The Biography Of Artefacts Framework

Draft of Chapter Three from: Neil Pollock and Robin Williams (2009) Software and Organizations: The Biography of the Enterprise Solution Or How SAP Conquered the World, London, Routledge.

Chapter Three: The Biography of Artefacts Framework Here we describe the approach that we will adopt to study these various technologies. Building on earlier studies that have examined the mutual adaptation of technology and organisation, we develop a framework for investigating the ‘biography’ of software systems. Drawing on work from Science and Technology Studies, Material Culture and Cultural History, amongst others, we suggest an approach that follows the actual packages themselves as they evolve and mature, progress along their lifecycle, and move across sectoral and organisational boundaries. In this endeavour we address multiple timeframes and locales.

HOW MODES OF RESEARCH FRAME THE ANALYSIS STS has from the earliest days been concerned to resolve the question of adequate models for the analysis of technological innovation and associated societal change – as these frame the analysis and guide the methodology adopted and thereby what it is we can and cannot find out. This project in particular seeks to apply and further develop the biography of technology perspective, which emerged from our earlier work on organisational technologies (Brady et al. 1992; Clausen and Williams 1997; Pollock et al. 2003; Pollock and Cornford 2004). Our aim is to build a comprehensive understanding of the evolution of a technology – encompassing both technology design and implementation/use - and how it is shaped by its specific historical context across multiple social locales.

This research project was designed to exploit the opportunity to achieve insights from a longitudinal contemporaneous study, building upon our earlier research, including studies of Computer-Aided Production Management (CAPM) and other integrated automation systems in the late 1980s and early 1990s that were the progenitors of ERP and other ‘enterprise’ software packages of today (Fleck et al. 1990; Webster and Williams 1993; Fleck 1993; Clausen and Williams 1997). Underpinning this endeavour was an attempt to develop modes of enquiry that might be adequate to

1

explore these complex technologies in terms of both the design of empirical research and the conceptual tools advanced to explore them.

The project was inspired by our unhappiness with the way that most existing research into packaged software and ERP in particular was framed. As we noted in the previous chapters, there is a huge literature addressing ERP and the development and adoption of workplace technologies more generally. This research is often weaker in theoretical and conceptual terms than STS (which only constitutes a small share of this literature), particularly in its understanding of innovation, and is less concerned to consider issues of methodology and epistemology. The bulk of these studies are framed, somewhat unreflexively, within particular well-established modes of research, constrained within particular loci, timeframes, disciplinary perspectives and concerns. Our contention is that the framing of these studies can produce unhelpful readings about the character and implications of these technologies. Moreover, as Grabot and Botta-Genoulaz (2005) observe, particular types of research (e.g. quantitative survey, qualitative case studies) give salience to certain kinds of issues. We need to be in a position to reflect upon the implications of research choices for the outcomes of an investigation.

In this chapter, we first advance a critique of the dominant modes of enquiry into ERP and similar workplace technologies. We then go on to develop a research framework – based on our Biography of Artefacts Framework – that will be more adequate to the task. Here we draw upon research within our own tradition of STS, and also on related work from Organisational Studies and Information Systems perspectives which shares some of our presumptions and concerns (and we note that there is not always a clear dividing line between these disciplines, particularly in relation to work informed by constructivist insights, much of which has been influenced by STS perspectives).

We want to explore how some of the general debates within STS, outlined at the end of the previous chapter, relate to the concept of the Biography of Artefacts, about what would constitute an adequate analytical framework, equipped to address the multiple interfaces between technological artefacts and society. Studies of particular socially/temporally bounded locales, for example, the typical ‘ERP implementation’ 2

case study or survey are, we contend, ill-equipped to get to grips with these complex technologies which are instantiated at multiple sites (Clausen and Williams 1997; Kallinikos 2004b). Koch (2007) suggests that we need better spatial metaphors for addressing such objects. He draws attention to the evolution of perspectives, moving away from single site studies to multi-locale studies, and has further advanced the suggestion that we should analyse ERP as a ‘community’ (Koch 2005, 2007).1 These suggestions are thoroughly congruent with the broader social learning perspective we outlined in the previous chapter (Sørenson 1996; Williams et al. 2006). The task of this chapter is to chart out an analytical framework for such an endeavour.

Existing studies of technology and work organisation have in general paid inadequate attention to these kinds of debates (which is not to overlook outstanding exceptions to this generalisation and the increasing strength of social science research in Business Schools for example). There is a large amount of unreflexive research, including the many ‘impact studies’, addressing the organisational consequences of particular technologies, which pay little attention to questions of research design and theory. Amongst academic research, disciplinary divides have served to separate those studying technology supply from its adoption/use and technical from organisational issues.

The success of the organisational case study as a research model within Business Studies (as well as Information Systems and Technology Studies research) valuably focuses attention on local negotiations and choices around the design or the implementation and use of new technologies. Studies of technology and work have benefited greatly from the growing influence of interactionist perspectives (inspired for example by the exciting work of authors such as Lucy Suchman [1987]) and an associated enthusiasm for local ethnographic studies, which has been very effective as a research methodology in producing a rich local picture. However, this emphasis on local processes and actors may be at the expense of paying less attention to more generalised and long-term shaping processes.

Our work however seeks to explore how these local struggles are taking place within broader circuits of knowledge and influence including economic and social structures and material structures (and we suggest that a study of technology needs to engage 3

with technology as a materialised institutional form) which mobilise beliefs and visions and provide various incentives, resources and penalties and which thus set the parameters in which local actions are played out. We are also keen to explore how local outcomes may react back on and transform the broader setting, through diffuse and gradual processes of influence, which may not readily be detectable within shortterm local studies (Williams 1997).

Within the study of technology, the SST perspective has been marked by its insistence on the need to pay attention:

i.

to the specific material characteristics of technological artefacts and systems; and,

ii.

to the influence of social structure and history which pattern innovation, and which explain the patterns of uneven access to resources and sites of influence.

These considerations inform our search for a research strategy that addresses the multiple locations and different timeframes in which technologies operate. To this end, we examine the utility of various conceptualisations of ‘arena’ (Fleck 1988b; Jorgensen and Sørensen 1999), to explore the hybrid spaces in which different actorworlds interact, and of an ‘agora’ of technological and organisational change (Kaniadakis 2006), which provides a framework for looking at the relationship between different arenas and levels, and how local actions are set within broader settings. The agora is conceived in a relational sense; it is a complex space captured through the viewpoints that different actors (and analysts) make of this. Within this framework, we may wish to focus upon local, immediate settings of action or more widely dispersed institutionalised patterns. And this endeavour also suggests that we need to provide a register of the multiple different historical timeframes at play: from the immediate moment of action to the long term in which institutions emerge and evolve.

Before we turn to these questions, we shall first review the different kinds of research that have been carried into technology and work organisation and in particular ERP. 4

We wish to explore how these characteristic modes of empirical study impinge upon the framing of the research and on their findings.

EXISTING STUDIES OF TECHNOLOGY AND WORK ORGANISATION AND THEIR SHORTCOMINGS Snapshot Studies When a new technology comes to the fore, many of the first papers appearing are from the trade press and practitioner journals. The concern is to demonstrate the benefits of the technology and how these benefits may be successfully achieved. As a result the focus is typically on what Botta-Genoulaz et al. (2005: 574) describe as ‘impact factors’. Moreover, impact studies are also often what might be called ‘snapshot studies’. This is research that typically describes the organisation before and after the implementation of the particular innovation in question.2 There is a tacit technological (or other) determinism embedded in the research design and temporal/organisational framing of such enquiries, in which observed changes are attributed to the effects (‘impacts’) of the technology, apparently disconnected, for example, from the prior organisational setting which motivated and shaped the innovation as well as the subsequent organisational processes through which the outcomes were achieved.

Importantly, snapshot studies are often conducted a relatively short time after the introduction of a new technology, arguably before the complete consequences of an innovation can be reasonably assessed. This is problematic because as we have already seen, it can take years before the full benefits and costs of major changes such as ERP or Business Process Redesign materialise. As a result, researchers may be forced to anticipate projected benefits/costs, and therefore run the risk of being unduly influenced by the programmatic goals and visions of the promoters and suppliers of a technology, as well as the hopes of the managers pushing these changes through within their own organisations.

As discussed in Chapter Two, these kinds of studies are important in developing and circulating ideas about an emerging technology. They thus contribute to the rhetorics of technology supply and often also suffer from its shortcomings. They are however influential in framing understandings – and many of these studies are geared towards 5

influence rather than analytical accuracy and distance. Even though they might provide interesting information and insights, snapshot studies are typically only of limited analytic value and the data and claims they contain need to be interpreted with caution.

Implementation studies We include under the heading of ‘implementation studies’ the large body of writings concerned with the introduction of new technologies into organisational settings, (whether from sub-disciplines within Business Studies, such as the Management of Technology and Innovation, or from Technology Studies and the Computer Sciences). This involves work with a stronger analytical grounding than the snapshot ‘impact studies’ characterised above, which addresses how the technical and business outcomes are closely related to (indeed generated in) the process of implementation (Clark and Staunton 1989).

Implementation studies account for the bulk of research into new organisational technologies such as ERP. There is a huge literature on ERP systems (for reviews see for example Esteves and Pastor [2001]; and Al-Mashari [2003]). The ERP Research Group (2006), for instance, has 600 articles in its on-line bibliography and the overwhelming majority, over 95% of these, correspond to what may broadly be described as ERP implementation studies (including also closely related topics such as the management of ERP adoption, organisational outcomes and ‘critical success factors’).3 Though most of the earlier papers were of the sort we have described as impact studies, scholarly research tends to appear later (Grabot and Botta-Genoulaz 2005). This recent work has a stronger social scientific grounding and has been more rigorous and offered more critical insights. We note the growth across a range of disciplines (including for example Information Systems, Organisation Studies, Management of Change etc.) of more sociological work informed by a processual understanding of technical and organisational change and deploying qualitative, often ethnographic, research methods, which yields a richer knowledge base, going beyond the standard unitary managerial view of the organisation and addressing different perspectives within the organisation and the particular processes which underlie these outcomes.4

6

Problems of the temporal and societal framing of implementation studies While welcoming these studies we draw attention to potential forms of bias that arise from the temporal and societal framing of implementation studies (Williams 1997).5 The severe limitations of impact studies and methodological shortcomings of snapshot studies have already been outlined. Though many implementation studies may be more grounded, they are still typically retrospective accounts and of short duration compared to the extended timeframes involved in the complete adoption cycle (involving the initiation, procurement, implementation, use and subsequent review) for such kinds of radical technological and organisational change.

Retrospective studies suffer the risk that respondents will modify their responses with hindsight and often align with a managerially sanctioned view. For example, as McLaughlin et al. (1999: 104) observe, the selection of an information system will typically be described by informants as having occurred in a ‘carefully managed’ and ‘rational’ manner. There are likely to be biases in empirical access, since user organisations, let alone supplier firms, are not necessarily keen to promulgate accounts of failure or allow researchers access to projects that are the subject of overt conflict and controversy.

It is moreover necessarily difficult for external researchers to gain access to the earliest stages of a project to address its inception, what Gerst (2006) calls the ‘preproject stage’. One obvious reason for this is that, in developing studies of the adoption of a particular technology, the selection of firms for inclusion in the casestudy or survey is on the grounds of their having already taken at least the ‘in principle’ decision to adopt. As Pozzebon & Pinsonneault (2005) point out, choices by organisational actors at this stage configure the initial context of an implementation project and can have important consequences for its conduct and outcomes.

Researchers, limited by the practicalities of research funding and agreeing access, are also liable to leave too soon as well as arriving too late. Implementation studies are still often based on short- or medium-term access, with fieldwork covering a few months or at most a year or two, and are therefore weak in terms of assessing longerterm outcomes of innovation episodes for organisational users. This may be important 7

given the large body of research, dating back to Arrow’s (1962) analysis of ‘learning by doing’, that stresses the significant post-implementation improvements in productivity as organisation members discover and refine ways of using artefacts more effectively. This process of trial and error learning and struggling is key to our social learning perspective, which highlights crucial innofusion and domestication processes. Recent studies of ERP implementation, and especially research concerned to assess its outcomes, increasingly stress the need to look at this extended ‘postimplementation’ phase (Berchet and Habchi 2005). Various analysts have further divided this into the ‘shakedown phase’, and the ‘onward and upward phase’ as these complex systems are coupled with organisational practices and as their further utility for the organisation are discovered and exploited (Markus & Tanis 2000; Somers and Nelson 2004; Robey et al. 2002).

Implementation studies that end too soon may thus underestimate the eventual organisational consequences of an innovation. Indeed many critical researchers, keen to highlight the gulf between promised benefits of an innovation and its outcomes, may have unintentionally replaced the ‘Can Do’ rhetoric of technology supply with a misleading ‘No Can Do’ scepticism about its ability to reshape organisational contexts, emphasising the barriers to fulfilling their promise of delivering rapid organisational transformation – barriers that are rooted in particular in the diversity of local working practices (Williams 1997). In the short term, organisational structures and practices appear to be more robust than the organisational templates embedded in the machine (Webster 1990). However, there is a danger, and this is a very real danger when extrapolating from individual implementation studies, of overlooking the gradual alignment and harmonisation of organisational practices that may occur around the organisational templates embedded in the technology (Williams 1997). For example, Webster and Williams (1993) describe a case in which a 2nd implementation of CAPM succeeded, after the first ‘failed’, aided by the informational and change management practices that had been put in place in the course of the original implementation.

Various writers pursuing a deeper understanding of the organisational consequences of technologies have sought a more intricate, dialectical understanding of the interplay between organisational structures and artefacts (Orlikowski 2006). Thus Kallinikos 8

(2004a) sees ERP systems as embodying organisation templates and taken for granted views of the firm and at the same time reinforcing the routinised view of organisational activities, thus conditioning the behaviour of organisation members. Benders et al. (2006) similarly suggest that standardised organisational technologies like ERP, with their general models of organisational centralisation and standardised business processes, encourage diverse organisations to align with their embedded organisational models. They argue that this may constitute a new form – that they term ‘technical isomorphism’ – of the isomorphic pressures asserted by neoinstitutional theory as causing organisations to become increasingly similar. We are sceptical, however, of accounts such as the neo-institutionalist isomorphism thesis, which simply emphasise stability or change. Our analysis seeks to explore the intricate interplay between stabilising and dynamising factors, which often lead to more uneven outcomes. Thus, changing managerial alliances and circumstances may promote ‘drift’ over time as a result of misalignment between an ERP programme and evolving organisational exigencies (Lee and Myers 2004; Nandhakumar et al. 2005).

Studies conducted from a technology-supply perspective tended to see the user organisation as a blank slate, amenable to technology-induced transformation. However, organisations are robust socio-technical systems, with their established technical and managerial divisions of labour and assemblages of routines and working practices, formed through their particular history and earlier phases of technical and organisational change. We therefore need better tools to characterise organisations. Clausen and Koch (1997), for instance, have analysed these under the heading the Company Social Constitution (CSC), pointing out that particular episodes of technological change are just small and localised moments in the evolution of the CSC.

Many other implementation studies demonstrate the reverse of this problem and have tended to treat the software and its supplier as a something of a ‘black box’. This is partly due to the framing of their research. Lacking access to sites of technology development, studies of implementation had little opportunities to scrutinise the development processes and history that had given rise to it.6 Any inference about supplier behaviour made in these studies is thus primarily derived from observations and perceptions within the user organisation. Implementation studies have been 9

mainly reticent about the world of technology design. Rather paradoxically, perhaps, the software package vendor appears to have been made ‘other’, and, where discussed, one-sided accounts, and on occasion negative stereotypes have often been deployed to characterise the behaviour of vendors and of consultants. Drawing perhaps on critical perceptions of supplier offerings and behaviour within the user organisation, these accounts often convey a negative sense of the role and contribution of external technology and knowledge providers that seems hard to reconcile with the fact that it is the user organisation that hires the vendor.

Technology design/development studies This brings us on to consider how to address technology supply. This is linked to our concern to understand the material properties of organisational technologies. This aspect has not received sufficient attention overall. Rather few studies have been undertaken of the contexts in which organisational technologies have been developed and have evolved. We can identify some practical reasons for this, which include the difficulties in obtaining access to commercially sensitive sites of technology development, and the fact that there are far fewer developer firms than users. No less important may be disciplinary divisions: Organisation and Management Studies are concerned with organisational process and outcomes and have therefore tended to focus on the organisational user and black-box the supplier and their technology.

However, a number of writers have shared our concern about the need to look at the historical development of these artefacts prior to their organisational implementation. This is in relation to earlier systems like MRP II (Clausen and Koch 1999) and current ERP systems (Kallinikos 2004a; Koch 2005, 2007). There has also been some research into technology development processes, notably within a Technology Studies perspective. These latter studies have, however, often been carried out in isolation from technology implementation and use (MacKay et al. 2000). This fragmentation and framing of enquiry has consequences. In particular those analysing design may succumb to the temptation of seeking to infer the implications of particular design choices for those using the artefact, as exemplified by Woolgar’s (1991a) much cited idea that the designer ‘configures the user’. This kind of approach to analysing technology design, with its rather simplistic presumptions about development 10

capabilities and how design choices will affect the user yields a bizarrely overpoliticised account of design (Stewart and Williams 2005). This temptation is understandable perhaps. Social scientists want to develop an understanding of design choices and their attendant social implications and outcomes but run into the problem, when studying sites of design, that most of the work is ‘mundane’. Design choices have a ‘taken for granted’ quality and not much seems to be happening in relation to the user. In this context, every comment made by designers about the user and their context is liable to be seized upon. However, an understanding of the everyday reality of design and development work and the internal exigencies that exist within supplier organisations is important – and it is one of the things our book has attempted to investigate especially in Chapter Eight.

In terms of exploring the implications of design choices for organisational users, one approach might be to look for internal technology controversies that have emerged, to find sites where competing options are being contested and where choices and their implications become highlighted. However, This, however, is perhaps more easily done in historical research (Hard 1993). Another approach, and one that we have followed in this study, is to focus upon the various interfaces between suppliers and users which constitute key nexuses in which competing requirements are presented and worked out. In our study, for instance, we focused on the ways in which an ERP supplier utilised groups of users (Chapter Five) to help guide the evolution of its packaged solution. Design – and the coupling between artefact design and its implementation/use - is thus being worked out through a range of different networks and intermediaries linking supplier and user. It is also being worked out over multiple settings of organisational implementation (implementation cycles) and in aggregated form over multiple product cycles. It is important to address these implementationdesign-implementation cycles. For example, as we have previously mentioned, numerous ERP implementation studies have highlighted the problem of ‘fit’ between the package and the various settings in which it is applied. However, upon examination, this problem can be seen to be rooted not simply in the lack of fit between ‘the technology’ and ‘the organisation’ but between the implementation sites for which the technology was initially developed and the sites in which it is currently being applied (Brady et al. 1992; Webster and Williams 1993; Soh and Sia 2004).

11

Need to attend to technology design and implementation in tandem These observations underline our argument about the need to attend to technology design and implementation in tandem. Yet there are almost no studies addressing design and implementation together. Why is this? One reason, as well as the abovementioned disciplinary divisions between those studying organisational users and software developers, is the very practical one that technology development is in most instances not only socially but also temporally separated from implementation. Moreover, the lag between design of a technology and its implementation typically exceeds the duration of most social science research projects (Williams et al. 2005).7 Researchers, contemplating the trade-off between depth of fieldwork and the number and range of fieldwork sites of technological innovation have tended to opt for one or other setting. This trade-off is arguably made more difficult by the emphasis within contemporary social science on ethnographic approaches. Though strong in capturing the richness of local processes in real time, ethnographic methods are labour intensive. Ethnographic researchers therefore have often therefore opted for simple research designs – mostly involving single site studies or studies of a number of closely-related settings.

In this study we sought to overcome these problems, going beyond the current fashion in qualitative social science with deliberately naïve methodologies and utilising our theoretical and substantive knowledge – in particular our theory of the biography of packaged software - to sample a selected array of locales of technology design and implementation. Our multi-site analysis thus addressed settings of technology development and implementation/use and focused in particular on nexuses between design and use and the interactions between them. It integrated the historical and contemporary; and it addressed different locales in the design implementation cycle. We further argue that researchers need to look at artefacts at different stages in their biography. We thus observed packaged developments at early and late stages of this cycle, studying both the birth and evolution of new artefacts, as these differ significantly in terms of the relations between actors and the institutional structure – in short, the mechanisms for mutual shaping of technology and society. Let us look at these points in turn.

12

Addressing the social fabric Let us recap our point of departure from mainstream constructivist approaches to the analysis of technology and work. Early contributors to the self-styled ‘New Sociology of Technology’ (Pinch and Bijker 1984) primarily addressed innovations and innovators that established a new field of techno-scientific practice. As these were often at the interstices of existing institutional structures, ANT and SCOT theorists were able in their analyses to ‘foreground’ the actors directly involved. These often focused particularly on ‘heroic’ technical specialists, who were conceived as ‘Sartrean engineers’ (Latour 1987), apparently outside or able to operate free from constraint from the social structure. They thereby relegated to the background, or ignored entirely, the historical and institutional factors which underpinned these developments. However, such actor-centred accounts yield unbalanced explanations. Their shortcomings are particularly problematic when we try to use these frameworks to analyse the development of workplace technologies and other instances of incremental innovation within well-established institutional settings. Local actions are sustained and constrained by an extensive network of technical, organisational and social arrangements whereby some (material, institutional) elements are difficult for local actors to change (Kallinikos 2004a; Koch 2005).

Clearly, we need a ‘contexted view’ (Morrison 2002) that can address the complex social fabric and its history which pattern the activities of those involved locally. Moreover, our explanatory frame needs to be one which avoids the simplifying logics of particular disciplinary approaches or schools, and which can match the intricacy of the settings and processes we are studying. We start by an observation that the character of these applications is being fought for, and shaped, at a number of levels ranging from local contestation around features of design or its organisational implementation to the broad macro-level concept. The complex web of relationships involved moreover changes over time; it is as Koch (2007) observes a ‘moving target’.

Thus far, we have established the importance of sites of technology development and its implementation and use in understanding innovation. This alerts us to the need to address how individual actors (e.g. suppliers, potential users, intermediaries) and the

13

relationships between them are conditioned by their broader setting. How then shall we conceptualise the broader setting?

Koch (ibid.) has criticised the existing explanatory frameworks of ERP studies as being too simple. In an important series of articles, he maps out a broad framework: local studies (e.g. implementation studies) will no longer suffice he argues; ERP is both ‘local’ and ‘an institution’. However, he also expresses dissatisfaction with dualistic analyses that counterpose local and institutional developments (2005). He proposes, instead, that we should examine ERP as ‘a community’ (Koch 2003) constituted by joint involvement in a technology. Koch (2007: 440) observes that in future ERP studies need to ‘…go beyond the single space enterprise, as well as moving away from implicit assumptions of stable states of the system’ and to adopt instead a ‘multi-local’ analysis of technology. To this end Koch (ibid.) proposes as a conceptual frame, a six-field matrix, encompassing short- and long-term, and Micro, Meso and Macro elements, of which implementation is but one out of six aspects. We find ourselves in complete agreement with Koch’s analytical project (and particularly his call for more effective analytical templates addressing multiple locales and histories). We hope in this work to contribute to this common goal in some way.

However, we have some reservations about the Macro-Meso-Micro distinction. Though these are very convenient and communicative labels (and we will use them as such) they run the risk of being mechanistically mis-construed as fixed and separate levels when in fact they are interpenetrating (a point which Koch of course recognises in his critique of the local-institutional dualism). Moreover, these need to be seen as relational categories. In other words, what is seen as Macro and Meso is relative to the sets of (‘local’) actors and issues under examination (Kaniadakis 2006). It is perhaps more useful to develop a generic model of the social shaping of complex, commodified organisational technologies such as ERP, rooted in empirical analyses, that can provide a complex template for our analysis.

Developers, users and the developer-user nexus The need to attend to the developer-user nexus is flagged by various approaches. It is for example axiomatic to the Social Shaping of Technology (SST) perspective which 14

insists that the development of industrial IT applications is not just shaped by strategies of designers/developers in the domain of technology supply but also, as previously noted, the settings of implementation and use. Fleck’s innofusion concept (Fleck 1988b, 1993; Fleck et al. 1990) flags the innovation processes taking place in the so-called diffusion stage in the process and arenas of technology implementation, and the possibility that these experiences may feedback into future technology supply. It also highlights instances where this experience was not effectively utilised. The social learning framework builds upon this important observation and proposes an integrated picture of the development and implementation of technologies (Williams et al. 2005). Like Evolutionary Economics, the Social Learning perspective draws attention to the importance of the coupling between supply and use. However, unlike the main tradition in Evolutionary Economics with its concern to assess the differential overall efficiency of such coupling in particular national or sectoral innovation systems, the Social Learning perspective seeks to explore the detailed processes and mechanisms for this coupling and its consequences for innovation pathways.

The importance of paying attention to these coupling mechanisms comes immediately to the fore when we examine COTS solutions. Here we find an extraordinarily intricate web of formal and informal linkages between package vendor and organisational users. Here our analysis takes us beyond the immediate interorganisational level of direct interaction between supplier and user. For example, our concern to analyse procurement stimulates us to address the broader terrain of suppliers of classes of products. New industrial technologies develop in parallel in a multitude of user sites and through the activities of many vendors and associated players (e.g. consultants). It is to this Meso-level, beyond the direct user-supplier linkage that we now turn.

The fine structure of external experts, intermediaries and knowledge networks We can track an array of relationships out from the organisations contemplating COTS adoption. Here we note first the various avenues of information the user organisation draws upon in deciding to resort to a packaged solution like ERP and in selecting a particular vendor and providers of complementary products and services. Finkelstein et al. (1996: 1) have drawn attention to the influence of advertisements, 15

supplier literature and demonstrations. They also point to the importance of observed use of the packages in other settings (for example, demonstration sites where COTS have been installed) and comparative studies provided by third parties (trade papers etc.). As well as drawing information from knowledge and technology suppliers, potential adopters seek more impartial information through informal links with similar organisations. In addition, our study (Chapters Six and Seven) will highlight the growing influence of industry analysts as providers of community information about the provenance of particular vendors and their offerings.

Once a packaged solution is determined upon, a more tightly coupled set of contractual relationships will be established by the adopting organisation with those charged to deliver an ERP implementation, involving a wide range of actors. As well as the vendor itself, there could be the suppliers of associated products (hardware and bolt-ons) and various sorts of external knowledge providers. These include groups such as consultants who offer expertise in ERP implementation and who, in turn, might be assisted by external providers of other sorts of relevant expertise such as systems integration or change management experts more generally.

The management fashion for outsourcing organisational functions increases the range of external experts utilised, which may also include, for example, training providers. This trend is reinforced in the case of ERP by the explosive growth of the market, which outstripped the capacities of suppliers to provide this expertise, as well as the exigencies of supply of generic solutions that motivated the deliberate decision by suppliers such as SAP to focus on generic supply. These decisions are buttressed by an array of firms, such as the suppliers of interoperable products, systems integrators, implementers, trainers, change management experts, and so on (Sawyer 2001; Koch 2005, 2007). Koch points out that this differentiation is crucial to the valuemaximising strategy of the package suppliers, for:

ERP is (big) business and design of these systems occur under strategies of mass customization, where the encoding of the generic user is a necessary tool to reduce development costs and time to market (Koch 2005: 43-44).

16

We can also track these webs of knowledge and influence from the vendors as they seek to manage and sustain their existing user base and expand their markets. It arises in the course of their efforts to provide service and support to users and also to guide the development of new products, and through linkages to existing customers. In this latter respect this is often indirectly through knowledge arising from their services (e.g. help-desk queries) and also more directly and explicitly through permanent knowledge linkages (e.g. through user-clubs; marketing department discussions with potential customers).

The increasing resort to outsourcing the supply of business solutions (and their supply as packaged rather than bespoke solutions) and the increasing role of consultants in the supply of knowledge services radically transforms the institutional terrain in which changes in industrial technologies are adopted (though there may be some similarities in the process of implementation [Brady et al. 1992; Fincham et al. 1994]). In particular, it changes the character of the conflicts of interest from a primarily intra-organisational contest for political legitimacy and access to resources to a primarily inter-organisational contest for contracts and streams of income and services. As already noted, this also changes the character and role of internal expertise and sets up complex alliances between organisational interests and external economic interests (such that public and private organisations carefully regulate such linkages) (Procter et al. 1996; Howcroft & Light 2006).

The selection of consultants and the role delegated to them, shaped by established ways of working and reputation from previous projects, configures the arena in which the project unfolds in ways that may give consultants more or less autonomy and influence over the outcomes (Hislop 2002; Pozzebon & Pinsonneault 2005). We find that change is taking place in a complex social setting, and one that is patterned by pre-existing social relationships. Clausen and Koch (1999) similarly identified more or less stable couplings between particular groups of user organisation and vendors, which they have described as ‘segments’ of the ERP market.

The segment Clausen and Koch (1999) explored how the shaping of ERP in the 1990s took place across a structure comprising the CSC of the adopters, with their own internal 17

dynamics, and various ‘segments’ of IT suppliers and customers.8 They suggest that knowledge flows within these segments were shaping the evolution of ERP. This included implementation experiences, and the new demands and visions circulated between suppliers and their customers. Drawing on theories of ‘organised capitalism’ (see Lundvall 1985), Clausen and Koch (1999) see the persistence of these segments in terms of the benefits of these knowledge flows and a coalescence of similar views about business improvement. Crucially, they argue that different segments, and the different procurement strategies and associated forms of supplier-user coupling, offer different opportunities for local influence over the design of the ERP system (see Figure 3.1).

FIGURE 3.1 NEAR HERE

Later work by Koch points to the influence of broader and longer term changes affecting these ‘Meso’ structures, including the restructuring of the ERP supply sector in the 2000 economic downturn which swept away some of the weaker and small players. These segments are not stable. There have also been some realignments in the constellations of players around ERP provision (generic solution providers, suppliers of complementary products, implementers and other Value Added Resellers) in a complex pattern which combines elements of stability as well as dynamism (Koch 2004, 2005).

Our earlier work on CAPM had highlighted the influence of similar Meso-level features on the way these artefacts were shaped and promoted in part through industrial and professional networks (notably in the UK by the British Production and Inventory Control Society,9 a body strongly influenced by vendors and consultants as well as industrial managers). CAPM was also influenced by more general (and changing) models of best industrial practice (Williams and Webster 1993; see also Swan and Newell 1995; Robertson et al. 1996; Hislop et al. 1997).

And today, as we already noted in Chapter One, there is a marked re-orientation of ERP supply. This is motivated, on the one hand, by the fact that it has largely saturated its original established markets in large manufacturing organisations and now must find new customers or develop new functionality for existing customers, 18

and, on the other, by a desire to respond to criticisms that its offerings have lagged behind the latest business concepts, in particular of value-networks. So we see ERP being re-oriented towards new targets (public services and SMEs whose requirements it has been criticised for neglecting), the development of new technology architectures (in particular of the internet and web-services), and new functionality (CRM and supply chain management) which brings it into competition with established players in these niches. Its future is being debated and contested in this inter-organisational space as much as through organisation level implementation.

Macro-level - Visions and Imaginaries This final observation forces us to consider developments at a more Macro-level. This is, first, in terms of the relationship between these changing conceptions of an organisational technology and the circulation of broader views of industrial improvement (which inform prescriptions of good/best practice). And, second, with visions of how these may be fulfilled by emerging technologies. Previous work has noted the correlation between technologies and business programmes. For example, in the 1990s, the European Union invested heavily in electronic trading as it matched their vision of an open international market (Williams 1997). Similarly, 1980s conceptions, such as integrated information systems, were the technological correlate of the integrated, customer-focused finance organisations (Fincham et al. 1994).

Swanson and Ramiller (1997) have highlighted the role of ‘organizing visions’ in information systems innovation, encompassing interpretation, legitimation and mobilisation, which help mobilise the material and intellectual resources needed for innovation. Wang and Ramiller (2004: 12) analyse the evolution of attention (in what they call an ‘innovation community of vendors, consultants, adopters’) from:

i. knowing-what: Interpretations that help to conceptualize the innovation;

ii. knowing-why: Rationales for adoption that help to justify the innovation;

iii. knowing-how: Implementation and utilization strategies and capabilities that capacitate the innovation. 19

Whilst initially the focus of attention is on what the technology is, and why it should be useful, later attention shifts to issues in its successful implementation and exploitation by user organisations.

Judith Gregory has developed the concept of ‘incomplete utopian project’ to ‘describe the phenomenon of envisioning as constructed, evoked, and employed within an innovative intra- and inter-organizational effort, and to open up theorizing about innovation, work practices, and technology’ (2000: 180). The word ‘utopian’ draws our attention to the influence of ‘longstanding deeply shared desires simultaneously characterized by their unrealizability and their devotees' tendencies to over-reach reality in their pursuit’ (ibid: 194). Building upon this as well as a shared tradition in activity theory, Hyysalo (2004, 2006) has developed the concept of ‘practice bound imaginaries’ to refer to such visions which provide an intellectual resource that helps frame, mobilise and pattern expectations around an array of players. Instead of portraying visions as disembodied, the concept of Practice Bound Imaginaries conveys the extent to which they are coupled with existing and achievable practices and artefacts:

relatively integrated sets of visions, concepts, objects and relations that are regarded as desirable, relevant and potentially realizable in and for a practice, and as having cognitive and motivational power for organizing this practice. (Hyysalo 2006: 602).

This bears directly upon our analysis of the evolution and biography of ERP (and for example Koch’s [2003] analysis of ERP as a political programme for organising change). A key part of the ‘heterogeneous assemblages’ (Koch 2005) of human and material elements that constitute ERP is comprised by inter-subjective elements: promises, visions of best practice and prescriptions for industrial improvement, and criteria for assessing technologies alongside artefacts, techniques and practices. The biographies perspective however helps us analyse how these communities operate across the diverse set of social actors involved: suppliers and users, consultants, industry analysts, policymakers and commentators (Koch 2007).

20

Macro-level – external developments These final observations force us also to consider changes in wider terrains that shape and are reshaped by these developments in the world of ERP. These include, for example, at the broader level:

i.

changing models of economic life in an increasingly globalised networked economy;

ii.

the structure of the information technology sector;

iii.

the emergence of new technological models such as web-service architectures, which could herald major changes in how the information infrastructure is created through third party hosting etc.;

iv.

the structure and recent restructuring of corporate knowledge services involving the increasing influence of a small number of global management and accountancy firms, and the convergence of information systems integration and change management organisations upon high value-added opportunities from deploying these in tandem (and the associated linkages and mergers between these firms).

This brief review of developments around ERP brings us nearer to being able to sketch out a schema for analysing these multi-local developments in more abstract terms.

ARENA AND AGORA: CONCEPTS FOR EXPLORING THIS COMPLEX SPACE Arenas of technology development and implementation How then can we conceptualise this complex space, linking together material artefacts, practices and visions within an extended fabric of individuals, organisations and inter-organisational structures and associations. As Koch (2007) argues, we need better spatial metaphors for addressing this rich tapestry, which is characterised by gaps in time and space (e.g. between developers and users, as well as by more or less sharp differences of interest, expertise and commitment). We could theorise this as a 21

‘distributed innovation process’ (a concept recently advanced by innovation economists) or as the operation of a ‘network’ (in the way Actor Network Theorists might do). These however represent a very imprecise way to characterise what is in fact a rather structured set of relationships. As others have noted (see Knorr-Cetina & Bruegger 2001, 2002), the concept of network has come to be widely adopted but in a remarkably loose as well as inconsistent manner.

We have been attracted by Jorgensen and Sørensen’s (1999) concept of ‘development arena’. The value of the concept for us is that, seeking to provide tools for ActorNetwork Theory based explanation to deal with the broader interactions evident in global technology developments, it conceives of the arena as a space, using the analogy of a circus ring drawn in the sand, in which a number of more or less conflicting actor-worlds collide. In addition, they flag the possibility of radical reconfigurations of an arena through changing boundaries and realignment of players, providing tools to explain destabilisation as well as alignment:

…a development arena is a visualizing spatial expression of processes of competition and co-operation. It should convey the idea that several actorworlds are being construed within the same problem area. It depicts the idea that several actor networks co-exist and interfere with each other within a certain problem space. A development arena is our attempt to bring together processes or entities that would otherwise seem to be dislocated. It can be seen as the place where actors relating to a certain set of problems meet and exchange ideas etc (ibid.: 417-8).

Fleck’s (1988a) innofusion framework had similarly flagged the arena of implementation as a key site of innovation in industrial technologies. He has a rather similar concept of the arena as an inter-organisational space comprising members of supplier and user organisations, that constitutes a setting for practical learning and struggling, in which different kinds of competence and knowledge are deployed (e.g. the engineers’ knowledge of computer science techniques and artefacts and the organisation members’ knowledge of their context and purposes). Indeed our initial concept of the biography of an artefact was based on the idea of an artefact alternating

22

between moments of innovation in technology supply and implementation (see Williams 1997).

This kind of formulation seems less adequare to us today as our attention has turned to look in more detail at the myriad forms of direct and indirect relationships linking supply and use and also shaping the overall character of offerings in a technological field. It would be possible to scale up Jorgensen and Sørensen’s (1999) development arena to include implementation, but this would be to overlook the asymmetries and tensions between development and implementation. It is useful to examine moments of design and development separately from implementation and domestication as we see these moving not always in synchronisation but often exhibiting different dynamics.10 Industrial and organisational technologies are characterised by a small number of global supply side players, a larger array of complementary product suppliers, and huge numbers of adopters.

Clarke and Casper (1996) have also deployed the concept of arena in their interactionist analysis of medical and scientific developments, as part of an endeavour, which parallels some of our concerns, to address power dynamics that have been muted in ANT and SCOT analyses. They address the relative power of actors by analysing outcomes: the consequences of the actions of social actors within their particular sub-cultures/social worlds and a shared arena. Their approach

…centres on grasping and representing the perspectives and properties of all the major actors (including collective social worlds and nonhuman actors) in a particular arena of mutual concern or in which certain actors are implicated (ibid. : 602).

How then can we characterise the complex spaces in which technologies like ERP emerge and evolve? To characterise these as a single arena may be to underplay the very different textures of the fabric of social relations (which for example range from contractual linkages between firms to weak associations of opinion across dispersed communities). We could alternatively describe the setting for development and evolution of organisational technologies such as ERP in terms of a multiplicity of overlapping arenas: these could be development arenas, implementation arenas and 23

specific Company Social Constitutions (for CSCs are surely arenas too). This however might distract attention from the fact that many players will appear in multiple arenas. Characterising these as separate spaces may not be helpful to our current concern to develop multi-local theorisation of both the many kinds of supplier-user relationships and of the overall development of a technological field. Instead, we want to look at the various different kinds of relationships established between broadly similar or at least strongly overlapping groups. We also need to be able to analyse these at different levels of generality and timeframes.

The ‘agora of technology and organisational change’ Antonios Kaniadakis (2006) has introduced the concept of the ‘agora of technology and organisational change’. He sees the agora as a meeting place and a market in which all producers and consumers of organisational technologies potentially interact. The agora, thus conceived, is diffuse and not clearly bounded. However various particular bounded perspectives on the agora are drawn (by actors and analysts) for different purposes (of action and analysis) depending upon their particular context and purpose. In other words, actors construct particular viewpoints of the agora: they see and engage with particular slices of the complex multi-local multi-actor space of the agora and set boundaries depending on their purposes and relevancies. How the agora is conceived depends upon the actor’s relationship with it. Thus, a user organisation has very different view and orientation towards the agora than a technology vendor.

Viewpoints are active constructs; it is not simply a question of where you stand. It is also a question of the purposes of players constructing it. The agora is also, and perhaps primarily, an analyst’s construct. The researcher makes choices about which tranches of this complex structure to sample and with what closeness of view.11 The arenas we have discussed may be seen as particular viewpoints within the agora. The agora has a structure (which we shall attempt to map empirically and conceptually). Thus, Kaniadakis sees the agora as having Micro-Meso- and Macro-levels. However, what appears as local and as broader context also depends crucially upon what is being examined and how. For a study of interactions in a particular workgroup, global technology developers may appear as established features of the macro-environment, along with other legal and institutional structures, that are not amenable to influence by the actors in the timeframes involved. In a study of these technology developers, 24

however, the market of (unknown, distant and thus impersonalised to them) potential users may appear as an obdurate and immovable constraint. This is then a relational (not a relativist) conception. The concept of agora would seem to meet Koch’s (2003) call to go beyond a dualistic local-institutional conceptualisation and address ERP as ‘a community’. Moreover, it opens up opportunities to address the intricate structure of this community and develop methodologies to capture this.

The concept of ‘viewpoint’ gives us a means to discuss the necessary choices in research design when trying to address complex social phenomena. It provides a means of steering between naïve undifferentiated approaches to approaching the social setting and adopting a particular conception of social structure. It links with our argument that research design needs instead to adopt a ‘variable geometry’.12 The viewpoint concept flags that choices must be made in terms of sites of access and tools for investigation, depending upon the phenomena being investigated, the kind of access the researchers have secured and their analytical purposes. Effective research design calls for thoughtful selection of sites and methods of investigation. The same phenomena could, for example, be addressed through large-scale survey methods or ethnography. Which is most appropriate depends on the research questions involved.

The dynamics of the agora The agora may be a diffuse and plural array of players. However, it is not an open and equal community – in the way in which we might conceive of scientific communities operating under the Mertonian ideal, for example. Its internal structure comprises not just peer-like communities of practice (a la Wenger 1998), but also communities of (often conflicting) interest. It is characterised by asymmetries and entrenched conflicts as well as alignments of interest. Criticisms have been advanced of the failure of community of practice theory to develop an adequate analysis of power, ideology and conflict, particularly in inter-organisational settings and despite its initial recognition that these were potentially important (Fox 2000; Roberts 2006). Many of the points made by Koch (2007), explaining why his analysis of ERP as a community could not be in terms of a ‘community of practice’, are applicable here.

Moreover, the agora is not only comprised of diverse and heterogeneous elements, it is also a disjointed space, perhaps better understood as an heterogeneous assembly of 25

physical and abstract spaces. For example, the agora of technology and organisational change though viewed perhaps from particular national settings is closely coupled with bodies and groups with international scope. However when we consider a developing country like China, for instance, we find a space only partly integrated with the Western agora for Enterprise systems and organisational technologies more generally, with very different traditions amongst ‘user organisations’ and national software suppliers despite the presence of some Western multinationals (Liang et al. 2004; Xue et al. 2005; Wang 2007b).

Commodification of knowledge networks The ERP community is a knowledge network. It is also, however, a locus of struggle and conflict. Many parts of this segment of the agora are subject to commodification. This imparts a complex dynamics to the agora. We have already discussed the difficulties of trading in informational and non-material products. Thus in the case of ERP, we find various commercial suppliers of knowledge based products (ERP solutions, and complementary products) and services (ERP consultants and other change management and integration specialists). In Chapter Six we will analyse in detail the attendant difficulties in establishing the provenance of a provider and its products, and of demonstrating the benefits of that solution to a particular user organisation and of overcoming mis-fits. Not least, because ERP is a generic product, a substantial investment must be made in implementing it within an organisation before its outcomes - before actual and achievable fit - can be realistically assessed. These difficulties in assessing the qualities of a product mean, on the one hand, that the market is a rather inefficient discovery mechanism, but must be supplemented by network or community types of relationships. On the other hand, outsourcing and commodification radically change the incentives faced by players in commercial relationships in the procurement of technology or (consultancy or integration) service, with sharp and very obvious conflicts of interest between competitors but also differences of interest and of commitment/world view between consumer and producer. Once the procurement process has been concluded, the arms-length externally-policed contractual relationships invoked by economists might be presumed (hypothetically) to apply. However, in a context of necessarily ‘incomplete contract’ issues, this strict contract relationship remains notional. Though the existence of contracts changes the legal and governance character of previously 26

voluntary relationships, the exigencies of joint learning in implementation are characterised by the erosion of boundaries and lines of responsibility. Here we may infer a spectrum of market relationships between what we might provocatively term, following Burns and Stalker (1961), mechanistic and organic relationships, between those in which a more strict versus a more collaborative relationship prevails.

Thus, we see that in both phases of the market relationship, market forms are supplemented by communitarian and network forms of relationship (Fincham et al. 1994; Adler 2001). This is one way in which the relationships of the agora are unlike a community of practice or a scientific community but are shaped by the dynamics of commodification.

Another important way in which commodification shapes the agora is in the making of markets through the alignment of expectations. Here the providers of knowledgebased goods and services find themselves facing another paradoxical situation in mobilising general expectations. On the one hand, they may be drawn to collaborate with their competitors in building expectations that a particular technology/technique represents the way forwards for business improvements. They have an incentive to raise collective expectations about this class of offerings, to establish a market for such offerings, and to promote its provenance amongst other possible ways forward as a road for organisational improvement. Simultaneously they need find some way to promote particular claims about their own offerings, to convey that their particular product offers competitive advantage over its rivals. Thus, we find developers, on the one hand, competing with other providers to build their own particular solutions, to position themselves within their market niche, and build their share of the market, but, on the other hand, also operating in tandem with their competitors to establish the generic idea of utility of classes of artefact.13

This paradoxical situation: the simultaneous necessity of self-promotion and building trust in the class of solutions places interesting constraints upon vendors and consultants. In practice, they often find themselves operating in a complex collaborative array with suppliers of competitive and complementary products and services of the sort that have been characterised as co-opetition. As Swanson and Ramiller (1997) point out, in discussing organisational visions, though some diversity 27

may lead to richness and robustness in innovation, too much diversity and competition may lead to dispersal and loss of commitment.

James Stewart (1999) coined the term ‘poles of attraction’ to explore the ways in which (ICT supplier) firms seek to mark out their plans and visions of future technology in very clear ways: that is to mobilise the expectations of potential customers and thereby build confidence in, and win commitments to, an emerging technology, or, at times, to ward off competitors, to mobilise fear, uncertainty and doubt and thus frustrate a competing technology. This concept highlights the influence powerful players can achieve within the agora based on resources they are able to mobilise.

That there are benefits in aligning expectations around new organisational technologies is evinced by the way in which a series of technologies and change management techniques have acquired solidity and momentum (often by identifying a class of more or less disparate vendor offerings as a ‘technology’, of the sort described by the succession of three letter acronyms such as CRM, ERP, MRP, and BPR), only later to see that term ultimately replaced as supplier strategies and user expectations migrate to the next new solution.

Power and conflicts within the agora The agora refers to a linked array of locales in which economic and organisational interests as well as meanings, are at play and are being played out. From our Social Shaping perspective, we are particularly concerned with economic interests and technological commitments. The agora is a site of conflict and struggle and of negotiation and alignment. The kinds of activities that we note within the agora include many of the processes that have been analysed effectively by ANT (e.g. enrolment and alignment). However, ANT, with its roots in Machiavellian political theory (see for example Latour [1988]), tends to portray these processes in markedly voluntaristic ways that are perhaps more characteristic of struggles in a political realm. However, the agora is a site of economic and political power - that is, of political economy.

28

This is not the place to argue for a particular theory of power or of society – and our biographies framework may be compatible with diverse approaches. However, in our analysis, we are seeking to address the ways in which local struggles are taking place within a broader institutional context - conceptualised in terms of circuits of knowledge and reward: creating visions and providing resources which constrain and enable local actions. We seek to draw insights from both semiotic and institutional accounts of power, and are attracted by Clegg’s (1989: 186) analysis of power relations within the organisation in terms of the operation of multiple ‘circuits of power’ at different (micro-macro) levels. Clegg’s analysis focuses upon the intergroup competition for resources and influence within an organisation, building upon a tradition that could be traced back to Burns and Stalker (1961). The resort to outsourcing of technology and knowledge-based products and services, coupled with transformations in the organisation of external supply, change the character of the actors’ interest-pursuing strategies. In the struggle for influence within the organisation, it places greater salience upon inter-organisational as well as intraorganisational relationships. Members of external suppliers equally must act on an inter-organisational as well as intra-organisational terrain to sustain themselves as ‘points of passage’ for their customers (Law and Callon 1992) and in this way to secure resources and commitments to their own project. There is thus a close interplay between organisational political and commercial struggles and between economic and organisational political power. For our purposes Clegg’s model needs to be adapted to consider the inter-organisational (e.g. Meso) structures as well as the organisational/occupational dimensions he focuses upon.

We seek to explore how local actions and outcomes depend upon a context of knowledge and beliefs and which, in contrast to a narrowly semiotic interpretation of power, provides material as well as intellectual resources which generate incentives and penalties for local players and pattern the conduct and outcome of local actions, by framing discussions. We are also seeking to explain the ways in which local actions collectively react back onto and produce/reproduce social structures.

In this analysis we are anxious to avoid dichotomising action and structure or reducing outcomes to the operation of one or the other – whether an actor-centred or institutional account. Unlike ANT, we do not wish to do this by dissolving everything 29

into the homogenising framework of the language of actor-networks, compressing all the different components into an actor-centred account and linked by ANT’s limited repertoire of relationships (such as Callon’s [1986b] generic translation strategies). Instead we wish to pay due attention to the complexity of operation of socio-technical phenomena, differently constituted and observed at multiple levels of generality.

In terms of our analysis of the agora, the distinctions between Micro, Meso, and Macro refer not to different things, insofar as they all are of the stuff of human sociotechnical action (a point made by ANT writers that we are happy to take on board). Furthermore, in this relational framework, scales and perspectives of analysis can shift: what appears as an external Macro constraint could figure as a Meso layer in another analysis, as may be evinced by considering the example of the status of corporate strategy in a study of the short-term behaviour of a work group or in a study of the management of a major technology change programme. The distinction instead relates to the very different practical experience from particular viewpoints of immediate short-term and local actions and the more generalised aggregate outcomes of those actions – embedded in norms, practices, habits, organisations and technologies. This Micro-Macro scale then is in part a distinction based upon the level of generalisation, whereby Macro analyses may address widely-adopted routinised behaviours that constitute institutional constraints, as well as being a question regarding their level of local malleability to local actors.14 Standardised technology artefacts thus present themselves as part of the objective landscape for firms choosing organisational technologies, though in a study of settings of technology design, their embedded technological choices will be more accessible and malleable to members of the vendor organisation.

The agora for technology and organisational change, analysed by Kaniadakis (2006), is a site for the mobilisation of promise and expectation (and likewise a site for counter-enrolments and mobilisation of uncertainty and doubt) at various different levels of generality. This may, for example, range from particular organisational implementations and supplier offerings, to classes of organisational technology, and ICT capabilities more generally.

30

We thus see the agora as itself a product of a series of enrolment efforts and struggles, which may be described from different perspectives and at various levels of generality/pervasiveness and historical timeframes, ranging from immediate contexts of local action to more generalised patterns of behaviour, sustained over longer-terms, which in turn constitute economic, technological and institutional structures. These broader ‘structures’ act to pattern innovation, providing resources and material constraints to actors in terms of their choices regarding which options appear doable, which factors can realistically be changed, and which are to be taken as part of the landscape.

This double-sided character of the agora, as both shaping and shaped by sociotechnical processes, may usefully be approached through the concept of negotiation, with its two distinct connotations. Firstly, this is negotiation as a meeting, a place for alliance building, conflict and struggle with more or less obdurate or amenable human and non-human elements; and secondly negotiation as a set of manoeuvres needed to accommodate or by-pass those elements which are effectively ‘non-negotiable’. This second usage of negotiation, which is akin to the way we might negotiate ourselves down a mountain pathway, is informed by the fact that some of the things we encounter present themselves as more solid and permanent from the view of particular local actors, including institutions and technologies, which in this sense, are a kind of materialised institution, and have to be negotiated around.

This is not to say that these external factors directly imposed determinate constraints on local actors. Instead, we note another form of negotiation, as a kind of bridging between these spaces. For example, whereby key organisational players/systems represented external factors within the organisation. Finally, though we have for illustrative purposes discussed the case of elements seen as malleable or as fixed to local actors, there is no necessary polarisation between these. Actors can exercise choice about which ‘black-boxes’ to accept as fixed and which to open up – depending, for example, on the relative costs and benefits of so doing. An organisation could, in principle, choose to develop a bespoke solution rather than be constrained by existing availability of packages but this would depend on the level of resources they can deploy to such ends and whether such an investment would appear justifiable. 31

Finally, we notice that the agora is a space where very different kinds of commitment are being played out. There are, in particular, important differences in orientation to the agora between suppliers and organisational users of technologies and other knowledge-based products/services. Thus for a technology supplier or a consultant the agora of technology and work organisation is the space where their commercial future is worked out. In contrast, a user organisation will have a more contingent orientation to the agora; they are keen to benefit from technologies but it is not normally part of their perceived mission to carry forward their experience with implementing technology as a resource for technology development.

Intermediaries as strategic players in the agora The agora concept provides a space for analysing the various kinds of social relationship beyond the immediate inter-organisational level of direct interaction between supplier and user. To be useful, however, the detailed operation of the agora needs to be filled out and explained. Our concern to analyse procurement stimulates us to address the broader terrain of suppliers of classes of products and the ways in which beliefs about the provenance of a technology are constructed across a community of supplier and user organisations. Our final addition to the framework is to examine the emergence of new kinds of intermediaries who are also market makers and conveyors of community information. In Chapter One we reviewed research findings regarding the powerful influence of the technology supply side over views or what constitutes ERP. In Chapter Seven we will draw attention to the role of various kinds of intermediaries, and in particular the growing importance of a relatively distinctive class of intermediary, industry analysts.

Our Social Learning framework flags the importance of different kinds of intermediary in innovation, in linking supply and consumption and the wide range of roles they play (Howells 2006). However, industry analysts appear to occupy particularly strategic sites in the agora. Our work will flag the importance of the Gartner Group in two ways:

i.

we will show how this actor acts as a repository and organiser of what we call ‘community knowledge’ about the implementation of particular products and about the reputations of their suppliers’; 32

ii.

we will see (as noted in Chapter One) how Gartner’s sectoral reviews consolidate the existence of a domain of technological activity (in this sense of constituting a technology like SAP’s R/3 as an instance of ERP), charting the overall development of particular technologies and their future development trajectory (Mabert et al 2001; Judd 2006).

The industry analysts seem thus to play a crucial role in configuring particular development arenas and in mobilising consensus. It might appear that in some instances it is they who hold the ropes and set the rules of game – defining the boundaries of technology and the criteria by which particular vendors and their offerings may be judged. However, it is also important, while addressing their influence, to attend to the limits on how industry analysts proceed. Thus, we find that they are not able to impose their views. Their ability to play their role (and sell their services) depends on their being seen to operate in a close relation to practice, reflected, as we shall see, in the strenuous attention they devote to legitimating their position as impartial bearers of community knowledge in the face of criticisms of partisanship.

We consider the existence and profile of industry analysts like Gartner to be indicative of a broader development. Our review of the difficulties of the operation of the market for complex technologies like ERP had pointed to the importance of indirect indicators of the behaviour or suppliers and their products – in particular of experience-based trust. However, this kind of trust is slow and expensive to acquire. In looking at the mechanisms for selecting management consultants, Glückler and Armbrüster (2003), as already noted, highlighted the value of networked reputation as a more effective mechanism for overcoming the buyer’s uncertainty. The role of industry analysts in IT procurement points to one mechanism for enhancing the efficiency of networked reputation formation through the commodification and canalisation of the circulation of community knowledge (and how this is subject to particular forms of accountability). We may see this as a response to the deep uncertainties surrounding the procurement of organisational technologies that seem to be compounded by the growing pace and increasing organisational significance of technological change. 33

The need to address multiple historical timeframes A corollary of our insistence upon the need to examine sociotechnical change at multiple levels of generality, in terms of addressing immediate contexts of action and broader contexts, is that we need to consider socio-technical processes temporally, in terms of:

i.

the unfolding of multiple histories; and also

ii.

the different historical timeframes around which an object, event or activity may need to be analysed.

Multiple histories and timeframes are intrinsic to our attempts to capture the evolution of a new technology, addressing, for example, both its development and adoption. In this way we seek to captures the complex set of developments taking place across a variety of locales, encompassing both the ‘local’ context of immediate action and interaction and its patterning by a broader context. This broader context is constituted by the aggregate outcomes of previous actions which, in turn, provide a less-readily negotiable set of factors that frame and pattern outcomes and which need to be analysed over longer term timescales.

It is important to pay attention to the multiple dynamics and timeframes surrounding innovation. We have noted that the dynamics of technology development and appropriation may differ. For example in the case of ICTs, where development cycles may have shortened to a year or two, appropriation cycles may be an order of magnitude greater, with new consumer products taking decades to diffuse into widespread use and having greater longevity (Williams et al. 2005) (though both timeframes are becoming shorter). This longevity in appropriation and replacement cycles is particularly marked in the case of organisational information infrastructures such as ERP.

Particular episodes form part of multiple histories. Thus the implementation of a technology constitutes a moment in the history of a particular Company Social Constitution (Clausen and Koch 1999). It is also one of a number of sites of implementation of a particular supplier offering, contributing through its innofusion 34

and appropriation to the further elaboration and wider adoption of that specific artefact. And that specific story in turn forms part of the evolution of the class of artefacts with which the supplier offering is associated. We have coined the concept ‘biography’ to refer to this history of relationships and sites implicated in the evolution of a specific artefact and a class of artefacts. And the latter can, at a more general level, be seen as a phase in the development of organisational technologies more generally. In the latter three cases, the specific history is nested inside another more long-term generalised set of relations. However, a technology implementation can also be seen as the linking together of two specific histories that may not have been previously conceptualised together: the Company Social Constitution of the organisational adopter and the biography of a specific artefact.

In theorising the multiple tempos that we may need to address in analysing particular episodes, we find considerable merit in the framework articulated by Hyysalo (2004). He draws on Hutchins’ (1995) study of how quartermasters learn naval navigation in a system of distributed action, which portrays the simultaneous unfolding of different histories:

any moment in human conduct is simultaneously a part of the unfolding of a task, the development of the individual doing it, the development of the work community, and the development of the professional practice (Hyysalo 2004: 12).

Hutchins' cube (Figure 3.2) represents these as different speeds of change within a single moment of practice (rather than portraying these as separate levels).

FIGURE 3.2 NEAR HERE

Hyysalo introduces us to attempts within Activity Theory to characterise time-scales for analysing social and technological development.15 His study of the development of new health care technology highlights three key time-scales in the coupling of design and use: 35

i. the prevailing ways of organizing design and use in industrial production. Hyysalo refers here to features of the innovation system liable to be stable over many decades: ‘pervasive and relatively slow changing ways in which design and use are generally organized in industrialized countries’ (Hyysalo 2004: 13);

ii. the coupling of a technological field and a societal practice. Which he sees as relatively stable institutions, potentially stable over years and decades, though noting the possibility of changes in practices, in technologies and in the ways these are coupled together; and

iii. the development of a particular innovation and the organizations and people connected to it.

We can adapt this schema to our own analytical concerns. Hyysalo’s longest timescale (i) prevailing ways of organizing design and use would perhaps correspond in our study to the resort to packaged solutions for organisational technologies. Our concept of biography would encompass his other shorter timescales: (ii) the technological field which corresponds to the biography of a class of artefacts (e.g. ERP systems in general); (iii) and the development of a particular innovation to address the biography of a specific artefact (e.g. SAP’s R/3 system).

The comments we made earlier, in discussing viewpoints and research design, about different ways of slicing through the complex social space represented by the agora, depending upon our location/orientation to it and our concerns, also apply to the historical framing and timescales of our research. Such choices about the temporal framing of enquiry have important implications for what may be viewed. For example, local studies of immediate settings of action inevitably draw attention to the scope for discretion (such as ‘user work-arounds’) but provide a poor vantage point for exploring longer-term processes of technology-organisational alignment (for example around common business process templates within enterprise systems). This may need to be captured by other modes of research (for example larger-scale surveys or longitudinal studies).

36

Rather than invoke one modality of research, our approach seeks to retain awareness of the multiple historical registers that surround a particular phenomena. The choices we make regarding which timeframes and historical registers are to be centrally addressed and parallels our earlier discussion of choices regarding the adoption of a local or of a more global gaze. Whilst the agora concept provides tools for looking at social space, the temporal distribution also needs attention.

We are minded here of the critique of constructivism made by Kallinikos (2004a: 12) on the grounds that the ‘study of technology and its social impact cannot be exhausted at the very interface upon which humans encounter technology. Essential strips of reality are not observable…’.16 Kallinikos is highlighting issues of social structure, of particular relevance when we consider technologies that typically come to us as the result of a more or less elaborate (occupational, organisational, and industrial) division of labour. If we are to address the material character of artefacts, many elements are developed at a remove (socially and temporally) from their sites of implementation and use and are not under the control of actors in user locales. This observation can also usefully be applied to the existing institutional context that provides resources and sets constraints for local action.

We are proposing a relational approach that brings to the foreground certain features for detailed analysis – but within a broader historical register that also records other levels of generality and tempi.17 Our work seeks to find ways of probing and addressing these other levels/tempi through the adoption of a complex methodology. We contrast this, inevitably messy, endeavour to other dominant social scientific research approaches which recognise only a single register for analysis (whether of immediate action or of broader structuring). We see this failing, for example, in economic accounts which are founded upon conceptions of human behaviour that are demonstrably flawed, and also in the ‘atomistic individualism’ which characterises much recent work from a constructivist background which only recognises immediate contexts of action. We contend that this yields an unhelpful reductionist account of complex social processes.

Rather than propose a particular level of analysis, we emphasise the benefits of multilevel analyses, which may have different depths and centres of focus depending on the 37

issue under analysis. The particular scope and framing of analysis selected depends upon the matters under examination. For us the matter of research design and epistemology should be driven by a critical reflection about which (spatial/temporal) slices of complex techno-social fabric are brought into the centre of our analytic gaze by particular modes of research and from what viewpoints.

MULTIPLE METHODS We propose the concept of biography as an instance of a ‘variable research geometry’ that can be applied to diverse issues and in differing contexts, depending in particular upon what issue(s) are being addressed and which entities are being tracked. The biographical approach focuses upon social (or rather sociotechnical) processes involved in innovation and how these are shaped by their context and history. Many kinds of biography are thus possible. For example we could address the biography of an organisation – indeed this is how we understand Clausen and Koch’s concept of Company Social Constitution (Clausen and Williams 1997; Clausen and Koch 1999); the biography of an occupation; or indeed the biography of a managerial innovation such as BPR or TQM (see Mueller & Carter [2005]).

Our concern here is to understand the biography of an artefact which may be conceived narrowly in terms of the development/implementation of a particular innovation, or more broadly of a class of artefacts, or of a technological field and their complex couplings with social institutions, actors and practices. This has been the (often tacit) objective of a diverse array of Social Shaping of Technology studies. These have deployed various research geometries in terms of the historical scale and the level of generality of the phenomena under study. However, what is at stake here is not only a matter of temporal and social framing – of zooming in and out, to use our photographic analogy – but also involves important choices also in terms of the methods and concepts deployed and the relationship of the study with existing knowledge. Multiple methods may be required, knitting together different kinds of evidence including historical studies, ethnographic research, qualitative studies of local and broader development and the use of larger-scale research instruments and quantitative data. These differing kinds of evidence have differing strengths and contributions to mapping the dimensions of an issue. For example local qualitative research may be provide better tools for drawing out intricacies and particularities of 38

social process and is particularly pertinent to exploratory research opening up new understandings of a novel and emerging phenomena, whilst larger-scale research provides a more effective base for addressing regularities and trends as well as for testing hypothesis and models and confirming findings from exploratory qualitative studies (MacKenzie 1988). It may be further adduced that combinations of different kinds of evidence are liable to produce more robust and richer understandings.

Multiple theoretical orientations As well as proposing a ‘variable geometry’ in relation to the temporal and technical/societal framing of research, we argue for a certain level of critical eclecticism in relation to broader worldview, and the theories and concepts that inform it. Of course, theories and methods cannot simply be combined on a pick and mix basis; they are underpinned by different and often incompatible presumptions and tools. Though some have interpreted this truism as constituting a case for sectarian theoretical purity, we suggest a different response. We argue instead that we can interrogate differing analytical traditions in terms of their robustness and applicability to the phenomena in question and their compatibility with other perspectives; we can reason and make judgements about these questions.

Though informed by our close association with the Social Shaping of Technology and Social Learning perspectives, particularly in our emphasis upon material and social structural influences, the biographies approach is not ‘hard-wired’ to a specific theoretical perspective. And many of the schools and analytical currents within STS have common and convergent concerns (Williams and Edge 1996).

We contrast our analysis with the widespread espousal within current STS of what we may call the ‘atomistic interactionism’ in many explanations of the world with roots in social constructivism and phenomenology, which see the world as constructed and reconstructed anew in sites of everyday action. A similar analytical consequence arises from the rejection by ANT of explanation in terms of the operation of broader social structures, accompanied by their rejection, as unwarranted generalisation, of social scientific theories regarding the operation of these structures. Though appealing, these exemplify what could be described as the ‘fairy cake theory of the

39

universe’. This theory was propounded in Douglas Adams’ (1979) now classic work The Hitchhikers Guide to the Galaxy that

…since every piece of matter in the Universe is in some way affected by every other piece of matter in the Universe, it is in theory possible to extrapolate the whole of creation - every Galaxy, every sun, every planet, their orbits, their composition, and their economic and social history from, say, one small piece of fairy cake.

In place of basing their choice of research setting and methodology upon social science theory, these actor-centred accounts generally resort, as noted in Chapter Two, to a ‘naturalistic’ (or perhaps empiricist) approach; seeing society constituted in the observable actions and interactions they study. For example, ethnomethodological research is founded on the view that the social order is constituted in social interactions: it selects particular sites of interaction for study and presumes that the relevant social relations will be present or represented there (Lynch 1993). Studies of particular sites and settings of action, what we may call ‘flat ethnography’, encounter the problem, that Kallinikos (2004a,b) also identified, that many issues regarding the material character of artefacts are determined outside the setting of technology adoption (including the availability of technologies as well as the institutional context which provides resources and sets constraints for local action). Perhaps as a result, ethnographic researchers frequently have the sense of not being in the right place or at the right time (Law 1994; Magolda 2000). One temptation faced with this incompleteness of vantage point is to elevate the importance of the particular settings and interactions studied. This could be exemplified by workplace studies of technology that present organisational information and communication processes, including the appropriation of IT, as of paramount importance, and correspondingly neglect technology design and other distal processes. We would instead propose an alternative solution involving what we describe as ‘strategic ethnography’, addressing multiple sites, selected according to the matter in hand based on our preliminary knowledge thereof. Such an analytical move requires researchers to explicitly recognise and make accountable the strategic choices involved when deciding upon the location and boundaries of ethnographic work. It would in turn require reflection upon the theoretical commitments and presumptions that inform these choices (rather 40

than pretend that it is possible to avoid such choices for example by empirical sensitivity). Some strands of ethnographic research, particularly associated with Ethnomethodology (see for instance the articles with Button [1993]), have difficulties in acknowledging such an approach (which, we would argue, is however more in line with the anthropological tradition).

ANT, with its nostrum of ‘following the actor’, does not limit itself to particular settings, but accepts that research involves making strategic choices about which sites and people should be tracked. It justifies these choices, however, in terms of empirical outcomes; in this sense ANT claims to be able to see ‘where the action is’ (Latour 1987). However, ANT does not provide tools to guide those choices or make them accountable. This claim to be able to resort to a naturalistic method leaves ANT open to criticisms of empiricism (Russell 1986; Williams and Russell 1988). Moreover, a multiplicity of accounts would be possible from different perspectives; any ANT account of necessity involves choices about which actors and perspectives to foreground (Sørensen and Levold 1992). Since ANT has rejected other theoretical knowledge, these choices are made based on largely unacknowledged presumptions (though see Law 1991) and common sense knowledge.

What is at issue here is a particular orientation to theory with which we differ. Across the social sciences, we can find a spectrum of styles and approaches to theorisation, between work that in its insistence upon particular theoretical and methodological approaches becomes purist, and more eclectic approaches.

Analytical purism has attractions. It is easier to justify the particular methodology and epistemology within a well-honed world-view. However, this strategy also brings weaknesses and intellectual rigidities. In particular such purism typically involves a process of simplification of the object of study (as we saw in for example in Chapter Two, in the construction of disciplinary domains in which phenomena were conceived in narrowly ‘economic’ or ‘technical’ terms). This has unhelpful consequences, particularly where it leads to the exclusion from consideration of other forms of knowledge, theoretical and methodological tools and empirical evidence. As a result, it generates accounts of the world that do not match the complexity of issues under examination. Of course, this kind of reductionist analytical strategy has been 41

extraordinarily successful, particularly in Science and Engineering. Within the social sciences, perhaps only Economics and Psychology have succeeded in replicating the reductionist disciplinary strategy of the natural sciences. Processes of disciplinary specialisation within other social sciences have more generally involved the formation of specialised schools of analysis rather than cohesive broader disciplines. The formation of these specialised schools has facilitated rigorous development of analytical concepts and instruments – but we would argue, at a cost of narrowing the frame of enquiry. In particular, we note the tendency for a single level/frame of analysis to predominate. This is perhaps most evident in relation to what is perhaps the key social scientific debate – between explanations of the world in terms of action and structure – where social sciences tend to coagulate into either action centred or institutional centred accounts.

There have at the same time been movements against theoretical purism in the social sciences that have led some to embrace thoroughgoing eclecticism. However, as the rather disappointing achievements of ‘systems theory’ indicates, this strategy has been demonstrably unsuccessful. Much of this work is marked by a loss of rigour in relation to the development of concepts and tools (Hoos 1973; Keat and Urry 1975). We therefore do not propose a retreat from theory. Understanding in social science proceeds by advancing competing interpretations of a complex and interconnected world. Theory is needed to provide critical insight. Atheoretical approaches perennially threaten to become overwhelmed by diversity – lacking the theoreticallyinformed tools to help the analyst impose order on unruly reality (put boundaries around problems, sort out and rank diverse potential influences, select sites for examination, and so on). In consequence, we would argue, they end up with empiricist and descriptive accounts; and they are hard pressed to extrapolate or draw general lessons for practice. A clear theoretical perspective seems to be a pre-requisite for effective social scientific enquiry.

Our work, and our view of the Biography of Artefact perspective, is rooted in and inspired by STS, most immediately social learning and social shaping analyses, but also deeply influenced by writings from ANT. However, we differ with the latter’s rejection of existing social scientific knowledge (despite the articulate defence of this approach found recently in Latour [2005]). What we are not proposing is not just an 42

‘in-between’ position – balancing between eclecticism and theoretical purism - but rather a different relationship to theory.

Our approach to understanding the Biography of Artefacts Framework addresses the technology-society relationship at multiple levels and timeframes and also acknowledges the multidimensional character of these phenomena and thus the potential pertinence of analyses of these phenomena from different (technical, economic etc.) analytical perspectives. The analysis of the biography of an artefact, by acknowledging these multiple dimensions of the phenomenon under study, brings the researcher into contact with other areas of (social and technical) knowledge that are pertinent to the questions under examination. A multiplicity of theories and methods may therefore be pertinent. Acknowledging this point does not, however, answer the question of how to bring them together. As is clear from the above, there are dangers in eclecticism.

What is at stake here is the question of how we relate to (heterogeneous) theory. This is a question of whether social science theory is to be used as an analytical machine or as a tool for understanding.18 Rather than admitting all knowledge claims as being of equal validity, we retain our STS perspective as our core analytical commitment. At the same time, we can hold a range of other findings as potentially pertinent forms of background knowledge that can inform a particular study. Social science theory and methods do not constitute some kind of analytical machine (you turn the handle and out come facts).19 Instead, these bodies of knowledge provide potentially valuable resources, sensitising and guiding the analyst. Rather than rejecting and ignoring other forms of knowledge (as espoused for example by ANT and some related approaches), or accepting that somehow ‘anything goes’, we can make some assessment of the relevance, robustness and pertinence of these theories. We can critically examine the presumptions underpinning these concepts and methods and the available evidential base, in terms of their consistency/compatibility and their applicability in other contexts or for other analytical purposes. For a particular analysis it is necessary to develop an analytical framework that advances our own understanding and acknowledges other areas of relevant knowledge. Rather than resort to a singlepurpose analytical schema, in the way propounded by ANT or Ethnomethodology, we

43

suggest that it is necessary to argue for the adequacy of the methods and concepts deployed according to the issue and phenomena in hand.

ANTHROPOLOGICAL PERSPECTIVES ON ARTEFACTS In a moment we shall explore how our biographical perspective can be applied to understand the history of ERP but before we do this we want to discuss some links with other fields which also share our interest in investigating technology over longer timeframes. This includes work on the social and cultural history of technology consumption (e.g. Marvin 1988; Pantzar 1997) and by cultural and economic anthropologists of technology (e.g. Appadurai 1988; Kopytoff 1988; Thomas 1991). The latter have coined a parallel usage of the term biography of artefact – in a discussion that we find informative but which differs from the aims of this book in important respects. The most well known contribution to this strand is, perhaps, Kopytoff (1988), who argues that an object cannot be properly understood at only one point in time. Rather we should look across its whole life-history to analyse its production, consumption and circulation. Just as we study the life course of people, we should also do the same for technologies. In the same way the status of people change over their life-time then so does the status of an artefact. Yet whilst Kopytoff sets out these general aims for uncovering how objects ‘accumulate histories’ he is in fact interested in objects for much more specific reasons; to explore how they mediate social relations in particular settings and what they reveal to anthropologists about those relations and places. He writes:

Biographies of things can make salient what might otherwise remain obscure. For example, in situations of culture contact, they can show what anthropologists have so often stressed: that what is significant about the adoption of alien objects – as of alien ideas – is not the fact that they are adopted, but the way they are culturally redefined and put to use. The biography of a car in Africa would reveal enormous amounts of information about the relationship of the seller to the buyer, the uses to which the car is regularly put, the identity of its most frequent passengers and those who borrow it, the frequency of borrowing, the garages to which it is taken and the owner’s relation to the mechanics, the movement of the car from hand to hand over the years, and in the end, when the car collapses, the final disposition of 44

its remains. All these details would reveal an entirely different biography from that of a middle-class American or Navajo, or French peasant car’ (1988: 67).

As we see it, Kopytoff’s conception of the term biography is useful for the following reasons. Firstly he is interested in how artefacts are transferred from one place to another, and denotes how they are ‘alien objects’ when they arrive in their new settings. Though he is also careful to point out that even though they are ‘alien’, the object is not imposed on actors since it is typically adapted and redefined according to the needs of each new place. This has obvious parallels with the STS and Social Learning work reviewed previously. Secondly it focuses on how various communities in domesticating these artefacts leave ‘traces’ in the sense that they shape the object in some form (though he falls short of suggesting that the users of the artefact actually contribute to the production of the object which is what we are suggesting with the development of software packages). Thirdly that the current significance and meaning of an object today is always related to the settings and communities to which it was once connected. Finally that the same objects would have different biographies depending on the range of settings through which they travel.

Whilst this form of analysis might be applied to the study of ERP systems – and we have attempted an initial analysis using this framework (Pollock et al. 2003; Pollock & Cornford 2004) - we think there are limitations with the term as it is currently conceived within anthropology. Kopytoff’s focus is principally on the significance and meaning of an artefact and how this changes through its career. It is the multiple ways an object is viewed and understood by the different communities who consume it that make up its biography (the more and varied these meanings the more ‘eventful’ the biography). By contrast, whilst we think ‘meaning’ is important in the context of generic packages we also want to show how systems are not only symbolically but also materially changed over time. A system like ERP, as we have already described, is a heterogeneous assemblage where the various affordances built into the technology are constantly developing and evolving. In addition, whilst reading Kopytoff’s work, it seems that the biography of his artefact turns out to be the story of a highly bounded object20. As he describes it the notion of the ‘artefact’ is the circumscribed one common to anthropology and not the messier notion of artefact found with STS. In this respect, echoing the point above, enterprise-wide solutions are 45

not material artefacts in Kopytoff’s narrow sense but can more usefully be thought as a heterogeneous assemblage (Koch 2007). Thus, we intend the biography of an artefact in this wider more encompassing sense.

Finally we are not wholly convinced the most useful way to study artefacts is solely at the place where the user encounters them. Nor are we convinced that these encounters tell us much about the wider contexts in which the objects are situated (which is Kopytoff’s intention). As we have suggested local studies of adoption offer an inadequate lens for exploring the longer term development of a complex organisational technology like ERP. Contrary to what Kopytoff suggests, extrapolating from the specific object to the wider context can only ever be done in quite general ways. Rather than study the variety of ways in which the same object is adopted by different groups of users (these ‘horizontal biographies’ if you like) we are interested in the production of ‘nested biographies’. Local adaptations, improvisations and work-arounds etc., should be seen as one moment in the biography of a specific object, which go onto form part of the evolution of the class of artefact, which in turn is part of the history of the resort to packaged solutions. Arguably, through moving out from the specific user’s interactions with a specific object in this way, the technology and society relationship is bridged in a more sophisticated way.

However there are other aspects to Kopytoff’s approach that we find more useful, such as his discussion of ‘commodities’, for instance, where he gives particular attention to the movement of an artefact between different economic states (see also Appadurai 1988). It was the way in which an object could be ‘commoditised’, ‘decommoditised’ and even ‘recommoditised’ that Kopytoff found interesting. Indeed it the movement of objects between different commodity states that constitutes the biography of a technology (Dant 2001). He cites the example of one of the most common and exchangeable of technologies, the mass-produced car and notes the process by which some vehicles grow older they may develop into a ‘vintage’. In so doing they become ‘unique’ and no longer a straightforwardly exchangeable commodity (a process Kopytoff describes as ‘singularisation’):

In the homogenised world of commodities, an eventful biography of a thing becomes the story of the various singularisations of it, of classifications and 46

reclassifications in an uncertain world of categories whose importance shifts with every minor change in context… (ibid.: 90).

Now it is clear an object might acquire new meanings (or further develop its biography) in other ways than through exchange (Dant 2001) but for Kopytoff, it is the economic life of an object that has particular significance. This is what he intends when he talks of things as having an ‘eventful biography’ in that artefacts are not always commodities but can (and often do) move between different states (thus the commodity phase is only one part of an object’s social life). Interestingly there is a somewhat similar discussion by the economic anthropologist Thomas (1991) only he reaches somewhat different conclusions. He is also interested in the process of commodification but rather than simply signifying one more stage in an object’s biography this is concerned with detaching an object from its history. For this reason in his book Entangled Objects he describe commodification as a process of ‘alienation’:

Commodities are here understood as objects, persons, or elements of persons, which are placed in a context in which they have exchange value and can be alienated. The alienation of a thing is its dissociation from producers, former users, or prior context (ibid.: 39).

His definition is interesting to us since it highlights the possibility that while commodification/alienation can occur there are sometimes opposing forces present; for some objects ‘disentangling’ can never fully occur. The notion of ‘entangled object’ suggests that some artefacts are inextricably connected to their place of birth. To exemplify this he contrasts those objects that have little difficulty in travelling i.e. those things that were specifically created for exchange - with those that have become so entangled with previous owners that there is the impossibility of exchange. There is, he writes, ‘[v]ery close association between people and some particular objects…’ such that the object ‘may not be transmitted at all’ (ibid.: 72-3).

This strand of work is relevant because software packages as a class of technology are, as we have already mentioned, not simply borne ‘commodities’ or as ‘generic’ products but only become so through various complex processes. One of which is that 47

software package vendors work to ensure that their systems are never too much like the place(s) for which they were built and never too connected to specific people or events. Their fear is that if the package becomes overly connected to, or identified with, these places then it will weaken the potential/ability of their systems to move beyond these places/settings. Thus, one of the things they do is to actively mange the biography of their system, this is the history of relationships and sites implicated in the evolution of a specific artefact, so that they avoid ending up with an ‘entangled artefact’.

REINTERPRETING ERP THROUGH A BIOGRAPHICAL LENS The starting point for this focus on biographies was the observation by Edinburgh scholars past and present (Fleck et al. 1990; Fleck 1993; Webster and Williams 1993; Pollock et al. 2003) that workplace technologies were often condensation of existing work practices, coupled with a view of achievable change geared towards current conceptions of best practice. In other words, information systems were not extrinsic developments coming from outside the industry but at least in part were intrinsic developments. This was obviously true in relation to the earliest phases of process innovation that arose within the ‘user organisations’, for example, in the industrial revolution (Rosenberg 1976) and in the earliest stages of the application of computing (von Hippel 1994). However it continued even after a specialist supply-side had emerged, which continued to be linked to the user inter-alia through the implementation process. Building upon these debates, Brady et al. (1992) suggested that packaged software artefacts had biographies. Williams (1997a) applied this concept to analysing the evolution of CAPM describing its historical evolution from a family of artefacts (MRP, MRPII etc.). And Pollock (Pollock & Cornford 2004) later used a version of the same argument to study the transfer of ERP across other sectors.

In this early work, whilst the influence of the institutional setting was highlighted – including the role of professional associations and of public policy in promoting ideas of best practice (Webster and Williams 1993), our initial explication of the biography framework did not include a comprehensive set of conceptual tools for analysing the social fabric beyond the supplier-user nexus. And in this respect some valuable further work elsewhere has been undertaken elsewhere, for example, by Swan et al. (1999) who noted how national differences in the structure and operation of 48

professional associations had consequences for the uptake and manner of utilisation of these technologies. Clausen and Koch (1999) likewise drew attention to the fragmentation of the supply-user nexus into a number of distinct segments with relatively stable linkages between suppliers and users within segments. However, the challenge now is to theorise in more detail the structuring and operation of this institutional setting.

This is what we will attempt to do now in the case of the emergence of ERP as a field of technology where we will reinterpret its history through the lens of our biographies framework.

Understanding the evolution of technological fields: the history of ERP We have already suggested that the concept of biography could be applied to analyse (in Hyysalo’s [2004] terms): the development of a particular innovation (and the organizations and people connected with it), and the coupling of a technological field and a societal practice. In relation to the former, this may encompass the evolution of a particular supplier offering as well as particular episodes of its design, implementation and use. In relation to the latter, the biography concept can be applied to explore the operation of relatively stable institutions over a period of years and decades in the emergence and evolution of particular technological fields. We can explore this in relation to packaged organisational software – which became known as ERP. Koch (2007: 428) reminds us that ‘ERP “systems” need to be understood as heterogeneous networks, assemblages of human and material elements’. The concept of biographies lets us explore the historical emergence and evolution of this heterogeneous assemblage.

We briefly reviewed in Chapter One the extended biography of ERP which can be traced back to early 1960s stock control systems in vehicle and aerospace production. ERP, like many other popular technologies, has a remarkably well rehearsed history; characterised by a clear succession of predecessors with their own acronyms (though there is less agreement about the distinctions between these stages).21 We found a process of incremental development of the artefact and conceptions of its business application, punctuated by more radical changes, with discontinuities often loosely associated with changes in terminology. We will discuss the role of a technology 49

name in more detail below. These labels refer not to specific homogeneous artefacts but to a more or less heterogeneous collection of artefacts (software, management techniques) which link a community (or rather several overlapping communities) of suppliers, intermediaries and adopters.

It is instructive to focus on these discontinuities – and changes in designation. They do not reflect simple ‘technical changes’, though they are often associated with changes in the underlying technical architecture. They are also associated with changes in overarching paradigms for business improvement. Perhaps the key periodic driver of a change in name arises when suppliers decide to port new developments to different technical architectures – and need to sell them to their customers. Perhaps the key event in the evolution of ERP from its predecessor, MRP II, was the 1992 launch by SAP of its R/3 product based upon client–server architecture. The current debate about the future of ERP (and the idea of extended ERP or ERP II) revolves around novel technology architectures based around the internet and web-service architectures, though also accompanied by a shift in views about the search for competitiveness from the enterprise to the value-network. Software suppliers and management consultants (and latterly, commentators and business analysts, as coordinators of community expectation) appear to exercise particular influence over these changing prescriptions. A combination of factors thus appears to be at play in these shifts. We shall explore the operation of these factors in more detail by revisiting the historical development of ERP and its predecessors.

The origins of Materials Requirements Planning (MRP): techniques, technologies, communities Robertson et al. (2002) and Jacobs and Weston (2007) have analysed the origins of MRP in a small group of US industrial practitioners and academics with a background in operation research techniques and interested in applying these in manufacturing organisations through computer-based systems. These accounts highlight the coming together in 1966 of three individuals - George Plossl, Joe Orlicky and Ollie Wight22 who collaborated in developing the conception of Materials Requirements Planning (MRP), and later became widely known as ‘MRP gurus’. Initially their emphasis was on the application of production planning techniques within firms rather than software – the systems depended upon computers but were developed in-house or provided as 50

bespoke systems, building upon existing stock control systems developed in the manufacturing of complex assemblages in vehicles and aerospace. This informal community (primarily of management specialists, connected to or employed by firms adopting MRP techniques) had links with IBM which invested in the development of solutions resulting in 1972 in a successful packaged solution. This was the Communications Oriented Production Information and Control System (COPICS) designed to run on their new IBM Model 360 series mainframe computers. Specialist knowledge was codified, for example, through COPICS handbooks and commodified not just in artefacts but also more saliently in consultancy. The shift to packaged solutions was less marked in this period however than the movement towards the education and professionalisation of managers and consultants (Mabey 2007), subject to some certification and quality control procedures. Ollie Wight, for example, set up a network of ‘Class A’ companies and consultants applying his concepts.23

Whilst the early stages of these innovations were characterised by opportunistic encounters, later developments were pursued in a more organised manner (Robertson et al. 2002). The three MRP gurus became the core of a broader network and in particular promoted their ideas through the American Production and Inventory Control Society (APICS) through what they described as an ‘MRP Crusade’ (Clark and Newell 1993; Robertson et al. 2002; Jacobs & Weston 2007).24 In that decade today’s major ERP providers were established including SAP (1972), JD Edwards and Oracle (1977), The Baan Corporation (1978) (Jacobs & Weston 2007) and a host of other providers which no longer exist today. APICS strengthened itself, building its membership significantly through the MRP campaign – and we see the development of the technology and of its supporting institutions proceeding hand in hand (Robertson et al. 2002; Jacobs & Weston 2007). This is linked with a concerted effort of education and professionalisation of production management, through the production of textbooks and courses (Mabey 2007).

It is instructive to note that, in the first phase of MRP, the main institutional repositories were practitioners: user organisations, management professions and professional associations. A market was also being built for knowledge-based products. We also note a familiar pattern as with other innovations: the establishment of a division of expert labour and the partial convergence of knowledge in specialised 51

supply. During the 1970s, we see the increasing influence of management consultants and of technology suppliers (and by the 1990s educationalists, suppliers and consultants made up over 40% of APICS members [Clark and Newell 1993]). The MRP crusade seems to have been successful. By the early 1970s only 150 US companies were using MRP techniques, a figure which rose to 750 by the mid 1970s (Robertson et al. 2002), and which by the 1990s had risen to over 60,000 (Mabey 2007).

From MRP Materials Requirements Planning to MRP II Manufacturing Resource Planning SAP launched its highly successful R2 software in 1978 (Jacobs and Weston 2007). New systems emerged that could run on cheaper minicomputers such as the IBM System 38 that were affordable by smaller firms. In the early 1980s, the progressive extension of the functionality of Material Requirements Planning systems inspired a search for a new name. Jacobs and Weston (ibid.: 360) describe how:

Ollie Wight began calling these new systems ‘Business Requirements Planning’ only to find that this name had already been registered as a trademark. So he referred to them as ‘MRP II’ systems, which by the late 1980s, was ‘translated’ as ‘Manufacturing Resource Planning’ to distinguish this new capability from the original, simpler, system.

An alternative account suggests that Wight was happy to stick to the MRP acronym with which he was closely associated.25 However MRP and MRPII are not wholly distinct: at ‘the heart of any MRP II system was the fundamental MRP logic, now typically re-written in modern code’ (ibid.: 360)26, and there was often ambiguity in discussions as to what was being referred to (MRP or MRP II?).

Computer-Aided Production Management: the vision that was not sustained In the late 1980s, we find a new terminology being introduced in the UK of Computer-Aided Production Management (CAPM), kicked off by a report published by the Institute of Production Engineers (Cork 1985). The ACME Directorate, the Science and Engineering Research Council’s Application of Computers in Manufacturing and Engineering Directorate subsequently launched an Initiative in 52

CAPM Research (Waterlow and Monniot 1987). In this period, MRPII systems were being strongly promoted (for example by BPICS, set up as the British subsidiary of APICS and strongly influenced by vendors) and were being implemented in mediumsized UK firms, with mixed results. In this context, we see attention directed towards overcoming some of the shortfalls of existing technology. Chief amongst these were the complexity and inflexibility of MRPII systems which made them difficult to implement successfully, particularly for smaller firms which were very different in their production environment and management systems from those in which MRP had emerged. There was also a debate about the appropriateness of these tools as vehicles for pursuing the current vogue for flexible Japanese-style Just-In-Time systems (Clark and Newell 1993; Webster and Williams 1993; Swan et al. 1999). Though MRP systems were designed to operate in stable and predictable batch manufacturing environments, there was an attempt to offer these as vehicles for Just-In-Time even though these were based on very different principles and approaches.27

The availability of research funds and other support attracted a range of suppliers of MRP and MRP II and related systems, which were motivated to rebrand their diverse offerings as CAPM. The CAPM terminology acquired a certain stability in academic writings particularly in the UK and in some associated research centres; but did not catch hold in the USA and was not sustained.28 In 1990-1 when our investigation of CAPM was concluded, there was no clear sense of the future of MRP.29 As suggested at the outset of the book, the keynote paper for a high-level European Workshop on the future of MRP (Wijngaard 1990) noted three competing scenarios: the evolution of MRP offerings by current large suppliers; partnerships of MRP suppliers and users; and the emergence of Factory Management Systems offered by systems integrators rather than today’s generic MRP suppliers. The workshop emphasised reducing software complexity for the user (not necessarily for the designer) through context specific systems.30 Though there was no clear vision in 1990 of where MRP was going, in the aftermath of SAP’s launch of R3 only two years later, we find the sudden and all-encompassing resort to the alternative terminology of Enterprise Resource Planning developed by Davenport and industrial analyst Gartner (Lopes 1992).

53

From Manufacturing Resource Planning (MRP II) to Enterprise Resource Planning (ERP) The factors underpinning the shift in terminology from MRPII to ERP seem rather amorphous. On the one hand, there was the widely expressed dissatisfaction with earlier MRPII systems regarding the shortcomings identified in discussions of CAPM (Maskell 1993). On the other, early ERP systems were not distinct from MRPII. Gartner’s key paper that coined the term ERP and proclaimed it as the new paradigm was entitled: ‘ERP: a vision of the next-generation MRPII’ (Wylie 1990).31 Definitions of ERP, despite the ambiguities we reviewed in Chapter One (Klaus et al. 2000), generally emphasise the addition of accounting and Human Resource Management functions onto the core MRPII systems (see for example Chung and Snyder [2000]). However, these were already a feature of MRPII systems in existence and being implemented in the early 1990s (Kampf 2001). These extensions did however take these systems beyond the manufacturing process; and the Manufacturing Resource Planning terminology no longer seemed appropriate.32 Perhaps the one clear technical feature is the adoption of client server architecture offering cost and implementation flexibility advantages (Lopes 1992; Knutton 1994), coupled with a high level of integration between modules whereby a transaction on one module would be accessible to all the system modules, achieved through an integrated database.33

After Gartner coined the term ERP, other players (most notably vendors and consultants) began to flesh out what ERP was and how it worked, followed by adopter accounts of the organisational benefits of its adoption (Wang and Ramiller 2007). The theme of integration and process orientation seems to have been very attractive to corporate managers, which appeared to dovetail with ideas about good industrial practice in the context of the prior espousal of Business Process Redesign, together with a view that packaged solutions represented the way to achieve these goals (Deloitte and Touche 1997).34 Wang’s (2007b) analysis of the (mainly trade/practitioner) press points to the explosive growth in discussion of ERP in 1997-9 coinciding with a corresponding reduction in papers on BPR. He concludes that the ERP community has benefited from recruiting members and attracting attention from

54

related innovation communities hitherto looking towards BPR and MRP (see Figure 3.3).

FIGURE 3.3 NEAR HERE

De facto what seems to be driving the shifting terminology and the new paradigm is the success of SAP’s newly launched R/3 system (launched in Europe in 1992 and in USA in 1995) which, along with other similar offerings, establish themselves as a standard (Lopes 1992; Davenport 1996; Pairat and Jungthirapanich 2005). Existing MRPII users were early adopters of these new packages, and SAP in particular became widely adopted by global organisations. Added to this, organisations such as APICS promoted ERP (Berchet and Habchi 2005). However, in contrast to the previous stages in the growth of the MRP (the so called ‘MRP crusade’), the institutional basis for promoting ERP had already been put into place. As we saw in Chapter One, a range of factors reinforced the adoption of ERP, leading to the explosive growth in uptake, which was initially within manufacturing, but was progressively extended to other areas as versions of these offerings were created for process industries and various service sectors.

From Enterprise Resource Planning (ERP) to Extended Enterprise Resource Planning (ERPII) While the shift from MRP II to ERP appeared seamless, its further development has been subject to much debate. We already noted how a number of inter-locking factors have underpinned calls for radical repositioning:

i.

technology push: in the form of radical new technical concepts; in particular of web-service architectures, and software as a service, and component-based architectures (Kumar and van Hillegersberg 2000; Markus et al. 2000);

ii.

market pull: the saturation of its core markets provoked a search for additional markets (notably in smaller firms for whom ERP had been to expensive and cumbersome) and for new value opportunities in linking ERP with related technologies concerned with decision support, customer relationship management, and supply chain support; and 55

iii.

new business concepts: notably the imputed shift from an enterprise to a value chain perspective (Moller 2006).

Industry analyst Gartner declared ERP dead in late 2000 and proposed a wholly different vision under the label of extended ERP or ERPII. The changes involved are summarised below in Table 3.1.

TABLE 3.1 NEAR HERE

Gartner’s death sentence has been shown to be premature. Rather than the radical reorientation of ERP proposed by industry analysts under the label Extended ERP, we find an incremental development described elsewhere as ‘ERP 1.5’ (Judd 2006). This involves the adding on of functions rather than Gartner’s vision of a wholesale shift towards componentisation and a value-chain focus (Light et al. 2000; Jakovljevic 2001; Pairat and Jungthirapanich 2005).

Looking at this brief review of ERP and its predecessors we note a constant pattern of incremental development, progressively extending the scope of integration, accompanied by more radical re-presentations of the technology. Though each of these transitions has been subject to historical contingencies we can see some overall homologies in these innovation processes. We note the discontinuities occasioned by a coincidence of changes in Business Prescriptions and in technology supply strategies (in relation to both the development of specific application elements product and in the overall Technology Architectures of the product). We find an interesting pattern of linkages between classes of technology/their nomenclature and managerial prescriptions of best practice and broader visions of business improvement. We can observe stable linkages for example between ERP and the idea of process improvement. We can also find instances in which looser, more opportunistic and ephemeral couplings are made (between JIT and CAPM or between ERP and e-business). Finally, we note the longer-term intertwining of more general conceptions of technology and organisation; above all regarding the evolving concepts on the one hand of information integration and on the other of process integration. Thus we found the linking of the idea of the flexible customer-oriented 56

firm and CAPM and broader visions of Computer Integrated Manufacture (Webster and Williams 1993) and again today linking ERP systems and BPR.

The timelines for the Evolution of ERP points to something like a 10 year cycle: a periodicity which (doubtless not coincidentally) broadly approximates to the replacement cycle for corporate information infrastructures (see Figure 3.4).

FIGURE 3.4 NEAR HERE

As well as these cyclical processes, there is also a clear progression in the process of building a technology and associated organisational and institutional framework. At the outset, the new institutions of information technology and business improvement were rudimentary and inchoate. The early accounts emphasise improvisation and chance encounter. As time progresses, we find the gradual laying down and elaboration of technical frameworks in terms, for example, of the sedimentation of component technologies and their incorporation into solutions. The first stock-control and MRP systems were written from scratch (to work on whatever mainframes were used by the firm concerned). Later we see the recycling of code. In other words, porting functionality from the particular technical and organisational context in which it arose to other areas.35

Importantly, over the decades, the institutional frameworks for promoting enterprise system become better established (evidenced by the shifting terminology of a crusade (1972) to a movement (2005)). We suggest that the weave of the ‘socio-technical fabric’ changes over this period from a loose and open weave to a denser and more intricate pattern. Finally, we can reflect upon changes in the processes of assessment of technologies in the course of procurement. Thus, in the 1990s, consultancy organisations were beginning to collate information about supplier offerings, whilst by the twenty first century we find a much more elaborate system of consultancy and advice, and the emergence of specialist industry analysts making available community experience on a more commodified basis, and providing the basis for more formalised and systematised assessment of particular vendors and their offerings.

57

What’s in a name? This analysis has pointed to technological continuities between ‘different technologies’ or to, be more precise, different terminologies applied to a class of broadly similar artefacts. What is at stake in these classifications and reclassifications? The name applied to a technology is far from trivial. It proposes boundaries that link a class of often quite various artefacts whilst differentiating them from others. As we shall see, the designation of a technology field reduces uncertainty for adopters and for developers:

i.

it allows adopters to develop a generic case for particular innovation pathways (based upon an analysis of the potential performativity of that class of technology for certain types of organisational challenge), and, once this is accepted, paves the way for a comparative analysis of the relative advantages of particular offerings for their specific organisation;

ii.

the designation of a technology draws boundaries around a set of artefacts and their suppliers, and thereby creates a space in which some ranking may be possible; and

iii.

it allows developers to assess their offerings, their promotion, and enhancement in relation to the features of broadly comparable products and their likely future development trajectories. In addition, we see a clustering of offerings that may serve to reinforce expectations about what functionality should be included and where the technology will go in future.

1 There has been a certain theoretical convergence of view between a number of analysis – despite competing terminologies – around the study of a range of supply-side players and users, consultants and others involved in ERP, conceived as a community (Koch 2007), as a movement (Grabot and Botta-Genoulaz 2005), or innovation community (Swanson and Ramiller 1997). 2 Here we discuss technological changes such as ERP adoption though the observation could equally be applied to studies of the introduction of a new management technique – for example BPR. 3 Al-Mashari (2003) has produced a taxonomy of the topics addressed in the ERP implementation literature. 4 See for example the Journal of Strategic Information Systems that recently published a special issue on Understanding the Contextual Influences on enterprise systems in 2005. 5 Many of these shortcomings are relevant to other research into technology and work organisation.

58

6 Moreover, the process of technology development is a key interest for researchers from Technology Studies, but not necessarily for those from Business Schools addressing technology implementation and change management. 7 As a result, almost the only exceptions to this generalisation about the absence of studies addressing design and use in tandem are cases in which design was organised within the context of the user organisation (e.g. Mackay et al. [2000]; Williams et al. [2005]). 8 This included the ‘IBM segment’ (which was often in partnership with local developers), the ‘SAP segment’, the ‘PC segment’ (a vendor named ‘Damgaard’), and the ‘self development segment’. 9 BPICS, the British Production and Inventory Control Society, emerged in 1975, having been a subsidiary of the American Production and Inventory Control Society since 1963, established with support from the American management consultants Arthur Anderson (Clark 1990). In 1996 it became the Institute of Operations Management http://www.iomnet.org.uk. Whilst APICS was comprised primarily of practicing production managers, these constitute only half BPICS membership – the remainder comprising consultants and suppliers who exercise considerable influence over its activities (ibid.). 10 These difficulties in applying Jorgensen & Sorensen’s (1999) development arena concept to organisational technologies highlight differences between industrial ICT applications and consumer electronics (the sector in which they conducted their study). In case of established organisational technologies like ERP, we find close coupling between suppliers and their users (and innovation is a joint process). Workplace technologies have relatively local audiences concerned with industrial improvement. By contrast, in the development of consumer ICT platforms, commodification involves a more separate development phase, which, especially for network technologies, may involve prior collaboration between suppliers and complementary product providers in standardisation (Williams et al. 2005). 11 Effective research design requires the researcher to consider these trade-offs in focus (for example between breadth versus depth of field) and clever ways to manage these trade-offs effectively. 12 Our use of this idea of research geometry does not imply that we can apply simple triangulation (in the sense of trigonometry or a physicist tracking an object in space); to the contrary, our choices about (the multiple) points of access to a phenomenon shape what we may observe. In addressing variability of viewpoint, we rather make explicit the need for complex triangulation that is aware of the ‘parallax effects’ that may result from particular points of insertion to a phenomenon. 13 Similar points have been made in relation to the selling other non-material goods – viz the change management techniques of Business Process Redesign (Williams 2000). Similarly, Benders and van Veen (2001: 283) in their discussion of the uptake of Business Process Redesign note ‘the importance of considering the process in which management ideas gain “good currency” within the system of knowledge supply’. Their work for example mentions the importance of an idea having wide interpretive flexibility. That is, that it can invoke meaning to diverse players, as well as establishing a standardised terminology. 14 And, as already noted, different tools (concepts, methodologies) may be needed to address these different (macro-meso-micro etc.) levels. 15 Hyysalo (2004) draws on Braudel’s (1995) study about the speeds of change in Mediterranean history, to identify longer time frames that may be at play beyond that of immediate action: these are the long-term change of societies and mentalities, and the much slower periodicity of geographical change. 16 It might be argued that perhaps the object of Kallinikos’s (2004) critique might be more precisely characterised as ‘atomistic interactionism’ rather than the more ambiguous term of constructivism, which has been applied in many different ways. 17 Goffman (1974) addresses these issues from his action-centred framework through the concept of frame analysis. He distinguishes a number of concentric frames, conceived as different layers or laminations that frame a ‘strip of reality’. The frame is thus moveable – and is conceived by Gamson (1975) as analogous to the zoom lens on a camera – an analogy that we find very useful with its potential implications of breadth of field and depth of focus. We wish to raise two questions: first, about how adequate the tools are for examining the framing process; and, second, if these tools are weakly developed, what are the risks of overemphasising the immediate local construction over the broader and longer-term structuring. 18 Though we note with regret that the ‘toolbox’ metaphor has often been used as a device for eclecticism – as if it were possible to simply combine tools on a pick and mix basis without interrogating their presumptions. To the extent that we are arguing for some degree of eclecticism, rather than thoroughgoing eclecticism, where ‘anything goes’, we are proposing a critical eclecticism which interrogates the different evidential bases and presumptions of theories and tools. 19 We note here that the Sociology of Scientific Knowledge, in its critique of the positivist model of science, tends to invoke, as the interpretation that needs to be refuted, this kind of mechanistic model of the scientific process: an account of scientific

59

method that provides epistemological guarantees. To our mind, this misconstrues the scientific project. This kind of absolutist search for truth, as Bloor (1999) points out, is a feature of theological propositions rather than scientific truths. If the alternative is theological absolutism of course, we must argue for relativism. However, we argue the need not to abandon the search for truth behind radical relativism but pursue more complex understandings of truth (see Restivo and Croissant 2007). 20 Thanks to Sampsa Hyysalo for bringing this point to our attention. 21 The potted histories of the evolution of ERP all agree on the succession of technologies: MRP, MRPII, CAPM (for UK and associated scholars), and finally ERP. They also agree about the core functionalities that have been accumulated. Intriguingly, however, they differ in their account of with which technology these particular functionalities were associated. Compare, for example, Chung and Snyder (2000), Robertson et al. (2002), Kuldeep and van Hillegersberg (2000) and Jacobs & Weston (2007). 22 Joe Orlicky and Ollie Wight worked for IBM – though Wight later set up his own management consultancy (Wight 1977). 23 Chris Turner, Managing Director, Class A Limited, interviewed by Juliet Webster 6 July 1990 under ESRC PICT study of Organisational Shaping of Integrated Automation. 24 IBM sponsored the production and distribution of a series of videos made by Plossl, Orlicky and Wight to explain the concept of MRP that were made available to APICS members and their companies (Robertson et al. 2002: 14). 25 MRPII ‘It’s a kind of umbrella title isn’t it. I mean it’s almost a recognition that we’ve got all of these different tools out there and we need to put them all in a tool bag…Although it was called Manufacturing Resource Planning, it wasn’t just for manufacturing… I don’t think it’s bad in terms of wanting to get over a common message; the problem is that they’ve all got a different idea of what it means. We’ve never got really worried about that…I personally wish it had a different name. We’ve often said amongst the group it should have been called something different…Even if you just called it Business Resource Planning (you know ‘manufacturing’ is the wrong word). It all started from this fact that it was called Material Requirements Planning and Ollie [Wight] basically established that as a trade name and he wasn’t going to let that go and resource planning was very good; manufacturing happened to be the only word that fitted. So that’s how it came about’, Chris Turner, Managing Director, Class A Limited, interviewed by Juliet Webster 6 July 1990. 26 The development from MRP to MRPII was associated not only with the addition of extra-functions, including especially the adoption of ‘closed-loop systems’ driven by changes in production progress, but also with changes in the underlying IT infrastructure (notably the shift from mainframe to minicomputers). Indeed hardware and operating system features predominate in discussion of differences between different offerings in this period (Juliet Webster interview with Anderson). 27 For example in 1991, a nationwide series of BPICS seminars was organised on MPS, MRP & JIT Today, claiming that ‘The benefits of good forecasting integrated to master scheduling, in turn driving Material Requirements Planning is a pre-requisite to a good production control. It is also a vital ingredient for a successful “Just-In-Time” implementation’ (BPICS 1991). 28 The lack of longer-term impact of this effort to promote and enhance CAPM highlights some of the difficulties surrounding attempts by public policymakers to intervene in dynamic processes of technology development – that the target of intervention may be by-passed by broader developments in the field (Williams and Webster 1993). 29 For example the Association Francaise de Gestion Industriel held a conference (Paris 21 November 1990) entitled ‘Three Approaches to Production’ involving a debate between the main proponents of three competing approaches to manufacturing improvement: Georg Plossl (‘father of MRP’ with Oliver Wight), Jamashim the exponent of JIT and Eli Goldratt whose Theory of Constraints proposed an alternative approach to MRP/MRPII based on eliminating bottlenecks. The emphasis was on competing models of business organisation rather than technology artefacts (Webster 1993). 30 The Association of Logistics Management in Holland and Eindhoven University of Technology, European Workshop on the future of MRP (The Hague, 21 – 3 November 1990). Different scenarios for the development of MRP were sketched out in a background paper by Wijngaard, ‘Beyond MRP – MRP and the Future of Standard Software for Production Planning and Control’. The UK Department of Trade and Industry, influenced by these views, subsequently launched the Manufacturing and Planning Initiative to promote cooperation between suppliers and users in developing applications. 31 The roots of ERP in the earlier discourse of MRP/CAPM as a vehicle towards Computer Integrated Manufacturing are revealed by the observation that the Gartner Group’s April 1990 internal report espousing ERP as a ‘vision of the next-generation MRP II’ (Wylie 1990) came under the generic heading ‘Computer integrated manufacturing’, a concept that has almost entirely disappeared from later management or technology discourse (Mendham 2003).

60

32 Discussions about CAPM had flagged the shortcomings of the existing terminology of Manufacturing Resource Planning (MRP II), Chris Turner notes: ‘even if you just called it Business Resource Planning. You know, manufacturing is the wrong word’. Chris Turner, Managing Director, Class A Limited, interviewed by Juliet Webster 6 July 1990. Given that this particular term was not available, the substitution of Enterprise for Business is perhaps unsurprising. 33 In 1990, Georg Plossl outlined a vision of the future of MRP II as ‘not MRPIII’ but ‘distributed processing…networks of computers working…from the same basic file of data… Better networks of information within the organisation and therefore smoother flow of data’ (Plossl 1990). 34 In the same way that MRPII/CAPM were promoted as a vehicle for Just-In-Time, despite their contradictory principles underpinning, we see ERP being proposed as a necessary back-end for e-business (Norris et al. 2000) (though this attempt to couple the two evaporates when the dot-com bubble bursts). 35 Though it should be noted that even as late as 1990, MRPII and CAPM systems available in the UK were typically linked to, if not provided by, particular minicomputer or mainframe suppliers.

61