rm produces andmaintains itself (Maturana and Varela, 1980). Before considering the concept of tness and tness landscape theory it isimportant to recognize that the complex systems view considers some systemsto have elements (i.e. people) which have a decision-making capability(McCarthy, 2003). These elements are referred to as agents and their systemsare referred to as complex adaptive systems. Agents are able to receive andprocess information according to a set of goal directed operating rules (schema)that the system may have. This decision-making capability creates the internaldynamic of the system and permits system adaptation (Wooldridge andJennings, 1995). Thus, manufacturing rms are complex adaptive systems thatconsciously evolve and self-organise (adapt) in response to certain goals orobjectives.Introduction to tness landscape theoryThe origins of tness landscape theory are attributed to Sewall Wright (1932),who created some of the rst mathematical models of Darwinian evolution. Heobserved a link between a micro property of organisms (interactions between
  5. 5. IJOPM genes) and a macro property of evolutionary dynamics (a population of24,2 organisms can evolve multiple new ways of existing). To describe this epistasis (the effect of one variable on another) Wright proposed a tness landscape metaphor in which a population of organisms would evolve by moving towards a higher tness peak, i.e. from population A to population B as shown in Figure 1.128 More recently, tness landscape theory has been used to investigate a number of life science problems including the structure of molecular sequences (Lewontin, 1974) and mathematical models of genome evolution (Macken and Perelson, 1989). One speci c model, the NK model,was devised to examine the way that epistasis controls the ªruggednessº of an adaptive landscape (Kauffman and Weinberger, 1989; Weinberger, 1991; Kauffman, 1993). With this model, N represents the number of elements in a system and K represents the number of linkages each element has to other elements in the same system. This formal, but simple representation allows the model to be applied to other complex systems. For example, management and organisational science researchers have discussed and advocated the use of tness landscape theory for investigating: . organisational development and change (Beinhocker, 1999; McKelvey, 1999; Reuf, 1997); . the evolution of organisational structures (Levinthal, 1996); . innovation networks in the aircraft industry (Frenken, 2000); and . technology selection (McCarthy and Tan, 2000; McCarthy, 2003).Figure 1.Evolution as athree-dimensionallandscape
  6. 6. Despite the contributions made by these works, the questions of what exactly is Manufacturing tness and how does it relate to the studies in question are not fully addressed, strategyand in some cases are avoided.A review of tnessAlthough the term tness is used regularly in biological and evolutionary 129publications, its de nition and use is unclear. This ambiguity has beentransferred to those management and strategy papers that discuss therelevance and insights that tness landscape theory could offer to managementscholars. It seems that most authors assume there is a universally understoodmeaning of the term and therefore do not provide a working de nition. Thisproblem was identi ed by Stearns (1976), who observed that the term tnesshas not been de ned precisely, but that everyone seems to understand it. In anattempt to avoid repeating this problem, this paper presents a review andexplanation of the term tness, which will be the basis for the proposedde nition and model of manufacturing tness. The term tness was rst used by Herbert Spencer in 1864 in the context ofªsurvival of the ttestº and ªnatural selectionº as proposed by Darwin in hisOrigin of Species four years beforehand(Gould, 1991). In a later edition of the samebook, Darwin used the two phrases interchangeably, and later, it became widelyknown as ªDarwinian tnessº, which generally meant the capacity to survive andreproduce. It was not until 1930 that Fisher (1930) related tness to an organism’sreproduction rate, although he himself did not formally de ne tness. To better understand the biological meaning of tness and its relevance tomanufacturing strategy and survival, Table I presents a de nition of tnessand four related terms (Endler, 1986). Each de ntion is translated into amanufacturing context. The de nitions in Table I show that tness is traditionally de ned as therelative reproductive success of a system as measured by fecundity or other lifehistory parameters. Yet, it also indicates that tness is a measure of a system’sability to survive. Thus, we have two dimensions to tness: (1) survival tness, which is the capability to adapt and exist; and (2) reproductive tness, which is an ability to endure and produce similar systems.Manufacturing rms do not sexually reproduce, but those that compete bycreating new strategic con gurations often inspire others to imitate theirstrategy and mode of working. Thus, it is proposed that manufacturing tness isthe capability to survive by demonstrating adaptability and durability to thechanging environment. This involves identifying and realising appropriatestrategies, which in turn are perceived by competitors to be successful, who thenadopt the same strategy. This process is similar to the biological view thatconsiders tness to be an observable effect (i.e. the reproduction rate) and is also
  7. 7. IJOPM Context De nition and measurement Manufacturing relevance24,2 1. Fitness The average contribution to the A successful manufacturing strategy breeding population by an organism, will spawn a host of imitators who or a class of organisms, relative to seek the same bene ts the contributions of other organisms130 2. Rate Coef cient The rate at which the process of The rate at which manufacturing natural selection occurs. Measured rms will successfully adopt a new by the average contribution to the strategy gene pool of the following generation, by the carriers of a genotype, or by a class of genotypes, relative to the contributions of other genotypes 3. Adaptedness The degree to which an organism is A form of absolute tness, that able to live and reproduce in a given relates to the ability to survive (an set of environments; the state of internal factor) and a rm’s being adapted. Measured by the perceived competitiveness (an average absolute contribution to the external factor) breeding population by an organisms or a class of organisms 4. Adaptability The degree to which an organism or The internal process by which species can remain or become manufacturing rms survive in the adapted to a wide range of long-term. It is based on environments by physiological or self-organisation learning, genetic means innovation and adaptation 5. Durability The probability that a carrier of an The robustness and longevity of a allele or genotype, a class of manufacturing rm’s genotypes, or a species will leave competitivenessTable I. descendants after a given long periodThe ve contexts of tness Source: Adapted from Endler (1986, p. 40) consistent with the notion of rm effectiveness. For example, Seashore and Yuchtman (1967, p. 898) describe the effectiveness of a rm as ªits ability to exploit its environment in the acquisition of scarce and valued resourceº. Therefore, rms with high tness are able to adapt to survive. When faced with dif culties, they do not just dissipate, but nd ways to overcome circumstances, even if this means sacri cing short-term objectives. This view is supported by Katz and Kahn (1978), who assert that the behaviour of a rm simply revolves around the primary goal of survival, i.e. ªthe continuation of existence without being liquidated, dissolved or discontinuedº (Kay, 1997, p. 78). The strategic management view of tness is concerned with the balance between environmental expectations placed on the rm (costs, delivery, quality, innovation, customisation, etc.) with the resources and capabilities available in the rm. This is a process of matching environmental t and internal t (Hamel and Prahalad, 1994; Miller, 1992) and is consistent with the
  8. 8. theory of congruence, where each element of the rm ts with, reinforces, or is Manufacturingconsistent with, other elements (Nadler and Tushman, 1980). Although these strategyuses of the term ª tº were developed independently of tness landscape theory,they are consistent with the biological view of tness and the concept ofepistasis (the effect of one variable on another). At this stage, it is concluded that the tness of any complex adaptive systemis a measure of its ability to survive and produce offspring. Ultimately, the term 131 tness is used tautologically, because what exists must be t by de nition. Thekey issue for managers is to recognise that manufacturing strategy formulationand competition is a complex systems issue. Changes in one part of their rmcan sometimes lead to non-linear and disproportional outcomes in other areas.As will be discussed, these changes also affect the shape and membership ofthe tness landscape in which they reside.The NK modelThis section will explain how the NK model can be used to better understandstrategy formulation as complex adapting system of capabilities and torecognise the epistasis between capabilities and competing strategies. To begin with, the system of study is a manufacturing strategy as de ned indetail in the next section. It is analysed and coded as a string of elements (N)where each element is a capability. For any element i, there exist a number ofpossible states which can be coded using integers 0, 1, 2, 3, etc. The totalnumber of states for a capability is described as Ai. Each system (strategy) s isdescribed by the chosen states s1 s2 . . .sN and is part of an N-dimensionallandscape or design space (S). The K parameter in the NK model indicates thedegree of connectivity between the system elements (capabilities). It suggeststhat the presence of one capability may have an in uence on one or more of theother capabilities in a rm’s manufacturing strategy. To understand the signi cance of this design space to manufacturingstrategy formulation, a seminal example is adopted and conceptually modi edfrom Kauffman’s work (Kauffman, 1993, McCarthy, 2003). Table II shows theNK model notation and outlines its relevance to manufacturing strategy.Table III provides the data for the example, which has the followingparameters: N = 3 (three capabilities such as quality, exibility and cost); A= 2 (two possible states such as the presence (1) or absence (0) of a capability); and K= N2 1= 2 (each capability will affect the other two capabilities in the strategy).With these parameters the design space is AN = 23 , which provides eightpossible manufacturing strategies, each of which is allocated a random tness
  9. 9. IJOPM Notations Evolutionary biology Manufacturing strategy24,2 N The number of elements or The number of capabilities that constitute the genes of the evolving genotype. strategy and the resulting con guration. A gene can exist in different These could include: exibility, facility forms or states location, technology management, degree of132 standardisation, process structure, approach to quality, etc. K The amount of epistatic The amount of interconnectedness among the interactions capabilities. This creates trade-offs or (interconnectedness) among the accumulative dependencies elements or genes A The number of alleles (the Number of possible states a capability might alternative forms or states) that have. For instance, the quality capability could a gene may have. have four states: inspection, quality control, quality assurance, and total quality managementTable II. C Coupledness of the genotype The co-evolution of one strategy with itsNK model notation with other genotype competitors value between 0 and 1 (see Table III). A value close to 0 indicates poor tness, while a value close to 1 indicates good tness. In principle, the tness values can then be plotted as heights on a multidimensional landscape, where the peaks represent high tness and the valleys represent low tness. In Kauffman’s model, the tness function f (x), is the average of the tness contributions, fi (x), from each element i, and is written as: 1X N f (x) = f i (x) N i= 1 As N = 3, a three-dimensional wire frame cube can be used to represent the possible combinations and their relationship to each other (see Figure 2). Each System Element 1 Element 2 Element 3 Assigned random (strategy) (capability X) (capability Y) (capability Z) tness value 000 Absent Absent Absent 0.0 001 Absent Absent Present 0.1 010 Absent Present Absent 0.3 011 Absent Present Present 0.5 100 Present Absent Absent 0.4Table III. 101 Present Absent Present 0.7Manufacturing strategy 110 Present Present Absent 0.8as a three bit string 111 Present Present Present 0.6
  10. 10. Manufacturing strategy 133 Figure 2. A tness landscape: N = 3 and K = 2corner point of the cube represents a manufacturing strategy and itshypothetical tness value. Strategic change is assumed to be a process ofmoving from one strategy to another in search of an improved tness. This isknown as the ªadaptive walkº. If we arbitrarily select a point on the cube (e.g.point 011), there are three ªone-mutation neighboursº. These are points 010, 111and 001. If point 011 has an immediate neighbour strategy with a higher tnessvalue then it is possible that a manufacturing rm would evolve to this tterstrategy (point 111). The arrows on the lines of Figure 2 represent either anuphill walk towards a greater tness value, or a downhill walk to a smaller tness value. A ªlocal peakº is a strategy (e.g. point 101) from which there is no tter point to move to in the immediate neighbourhood. A ªglobal peakº is the ttest strategy (point 110) on the entire landscape. As this is a simple example consisting of three capabilities, it is relativelyeasy to visualise the space of strategic options using a wire frame cube. If theexample dealt with several capabilities, it then becomes harder to visualise thedesign space using a multi-dimensional cube. To overcome this problem aBoolean hypercube can be used to map the strategic design space. Figure 3illustrates the landscape of strategic options generated by four capabilities(cost, quality, exibility and delivery). The tness values shown in Figure 3 aretaken from the work of Tan (2001), who carried out an NK analysis of theManufacturing Excellence 2000 competition data in the UK. As with the Figure 2 example, Figure 3 uses a binary notation to representthe presence (1) or absence (0) of a capability. For example, strategy 0011indicates that the capabilities exibility and delivery are present, while thecapabilities cost and quality are absent. The base strategy 0000 is at the top ofthe diagram, while the maximum strategy 1111 is at the bottom of the diagram.
  11. 11. IJOPM24,2134Figure 3.A Boolean hypercube offour manufacturingcapabilities As a manufacturing rm’s strategy aggregates additional capabilities, it descends into the lower parts of the diagram. The assigned tness value for the various combinations of capabilities is represented by the bracketed gure. Lines are used to connect two immediate neighbours and the direction of the arrowhead indicates an increase in tness. The dotted lines represent the route from 0000 to 1111 that has the greatest gain in tness with each move. The dashed lines with double arrows indicate two neighbouring strategies with the same tness. When all the arrowheads are directed to a single strategy, this is considered an optimal strategy (either local or global). In Figure 3, there are two optimal points, 1101 and 1111, both with tness values of 0.67. The K and C parameters As mentioned in the previous section, the K parameter is an indicator of a system’s (a strategy’s) connectivity. It represents the epistatic interactions between each system element (capability) and can range from K = 0 to K = N 2 1. The former being the least complex system, where each element is independent from all other elements; and the latter being the most complex system; where each element is connected in some way to all other elements. For K = 0, the resultant landscape is relatively simple and smooth, except for one single global peak. This suggests that one single strategy dominates the
  12. 12. competitive landscape (see Figure 4). As K increases from 0 towards its Manufacturingmaximum of N 2 1, the tness landscape changes to an increasingly rugged, strategyuncorrelated, and multi-peaked landscape (see Figure 5). This level ofconnectivity indicates frustration in the system, because it can lead to manylocal tness maxima on the landscape. If the NK model is applied to the processof manufacturing strategy formulation it is assumed that the contribution of 135any capability to the overall tness of a manufacturing strategy depends on thestatus of that capability and its in uence on the status of the other capabilitiesin the strategy. Figure 4. Fitness landscape for K= 0 Figure 5. Fitness landscape for K= N2 1
  13. 13. IJOPM Kauffman’s NK model was originally a xed structure model, in that the24,2 system under study was not be in uenced by factors outside of its system boundary. In other words, it was a closed system in a static environment. In practice, this assumption is simplistic and invalid for complex systems. Therefore, Kauffman introduced a C parameter, to indicate coupledness between the system and other systems in the environment. Coupledness means136 that any system will not just depend on internal factors, but also the behaviour and performance of the systems in the same environment. This notion is central to competition, because if the tness of one rm’s manufacturing strategy is increased, it is almost certain to affect the tness of other rms’ manufacturing strategies. In summary, manufacturing rms are complex adaptive systems that aim to consciously evolve by seeking new strategic con gurations. Fitness landscape theory and the NK model offer an approach by which to map, quantify and visualise manufacturing strategy formulation as a search process that takes place within a design space of strategic possibilities, whose elements are different combinations of manufacturing capabilities. A de nition and model of manufacturing tness At this point, the paper has discussed the concept of manufacturing rms as complex adaptive systems. It has introduced tness landscape theory and the NK model, provided a review of the term tness and brie y examined the relevance of the NK model to manufacturing strategy. The following sections of this paper develop these discussions by providing a de nition and model of manufacturing tness. Whilst not presenting a systematic review as such (Tran eld et al., 2003), a relatively comprehensive review of manufacturing strategy is offered. A theory of evolution is then presented to help understand how manufacturing strategies and their capabilities evolve according to ªvariation, selection, retentionº and ªstruggleº. This theory provides the basis for the proposed de nition and model of manufacturing tness. The anatomy of a manufacturing strategy The previous sections view manufacturing strategy as a system of connected capabilities. Before providing a de nition of manufacturing tness, it is important to con rm and justify this view. Skinner (1969) proposed manufacturing strategy as a process to help rms de ne the manufacturing capabilities needed to support their corporate strategy. He argued that an appropriate manufacturing strategy could provide a competitive advantage in terms of cost, delivery, quality, innovation, exibility, etc. Since Skinner’s article, numerous other terms have been proposed by operations management researchers for describing capabilities. These include competitive priorities (Hayes and Wheelwright, 1984; Boyer,
  14. 14. 1998), order winner and quali ers (Hill, 1994), and competitive capabilities Manufacturing(Roth and Miller, 1992). strategy The eld of strategic management has also made important contributions tothe concept of rm capabilities, speci cally through work dealing with thedistinctive competences (Selznick, 1957) and resource-based perspectives(Penrose, 1959; Barney, 1991; Peteraf, 1993). To relate this and recent work to 137the anatomy of a manufacturing strategy and tness landscape theory, thispaper adopts and develops the dynamic capabilities view (Teece et al., 1997) byde ning the following terms: . Resources are the basic constituents of a manufacturing rm. They are the tangible assets such as labour and capital, and the intangible and tacit assets such as knowledge and experience. . Routines are the norms, rules, procedures, conventions, and technologies around which manufacturing rms are constructed and through which they operate (Levitt and March, 1988, p. 320). . Core competencies are created by developing and combining resources and routines. They in uence performance and de ne and differentiate a rm from its competitors (Prahalad and Hamel, 1990). . Capabilities are a collection of competencies (core or otherwise) that provide competitive advantage in terms of cost, delivery, quality, innovation, etc. (Skinner, 1969; Stalk et al., 1992). . Dynamic capabilities provide a manufacturing rm with the ability to integrate, build, and recon gure resources, routines and competencies that will create new capabilities and a competitive advantage (Teece and Pisano, 1994; Teece et al., 1997; Eisenhardt and Martin, 2000). . Con gurations are the resultant form or type of manufacturing rm. They are de ned by the collection of resources, routines and resulting competencies and capabilities (Miller, 1996).With these de nitions, capabilities are considered the basic elements of amanufacturing strategy, while a dynamic capability is the collective activitythrough which a manufacturing rm systematically generates and modi es itsresources and routines to improve tness (see Figure 6). Dynamic capabilitiesenable strategic choice and permit manufacturing rms to move from oneposition on the tness landscape to another by re-deploying resources(Lefebvre and Lefebvre, 1998). This process of resource deployment is achievedby the rm’s routines, which connect, manage and co-ordinate the resources ina particular fashion. The importance of routines to manufacturing rms is suchthat Tran eld and Smith (1998) outline how strategic regeneration andperformance improvement are underpinned by the routines found in amanufacturing rm. Thus, if competitive manufacturing rms inspire others toimitate their strategy and mode of working, then this is a process of
  15. 15. IJOPM24,2138Figure 6.The anatomy of amanufacturing strategy
  16. 16. organisational learning and evolution where routines become ªtransmitted Manufacturingthrough socialisation, education, imitation, professionalisation, staff strategymovement, mergers, and acquisitionsº (March, 1999, p. 76). The notion of interconnectedness (the K parameter) can be found inmanufacturing strategy. For instance, Skinner (1974) argued that it would bedif cult for a manufacturing rm to perform well if it adopted all capabilities,and that the rms should focus on a selection of capabilities only. This view 139implied that some form of trade-off or negative connectivity betweencapabilities was unavoidable (Corbett and Vanwassenhove, 1993; Mapes et al.,1997), while others argue that capabilities are positively connected and thatcertain capabilities must be in place before another can be adopted. Hence,capabilities can often reinforce each other, creating a strategy that is asequential, cumulative and dependent system (Ferdows and De Meyer, 1990).Understanding and managing this connectivity is dif cult, because strategyformulation attempts to serve an unpredictable environment and the processoften leads to emergent strategies (Mintzberg, 1978). Also, a major constraintfor strategy formulation is the inherent and incorrect assumption that thestrategic options available on the known landscape are xed. This assumptionis false, because the size and shape of the landscape, along with the de ningenvironment, is continuously changing. This creates new and unexploredniches for rms to discover or create. It is these territories that the rm shouldexplore to ensure that maximum bene ts are gained (Hamel and Prahalad,1989).Variation, selection, retention and struggleThese four processes underpin the evolution of a population of organisations(Campbell, 1969; Pfeffer, 1982; Aldrich, 1999). Though they will be presentedand discussed individually, it is important to note that they act simultaneouslyand are coupled to each other. Using these evolutionary concepts, this paper proposes Figure 7 as a modelof manufacturing tness. The model assumes that manufacturing strategyformulation involves populations of manufacturing con gurations respondingto and creating manufacturing systems around speci c socio-technicalcon gurations. It is important to note that the population concept assertsthat for the con gurations under study to follow an evolutionary pattern, theymust exist in populations. That is, they must be a group of similar entities,which co-exist on a particular area of the landscape (Allaby, 1999). Apopulation could be an industry or market sector, but is ultimately a collectionof con gurations grouped because they compete in and serve a commonenvironment. Thus, the boundaries of a population can often exceed that of asingle sector, and the criterion for membership is simply that a rm facessimilar evolutionary and competitive forces to other rms in the population(McCarthy et al., 2000b).
  17. 17. IJOPM24,2140Figure 7.Model of manufacturing tness
  18. 18. The following sections describe Figure 7 by explaining variation, selection, Manufacturingretention and struggle. strategyVariationThis process is consistent with the concept of dynamic capabilities, as itinvolves changing resources, routines, competencies and capabilities to create a 141new strategy and a resulting con guration. Variations can be either intentional(planned) or blind (unplanned). They are intentional when decision makers inthe rm deliberately seek new strategies and ways of competing. For instance, rms may have formal programs of experimentation and imitation such asbenchmarking, internal change agents, research and development, the hiring ofexternal consultants and innovation incentives for employees. Such programsare intentionally created to promote innovative activities that could change thecurrent con guration of a rm. Blind variation occurs when environmental orselection pressures govern the process of change. This includes trial and errorlearning, serendipity, mistakes, misunderstanding, surprises, idle curiosity andso forth. It can also take the form of new knowledge or experience introducedinto the rm by newly recruited employees.SelectionThis process eliminates certain variations. It is a ltering function that removesineffective strategies and their routines, competencies and capabilities. Theselection forces can be internal or external. For example, external selectionoccurs when customers request a certain management practice or an approachto quality, or when industry norms and regulations demand certainperformance standards. Internal selection refers to intra-organisational forcessuch as policy, group behaviours and culture. Such forces not only selectvariations, but also create a positive reinforcement of old innovations andpractices. The result is that manufacturing rms can sometimes carry on doingwhat they know best, and maintain their existing strategy rather thanexploring the landscape for alternatives.RetentionOnce variations have been selected, the process of retention preserves andduplicates the strategy. The strategy and its elements are replicated andrepeated, in a fashion that is consistent with the concept of tness and theability to reproduce. For example, the JIT practices that existed in the USsupermarket industry in the 1950s were positively selected by Japaneseautomotive rms, who then demonstrated the competitive value of thisapproach to other manufacturers, and this led to further selection and retentionof JIT con gurations across a wide range of industries. The retention processallows rms to capture value from existing routines that have proved or areperceived to be successful (Miner, 1994).
  19. 19. IJOPM Retention can occur at two levels, the organisational and the population24,2 level. Organisational retention occurs through the industrialisation and documentation of successful routines, and by existing personnel transferring knowledge about the routines to new personnel. Population level retention takes place by spreading new routines from one manufacturing rm to another. This can happen through personal contacts, or through observers, such as142 academics or consultants publishing successful new technologies or management practices. Retention is the process that promotes capabilities and routines that are perceived to be bene cial, because rms unlike biological systems, have the capacity to observe and imitate successful rms. Struggle Struggle occurs because the resources on offer to manufacturing rms are not unlimited. This process governs the other three evolutionary processes by fuelling or limiting their potential. For example, during the industrial revolution, raw material and energy were key resources, while the present need is for knowledge-based resources such as skilled workers, research partners and value adding suppliers. In new industries, the leading rms have ample gain and enjoy fast growth. As competition and volume in the industry grows, the resources become more limited, and failure rates increase. In summary, Figure 7 helps represent how manufacturing rms evolve strategies and con gurations to serve different environments or niches. It shows that variation, selection and struggle govern survival tness and that selection, retention, and struggle govern reproductive tness. To a degree, this is consistent with aspects of the institutional view of strategic evolution (Meyer, 1977; Scott and Meyer, 1994; Tran eld and Smith, 2002) which states that variations are introduced primarily by mimetic in uences; selection is due to business conformity (regulative and normative) and retention occurs through the diffusion of common understanding. Figure 7 is the basis for the following de nition of manufacturing tness: The capability to survive in one or more populations, and imitate and/or innovate combinations of capabilities, which will satisfy corporate objectives and market needs, and be desirable to competing rms. Conclusions So what is the signi cance of tness landscape theory and the NK model to the process of manufacturing strategy formulation? To address this question, this concluding section reviews the implications and relevance of these concepts under three headings. Central to each is the view that manufacturing strategy formulation is a combinatorial system design problem. It involves identifying the elements of the strategy and recognising that the connectivity between the
  20. 20. elements and the coupledness between competing strategies will in uence the Manufacturingtopology of the tness landscape. strategyThe Red Queen effectThe complex adaptive systems view asserts that manufacturing strategy is aconsciously evolving system of resources, routines, competencies andcapabilities, which co-evolves with similar competing strategies. Thus, any 143improvement in one manufacturing rm’s tness will provide a selectiveadvantage over that rm’s competitors. Thus, a tness increase by onemanufacturing rm will lead to a relative tness decrease in other competing rms. The result is that competing rms take steps to improve their strategyand maintain their relative tness. This process is central to the populationconcept and was termed the ªRed Queen effectº by the evolutionary biologistVan Valen (1973). The Red Queen refers to a character from Lewis Carroll’sThrough the Looking Glass, in which Alice comments, that although she isrunning, she does not appear to be moving. The Red Queen in the novelresponds that in a fast-moving world ªit takes all the running you can do, tokeep in the same place.º Thus, the Red Queen metaphor represents theco-evolutionary process where t manufacturing rms will increase selectionpressures, and those competing rms that survive by adapting and enduringwill be tter, which in turn creates a self-reinforcing loop of competition. For leaders of manufacturing rms, traditional strategic managementtheory and practice advocate avoiding the Red Queen effect by nding niche ormonopolistic positions on the tness landscape. However, isolation fromcompetition tends to be temporary, and as reported by Barnett and Sorenson(2002), it has a less-obvious downside, in that it deprives a rm of the engine ofdevelopment. This results in a trade-off in which those rms occupying safeplaces on the tness landscape eventually suffer over time as they fall behindthose who remain in the race.Appropriate system varietyThe ability to create new manufacturing strategies and resultingcon gurations is related to a manufacturing rm’s ability to understand andmanage its system of routines and resources. Fitness landscape theory anddynamic capability theory state that systems must recon gure themselves torespond to the challenges and opportunities posed by the environment. Thiscapability to create strategic variations is dependent on the system having avariety that matches the array of changes an environment may create (Ashby’slaw of requisite variety, Ashby (1970, p. 105)). In terms of innovation strategies, this notion is well known and hasdeveloped into principles such as the law of excess diversity (Allen, 2001) andthe rule of organisation slack (Nohria and Gulati, 1996). Both these principlesassert that the long-term survival of any system designed to innovate requiresmore internal variety than appears requisite at any time. Appropriate system
  21. 21. IJOPM variety facilitates exploratory behaviour (Bourgeois, 1981; Sharfman et al.,24,2 1988) and is a necessary attribute for tness and a dynamic capability. The implication of system variety for leaders of manufacturing rms is that they should recognise the connection and trade-off between system ef ciency and system adaptability. Any effort to reduce system diversity and increase system standardisation could restrict the potential for innovation. This is144 because the evolutionary process of variation (especially blind variation) requires excess system diversity to fuel evolutionary adaptation (David and Rothwell, 1996). This ability to create blind variations is linked to the talent of producing innovative strategies. This claim is supported by a study of successful rms by Collins and Porras (1997, p. 141), who concluded: In examining the history of visionary companies, we were struck by how often they made some of their best moves not by detailed strategic planning, but rather by experimentation, trial and error, opportunism and quite literally accident. What looks in hindsight like a brilliant strategy was often the residual result of opportunistic experimentation and purposeful accidents. Understanding and exploring the landscape Understanding the topology of a tness landscape can help the manufacturing rms address the three questions that underpin the strategy process: (1) What is our current position on the landscape? (Strategic analysis). (2) Where should we be on the landscape? (Strategic choice). (3) How will we get there? (Implementation). Figure 8 shows a highly rugged landscape with two manufacturing strategies, strategy A and strategy B. The route from strategy A to strategy B is represented by a dashed line. This route initially requires a downhill journey that is often accompanied by a reduction in rm performance, which related to the learning curve challenge and organisational disruption associated with the change. With this reduction in performance, a rm often stops the strategic change and returns to its original position on the landscape. Thus, for a manufacturing rm to successfully explore and achieve new strategies, it must recognise that: . this often involves the removal of one or more of the capabilities and de ning routines and resources that dictate its current strategy and position on the landscape; . even though the landscape is posited as being static, when any rm moves or makes a change, the topology of the landscape and associated performance will also change. Exploration of the landscape is a search activity and there are two basic search strategies. The rst is a local search that enables manufacturing rms to build
  22. 22. Manufacturing strategy 145 Figure 8. A route or adaptive walk between strategiesupon their current capabilities. It involves investigating those manufacturingstrategies in the immediate vicinity (the one-mutation neighbour strategies).The second search strategy is a long distance search, i.e. looking for strategiesbeyond the local area. This involves a relatively signi cant recon guration ofthe strategy and is likely to arise due to previous failure-induced searches(Tushman and Romanelli, 1985) or because of the innovative nature of the rm(Nelson and Winter, 1982). However, long distance searches rarely occur inreality (Cyert and March, 1963; Nelson and Winter, 1982), because the longerdistance, the less time ef cient and less cost ef cient the search becomes. Also, rms that already have a relatively t strategy are unlikely to risk a signi cantrecon guration. Studies, practice and history show that a rms’ currentstrategic con guration frequently constrains a rm’s dynamic capability toremain focused on those resources and routines which are current and familiarto the rm. Manufacturing strategy formulation can also involve multiple and constantsearches, as suggested by Beinhocker (1999). This approach has directrelevance to strategy formulation as a process of organisationalresource-investment choices or options (Bowman and Hurry, 1993). However,the capability to have options requires appropriate system variety.SummaryThis paper has reviewed, developed and synthesized a range of literature topresent a de nition and a conceptual model of manufacturing tness. It isbased on survival tness; the capability to adapt and exist, and reproductive tness; the ability to endure and produce similar systems. These two
  23. 23. IJOPM dimensions of tness are governed by the evolutionary forces ± variation,24,2 selection, retention and struggle. The de nition and model offer a starting point for further research on how factors such as landscape topology, population and rm dynamics, the type and number of searches, and the associated costs and time to search would affect manufacturing strategy formulation and the propositions and ideas146 presented. To progress this work it is necessary to conduct empirical studies that measure manufacturing tness as part of a longitudinal assessment of the changes within and between the manufacturing rms in a de ned population. This type of work would provide a quantitative analysis of the claim that rms occupying a global peak on a K = 0 landscape gain bene ts from this monopolistic position, but at the expense of maintaining and developing a dynamic capability. References Aldrich, H.E. (1999), Organizations Evolving, Sage Publications, London. Allaby, M. (1999), A Dictionary of Zoology, Oxford University Press, Oxford. Allen, P.A. (2001), ªA complex systems approach to learning in adaptive networksº, International Journal of Innovation Management, Vol. 5 No. 2, pp. 149-80. Anderson, P. (1999), ªComplexity theory and organization scienceº, Organization Science, Vol. 10 No. 3, pp. 216-32. Ashby, W.R. (1970), ªSelf-regulation and requisite varietyº, in Ashby, W.R. (Ed.), Introduction to Cybernetics, reprinted in Emery, F.E. (Ed.) (1970), Systems Thinking, Penguin Books, Harmondsworth, Wiley, New York, NY, pp. 105-24. Barnett, W.P. and Sorenson, O. (2002), ªThe Red Queen in organizational creation and developmentº, Industrial and Corporate Change, Vol. 11 No. 2, pp. 289-325. Barney, J.B. (1991), ªFirm resources and sustained competitive advantageº, Journal of Management, Vol. 17, pp. 99-120. Beinhocker, E.D. (1999), ªRobust adaptive strategiesº, Sloan Management Review, Vol. 40 No. 3, pp. 95-106. Bourgeois, L.J. (1981), ªOn the measurement of organizational slackº, Academy of Management Review, Vol. 6, pp. 29-39. Bowman, E.H. and Hurry, D. (1993), ªStrategy through the option lens: an integrated view of resource investments and the incremental-choice processº, Academy of Management Review, Vol. 1, pp. 760-82. Boyer, K.K. (1998), ªLongitudinal linkages between intended and realized operations strategiesº, International Journal of Operations & Production Management, Vol. 18 No. 4, pp. 356-73. Brown, L. (Ed.) (1993), The New Shorter Oxford English Dictionary on Historical Principles, Clarendon Press, Oxford. Campbell, D.T. (1969), ªVariation and selective retention in socio-cultural evolutionº, General Systems, Vol. 14, pp. 69-85. Capra, F. (1986), ªThe concept of paradigm and paradigm shiftº, Re-Vision, Vol. 9, pp. 11-12. Choi, T.Y., Dooley, K.J. and Rungtusanatham, M. (2001), ªSupply networks and complex adaptive systems: control versus emergenceº, Journal of Operations Management, Vol. 19, pp. 351-66.
  24. 24. Collins, J.C. and Porras, J.I. (1997), Built to Last: Successful Habits of Visionary Companies, Harper Manufacturing Business, New York, NY. strategyConfederation of British Industry (1997), Fit For The Future: How Competitive Is UK Manufacturing?, Confederation of British Industry, London.Corbett, C. and Vanwassenhove, L. (1993), ªTrade-offs ± what trade-offs ± competence and competitiveness in manufacturing strategyº, California Management Review, Vol. 35 No. 4, pp. 107-22. 147Cyert, R.M. and March, J.G. (1963), A Behavorial Theory of the Firm, Prentice-Hall, Englewood-Cliffs, NJ.David, P.A. and Rothwell, G.S. (1996), ªStandardization, diversity and learning: strategies for the co-evolution of technology and industrial capacityº, International Journal of Industrial Organization, Vol. 14 No. 2, pp. 181-201.Dooley, K. and Van de Ven, A. (1999), ªExplaining complex organizational dynamicsº, Organization Science, Vol. 10 No. 3, pp. 358-72.Eisenhardt, K.M. and Martin, J.A. (2000), ªDynamic capabilities: what are they?º, Strategic Management Journal, Vol. 21, pp. 1105-21.Endler, J.A. (1986), Natural Selection in The Wild, Princeton University Press, Oxford.Ferdows, K. and De Meyer, A. (1990), ªLasting improvements in manufacturing performance: in search of a new theoryº, Journal of Operations Management, Vol. 9 No. 2, pp. 168-84.Fisher, R.A. (1930), The Genetical Theory of Natural Selection, The Clarendon Press, Oxford.Frenken, K. (2000), ªA complexity approach to innovation networksº, Research Policy, Vol. 29, pp. 257-72.Gould, S.J. (1991), Ever Since Darwin: Re ections In Natural History, Penguin Books, London.Hamel, G. and Prahalad, C.K. (1989), ªStrategic intentº, Harvard Business Review, Vol. 67 No. 3, pp. 63-76.Hamel, G. and Prahalad, C.K. (1994), Competing for the Future, Harvard Business School Press, Boston, MA.Hayes, R.H. and Wheelwright, S.C. (1984), Restoring Our Competitive Edge: Competing Through Manufacturing, John Wiley & Sons, New York, NY.Hill, T. (1994), Manufacturing Strategy: Text And Cases, Macmillan Press, London.Katz, D. and Kahn, R.L. (1978), The Social Psychology of Organizations, John Wiley, New York, NY.Kauffman, S.A. (1993), The Origins of Order: Self Organization and Selection in Evolution, Oxford University Press, New York, NY.Kauffman, S.A. and MacReady, W. (1995), ªTechnological evolution and adaptive organizationsº, Complexity, Vol. 1 No. 2, pp. 26-43.Kauffman, S.A. and Weinberger, E.D. (1989), ªThe NK model of rugged tness landscapes and its application to maturation of the immune-responseº, Journal of Theoretical Biology, Vol. 141 No. 2, pp. 211-45.Kay, N.M. (1997), Pattern In Corporate Evolution, Oxford University Press, Oxford.Kuhn, T.S. (1962), The Structure of Scienti c Revolutions, University of Chicago Press, Chicago, IL.Lazarsfeld, P.F. and Menzel, H. (1961), ªOn the relation between individual and collective propertiesº, in Etzioni, A. (Ed.), Complex Organizations, Holt, Reinhart and Winston, New York, NY, pp. 422-40.
  25. 25. IJOPM Lefebvre, E. and Lefebvre, L.A. (1998), ªGlobal strategic benchmarking, critical capabilities and performance of aerospace subcontractorsº, Technovation, Vol. 18 No. 4, pp. 223-34.24,2 Levinthal, D. (1996), ªLearning and Schumpeterian dynamicsº, in Malerba, G.D. (Ed.), Organization and Strategy in The Evolution of The Enterprise, Macmillan Press Ltd, Basingstoke. Levitt, B. and March, J.G. (1988), ªOrganizational learningº, Annual Review of Sociology, Vol. 14,148 pp. 319-40. Lewontin, R.C. (1974), The Genetic Basis of Evolutionary Change, Columbia University Press, New York, NY. McCarthy, I.P. (2003), ªTechnology management ± a complex adaptive systems approachº, International Journal of Technology Management, Vol. 25 No. 8, pp. 728-45. McCarthy, I.P. and Tan, Y.K. (2000), ªManufacturing competitiveness and tness landscape theoryº, Journal of Materials Processing Technology, Vol. 107 No. 1-3, pp. 347-52. McCarthy, I.P., Frizelle, G. and Rakotobe-Joel, T. (2000a), ªComplex systems theory ± implications and promises for manufacturing organizationsº, International Journal of Technology Management, Vol. 2 No. 1-7, pp. 559-79. McCarthy, I.P., Leseure, M., Ridgway, K. and Fieller, N. (2000b), ªOrganisational diversity, evolution and cladistic classi cationsº, The International Journal of Management Science (OMEGA), Vol. 28, pp. 77-95. McKelvey, B. (1999), ªSelf-organization, complexity, catastrophe, and microstate models at the edge of chaosº, in Baum, J.A.C. and McKelvey, B. (Eds), Variations in Organization Science ± in Honor of Donald T. Campbell, Sage Publications, Thousand Oaks, CA, pp. 279-307. Macken, C.A. and Perelson, A.S. (1989), ªProtein evolution on rugged landscapesº, Proceedings of the National Academy of Sciences of the United States of America, Vol. 86 No. 16, pp. 6191-5. Mapes, J., New, C. and Szwejczewski, M. (1997), ªPerformance trade-offs in manufacturing plantsº, International Journal of Operations & Production Management, Vol. 17 No. 9-10, pp. 1020-33. March, J.G. (1999), The Pursuit of Organizational Intelligence, Blackwell, Oxford. Maturana, H. and Varela, F. (1980), ªAutopoiesis and cognition: the realization of the living Boston studiesº, in Cohen, R.S. and Marx, W.W. (Eds), Philosophy of Science, 42, D. Reidel Publishing Co., Dordecht. Meyer, J.W. (1977), ªThe effects of education as an institutionº, American Journal of Sociology, Vol. 83 No. 1, pp. 55-77. Miller, D. (1992), ªEnvironmental t versus internal tº, Organization Science, Vol. 3 No. 2, pp. 159-78. Miller, D. (1996), ªCon gurations revisitedº, Strategy Management Journal, Vol. 17, pp. 505-12. Miner, A. (1994), ªSeeking adaptive advantage: evolutionary theory and managerial actionº, in Baum, J.C. and Singh, J.V. (Eds), Evolutionary Dynamics of Organizations, Oxford University Press, Oxford. Mintzberg, H. (1978), ªPatterns in strategy formationº, Management Science, Vol. 24, pp. 934-48. Morel, B. and Ramanujam, R. (1999), ªThrough the looking glass of complexity: the dynamics of organizations as adaptive and evolving systems complexityº, Organization Science, Vol. 10 No. 3, pp. 278-93. Nadler, D.A. and Tushman, M.L. (1980), ªA model for diagnosing organizational behavior: applying the congruence perspectiveº, Organizational Dynamics, Vol. 9 No. 2, pp. 35-51.
  26. 26. Nelson, R.R. and Winter, S.G. (1982), An Evolutionary Theory of Economic Change, Harvard Manufacturing University Press, Cambridge. strategyNohria, N. and Gulati, R. (1996), ªIs slack good or bad for innovation?º, Academy of Management Journal, Vol. 39, pp. 1245-64.Penrose, E. (1959), The Theory of the Growth of the Firm, Basil Blackwell, Oxford.Peteraf, M. (1993), ªThe cornerstones of competitive advantage: a resource-based viewº, Strategic Management Journal, Vol. 14, pp. 179-91. 149Pfeffer, J. (1982), Organizations and Organization Theory, Pitman, Boston, MA.Prahalad, C.K. and Hamel, G. (1990), ªThe core competences of the corporationº, Harvard Business Review, Vol. 30, May-June, pp. 79-91.Rakotobe-Joel, T., McCarthy, I.P. and Tran eld, D. (2002), ªEliciting organisational cladistics through Q-analysis as a basis for the rational planning of change managementº, Journal ± Computational & Mathematical Organization Theory, Vol. 8 No. 4, pp. 337-64.Reuf, M. (1997), ªAssessing organizational tness on a dynamic landscape: an empirical test of the relative inertia thesisº, Strategic Management Journal, Vol. 18 No. 11, pp. 837-53.Roth, A.V. and Miller, J.G. (1992), ªSuccess factors in manufacturingº, Business Horizons, Vol. 35 No. 4, pp. 73-81.Scott, R.W. and Meyer, J.W. (1994), Institutional Environments and Organizations: Structural Complexity and Individualism, Sage, Thousand Oaks, CA.Seashore, S.E. and Yuchtman, E. (1967), ªFactorial analysis of organizational performanceº, Administrative Science Quarterly, Vol. 12, pp. 377-95.Selznick, P. (1957), Leadership in Administration, A Sociological Interpretation, Harper & Row, New York, NY.Sharfman, M.P., Wolf, G., Chase, R.B. and Tansik, D.A. (1988), ªAntecedents of organizational slackº, Academy of Management Review, Vol. 13, pp. 601-14.Skinner, W. (1969), ªManufacturing: missing link in corporate strategyº, Harvard Business Review, Vol. 47 No. 3, pp. 136-45.Skinner, W. (1974), ªThe focused factoryº, Harvard Business Review, Vol. 52 No. 3, pp. 113-21.Stacey, R.D. (1995), ªThe science of complexity: an alternative perspective for strategic changeº, Strategic Management Journal, Vol. 16, pp. 477-95.Stalk, G., Evans, P. and Shulman, L.E. (1992), ªCompeting on capabilities: the new rules of corporate strategyº, Harvard Business Review, March-April, pp. 57-69.Stearns, S.C. (1976), ªLife history tactics: review of the ideasº, Quarterly Review of Biology, Vol. 51 No. 1, pp. 3-47.Sterman, J.D. (2002), Business Dynamics: Systems Thinking and Modeling for a Complex World, McGraw-Hill, Irwin.Tan, Y.K. (2001), ªA tness landscape modelº, PhD thesis, University of Shef eld, Shef eld.Teece, D.J. and Pisano, G. (1994), ªThe dynamic capabilities of rms: an introductionº, Industrial and Corporate Change, Vol. 3, pp. 537-56.Teece, D.J., Pisano, G. and Shuen, A. (1997), ªDynamic capabilities and strategic managementº, Strategic Management Journal, Vol. 18 No. 7, pp. 509-33.Tran eld, D. and Smith, S. (1998), ªThe strategic regeneration of manufacturing by changing routinesº, International Journal of Operations & Production Management, Vol. 18 No. 2, pp. 114-29.
  27. 27. IJOPM Tran eld, D. and Smith, S. (2002), ªOrganizational designs for team workingº, International Journal of Operations & Production Management, Vol. 22 No. 5, pp. 471-9.24,2 Tran eld, D., Denyer, D. and Smart, P. (2003), ªTowards a methodology for developing evidence informed management knowledge by means of a systematic reviewº, British Journal of Management, Vol. 14 No. 3, pp. 207-22. Tushman, M. and Romanelli, E. (1985), ªOrganizational evolution: a metamorphism model of150 convergence and reorientationº, in Cummings, L. and Straw, B. (Eds), Research in Organizational Behavior, JAI Press, Greenwich, CT, Chapter 7, pp. 171-222. Van Valen, L. (1973), ªA new evolutionary lawº, Evolutionary Theory, Vol. 1, pp. 1-30. Von Foerster, H. (1960), ªOn self-organizing systems and their environmentsº, in Yovitts, M.C. and Cameron, S. (Eds), Self-Organizing Systems, Pergamon, New York, NY, pp. 31-50. Weinberger, E.D. (1991), ªLocal properties of Kauffman N-K model ± a tunably rugged energy landscapeº, Physical Review A, Vol. 44 No. 10, pp. 6399-413. Wooldridge, M. and Jennings, N.R. (1995), ªIntelligent agents: theory and practiceº, The Knowledge Engineering Review, Vol. 10 No. 2, pp. 115-52. Wright, S. (1932), ªThe roles of mutation, inbreeding, crossbreeding and selection in evolutionº, Proceedings of the Sixth International Congress of Genetics, pp. 356-66, reprinted in Wright, S. (1986), in Provine, W.B. (Ed.), Evolution: Selected Papers, University of Chicago Press, Chicago, IL, 161-71.