|Subject Terms:||Knowledge management|
|Classification Codes:||9175: Western Europe|
9130: Experimental/theoretical treatment
8100: Financial services industry
|Copyright Walter De Gruyter & Company 1999|
As Kay (1993) has pointed out, the exact reasons why some companies have been successful in managing technological innovation and change remain at least partially opaque, otherwise every company would have been as successful as 3M, Glaxo or Toyota. One reason for this opacity is that organizational and human resource aspects play a significant role in the successful management of technology and its implementation within the firm. The challenge to any modern organization is, as Drucker (1990) has observed, the `integration of specialised knowledges into a common task' (p. 4). However, the flowering of expertise through teams and individuals presents a challenge to the organization because the development of expertise raises the marketability of the individual or group to external agencies, rendering the organization strategically vulnerable to the threat of exit. Typically, organizations influence the inputs (training, information) that employees receive in the hope of benefiting more or less directly from the outputs (skills and expertise). An organization that invests in expertise may, however, find that the returns from that investment may not be fully captured or easily appropriated. In this paper, we argue that management needs to be aware of a number of typical appropriation challenges and adopt suitable policies for addressing them.
A diverse, specialized knowledge base poses a challenge to management because as Williamson (1985) and others (Willman 1991; Kay and Willman 1991; Scarbrough 1993, 1995) have emphasized, problems of appropriation arise when exchange-specific investments are required by one of the contracting parties. In this paper, we explore aspects of appropriability that arise in the context of technological change, and within this, we focus on the active role that managers can play in enhancing and protecting the organization's knowledge base. Physically, if an organization purchases a computer system, then one can say that the organization effectively owns (has title to) the system. The organization per se can control the computer's location, its inputs (programmes and databases) and outputs (reports and actions). A similar level of control cannot be applied to the computer's human operators or those charged with its implementation. Since the organization (slavery excepted) does not enjoy complete property rights over employees, absolute control over inputs, outputs and location cannot be asserted. Employees, at the end of the day, have the option of walking away from the organization. External consultants present additional problems because here the organization may not understand or properly evaluate the inputs and outputs that it is receiving. Consultants, even more so than employees, have every incentive to utilize their knowledge strategically to build in dependency.
Contingency theory suggests that rising environmental complexity finds its correlation in increasing organizational complexity and differentiation (Galbraith 1973). Sources of environmental change include technological innovation, the move to large-scale computerization, for example, or institutional change such as deregulation or privatization. The organization will try and tackle problems in its `relevant environment' by securing the services of intermediaries or experts such as computer consultants, engineers and lawyers (Scott 1981,1990). However, as organizational processes have become increasingly complex and diverse, greater management attention has had to be directed at the informational interdependence between specialities, functions and projects (Lawrence and Lorsch 1967; Tushman and Nadler 1978; Tushman and Nadler 1986; Rockart and Short 1989). This has prompted many suggestions on how the lateral flow of knowledge within organizations might be enhanced (Liu et al. 1990). McKersie and Walton (1991: 248-249) referred to the need for 'alignment' between technological requirements and organizational capabilities, because the former can only be fully realized through the latter. Pasmore (1994) suggested that organizations should be 'fractal' in design: such organizations should be viewed not so much as a series of functions and departments but rather as an inventory of talents and skills. Within fractal organizations, temporary concentrations of expertise can be created (ibid.: 225-226). In these organizations, the managers' role is increasingly one of co-ordinating different sources of expertise with open access to shared information (Webb and Cleary 1994: 30; cf. Grandori 1996) Arguably, however, the diffusion of fractal-type organizations is being slowed down because questions of appropriability are being neglected: for example, how can experts who are encouraged to engage in free information sharing with their colleagues in temporary project teams be at all sure that employing their knowledge in this way will not have any negative effect on furthering their own career? What if this sharing takes place in an overall context of organizational restructuring and downsizing? Whilst it might be true that the organizational role of integrating teams is to enable participants `to emphasise knowledge in particular areas while keeping abreast of developments in related areas' (Appelbaum and Batt 1994: 103) surely an important question is whether this is in the participants' own interest (Townley 1994).
At this point it is useful to reflect briefly on earlier research which has dealt with such issues. For example, Bjorn-Andersen et al.'s (1979) socio-technical study on job satisfaction and participation in technological change was generally concerned with recruitment and retention problems. At that time, a common management expectation underlying automation was that a computer-based clearing system would not only reduce boring clerical work but also increase the attractiveness of banking as a career (Harris and Mercer 1979). Subsequently, a number of other studies have explored these issues within a socio-technical framework (e.g. Child and Loveridge 1990; Kriegler et al. 1988). Amongst sociologists, a persistent concern since the early 1970s has been the analysis of the impact of technological change on employment levels and skill profiles (e.g. Crompton and Jones 1984; Mirris et al. 1991; Child and Loveridge 1990:112-118; Willman and Cowan 1985; Cressey and Scott 1992; Clark 1993). Authors including Evans et al. (1989), McKersie and Walton (1991), Walton (1989) and Osterman (1991) have focused on the role of work organization, training and employment security in technological change. Our own interest here is not so much in the `labour process' issues but in the consequences of certain managerial strategic choices affecting employees with skills that are scarce and/or difficult to replace (e.g. Lloyd and Rawlinson 1993).
The process of incremental improvements in information technology itself threatens organizations with atrophy and obsolescence if those relationships are not continually renegotiated to incorporate new developments. Organizations need to cope with a paradoxical situation: effective appropriation of organizational knowledge requires an environment of stability and commitment; but this risks the atrophy of the organization's effective skills base. Over time, the organization may lose the ability to draw in new knowledge from outside its boundaries, such as in the design and implementation of new computer systems. Furthermore, one cannot even assume that knowledge inputs will yield organizational benefits: employees for instance can use their newly acquired knowledge simply to demonstrate how up-to-date they are without, however, making it part of their daily routines. Also, in complex and fast changing environments, innovations may be adopted for their role as a `managerial fad' rather than for achieving real change (Abrahamson 1991); or in order to help management to make sense of their surrounding 'environments' (Daft and Weick 1984). Additionally, knowledge may simply 'slumber' in the organization - for example on the shopfloor -- without ever being utilized for organizational purposes. Therefore, acquiring a technology and training employees to use it is not enough if there are insufficient pre-existing mechanisms to retain key staff, and second, to integrate the newly acquired knowledge and skills into the existing set of routines (Kamoche and Mueller 1998). This poses various questions of appropriability that we need to discuss in the following section.
Problems of Appropriability: Developing a Taxonomy
From the organization's perspective, we can distinguish between differing aspects of appropriability that have to be recognized and controlled. At one level, the organization has to mediate between arguments that arise over how rewards should be divided between contesting rent-seeking claimants: we can label this as essentially a distributive appropriability problem. This takes two main forms: internal and external. Internal appropriability problems (IAP) arise from the organization's inability to fully appropriate all the knowledge, skills and abilities that it stores (Kay 1993; Mueller 1996). External appropriability problems (EAP) arise from an organization's inability to 'own' people, making it problematic to secure its interface with the external environment -- people can simply walk away. We argue that in contrast to the disjointed and separate discussion of these issues in the literature they, in fact, pose a management problem that requires simultaneous consideration of both dimensions.
Arguably, problems of distribution are complicated by the `ephemeral and organic' nature of organizational knowledge (Quinn 1992: 259) making it susceptible to withering and to becoming out-dated. This can be played out at both the individual and organizational level. The effects of atrophy are most visible at the individual level, for example very few gaslighters today find gainful employment. In other words, individual skills can become obsolete. Less visibly, the effects of atrophy can also strike the organization. For example, in the Public Sector several government departments found their manual processing of volume transactions increasingly outdated by advances in computer technology. We propose calling such problems arising from the up-to-dateness of organizational knowledge structural appropriability problems (SAP). According to Dierickx and Cool (1989), strategic asset stocks are threatened by decay unless proportionate maintenance expenditure is directed at nurturing and sustaining them. To put it differently, complacency and neglect of maintenance efforts can turn competencies into incompetence (Kamoche 1996). Where there is an SAP, which remains un-addressed, it is likely or perhaps even certain that all distributive claimants will lose out. Arguably, in order to perceive and address the SAP, organizations need to conceive of themselves as `open systems'. Flowing into and out of this open system are consultants and employees. In order to address the SAP, continuing additions to organizational knowledge, skills and abilities from inside and outside the organization are necessary, and organizations therefore have to manage this flow as a strategically important activity (Ginsberg and Abrahamson 1991). The desired outcome of this process is an internally coherent, tightly integrated and protected base of knowledge; the unintended outcome is often an internally incoherent, loosely integrated and leaking base of knowledge. Tapping external knowledge, however, poses the challenge of integrating the new into the existing knowledge base to combine the newly learnt with the already known (Thomas 1992: 289; Kanter 1985: 278). Clark and Staunton (1989: 12) suggest a new term, 'exnovation', to refer to the problems associated with the removal and replacement of existing practices by new ones. Integrating new knowledge into existing organizational routines does not take place without difficulty. Expensive consultants, for example, may be hired, only to have their recommendations disregarded and shelved away. In one of our cases, the Halifax Building Society, a newly established marketing department found its recommendation for a new cheque account resisted by longer established departments. Although difficult, in an era in which technological and regulatory change predisposes the rapid obsolescence of organizational knowledge, the active tapping and integration of new knowledge into daily routines is an exercise with high potential rewards.
However, the rapid obsolescence of knowledge is a problem not just of scale but also of scope. In practice, no organization could ever hope to tap into and integrate the full spectrum of changes taking place within its environment. Consequently, organizations have to learn to be selective in their pattern of interventions. In the financial services for example, it would clearly be important to be up-to-date with changes in the regulatory framework, but it may not be necessary to be as up-to-date with regard to the latest computer software or hardware. Technology, though, can be problematic for both academics and organizations: capital investments in tangible assets can seldom explain more than half of the resulting performance or productivity gains, the rest being a `sociological black hole'. The latter might stand for employee commitment that has been energized or it might stand for useful changes in organizational arrangements. In this vein, Appelbaum and Batt (1994) recently observed that an increasing number of service firms recognize `that office automation alone does not improve productivity and that it will be necessary to restructure and change human resource strategies' (ibid.: 100, emphasis added). With obvious human resource management implications, another prominent writer in the information technology field argued that successful process change requires not just restructuring but also `greater depth of job knowledge and greater breadth of task expertise' (Davenport 1993: 107). However, if one accepts that 'successful' usage of information technology often results from the patterns of social relations and tacit knowledge that sustain it - Kay (1993), for instance, refers to an established `context of reiteration and reciprocation' (ibid.: 77) which he calls 'architecture', but which Barney (1986) more conventionally calls 'culture' - then it becomes even more doubtful whether decisions over external aspects - e.g. `up-to-dateness' - can be seen to be at all separate from internal aspects. More specifically, where an organization does not operate in a competitive product market, such as the Department of Social Security or the Inland Revenue, an architecturally-based sound usage of a - strictly speaking -- outdated technology might be difficult to classify as anything else but a `success story'. Is our introduction of these various types of appropriation problems (APs) an unnecessary and perhaps arbitrary complication, or have these underlying issues actually arisen as 'problems' in the existing literature? The IAP has indeed been analyzed by various disciplines, including HRM, but typically only indirectly or implicitly, and never in explicit conjunction with the EAP.
There has been a gradual recognition that appropriating the added value emanating from the employment relationship is a task subject to uncertainty and tensions, especially during periods of technological change. The problem for the organization is that this added value is made up of a bundle of explicit and internal components difficult to fully specify in a contract or authority relation (Williamson 1985). There are two ways of interpreting this, one psychological and the other, organizational. At a psychological level, the explicit part of the contract is in constant tension with the implicit one. However, at an organizational level, formal organizational goals are continuously challenged by emerging informal parochial interests, which are subject to analyses in terms of organizational politics and/or industrial sociology. In more simple terms it can be put as:
`When information is the primary unit of organizational currency, we should not expect its owners to give it away.' (Davenport et al. 1992)
This is sometimes neglected by those who analyze `organizational learning systems', such as top management, informal networks, problemspecific, and department- specific OLSs (Shrivastava 1983: 21) without discussing the appropriation issues involved. To put it bluntly, innovation may be used for predominantly 'private' purposes. This might create an obstacle when specialist managers and niche professionals need to integrate their different types of expertise. An organization can never fully rely on the complete co-operation of its members, because individuals retain a degree of discretion to deviate from official goals, norms and policies (Crozier and Friedberg 1980). An organization's ability to access its employees' skills and knowledge is thus continually challenged. This is a theme that has been analyzed in remarkably similar terms by both organizational theorists and economists. Schein (1983) pointed out that increasing complexity in both technology and the economic and socio-cultural environment has made organizations more dependent upon the fullest possible contributions from their employees. In a situation of mutual dependency, a long-term relationship is in the interests of both sides, because no one wants to become subject to `hostage-taking' (Wilkins and Ouchi 1983; Williamson and Ouchi 1981; Ouchi 1980). As dependency increases, `perfunctory co-operation', i.e. just fulfilling the basic formal contract, is not sufficient, and `consummate co-operation' becomes necessary, i.e. showing loyalty and commitment on both sides (Williamson 1985). However, it is the same `external environment' which -- seemingly paradoxically -- makes the continuing co-operation of internal staff both more necessary but also less certain. Indeed, in the changed environmental context of the 1990s, the firm is confronted by a `post-entrepreneurial career logic' by which employees take their revenge on the re-engineering and downsizing organization:
`The benefits of new ideas are often retained by the person ... "Free time" for learning, creating, or inventing is channeled toward ends that enhance the person. It is uncertain whether any of this is channelled back to the organization.' (Kanter 1990: 331)
The discussion so far has highlighted the constraints faced by organizations in their ability to appropriate the added value arising from their stock of human resources. The IAP can also take the form of the organization being less than the sum of its parts, for example, due to the failure of employee involvement. After surveying much of the literature, Dodgson (1993) described individuals as the primary learning entity in a firm. Similarly, Kanter argued that organizations `know how to learn because their people know how to learn' (1990: 320). However, organizations can also be seen as knowing less than their members because of problems in communication - such as filtering and distortion problems or insufficient `channel capacity' - or because rapid environmental changes may overload communication networks (Argyris and Schon 1978; Senge 1990). There is a further dimension to the IAP: because of the difficulty of establishing full control, organizations undergoing technological change may find that their knowledge base has become politicized (Scarbrough and Lannon 1988). Lyles and Schwenk (1992) distinguished between a core knowledge framework that is widely agreed upon within the organization, and supporting knowledge elements which are linked to the core set. Different groups within the organization hold alternative, often politically competing agendas, and may try to enhance their position, for example, by means of task forces or special reports. Similarly, Knights and Murray (1992) see organizational groups as continually re-constructing technology and the market in a way that allows them to secure advantage. For example, marketing could try to influence the IT-system design in order to better satisfy the final consumer. Webb and Dawson (1991), in their research, found that there is a socio-political as well as a purely technical dimension to the problem of managing expertise. There was no `simple progression from one strategic recipe to another, but instead a continuing tension between manifest and latent formulae, lasting more than a decade' (ibid.: 203). Expertise should be seen in connection with the external infrastructure of occupational and technological forces, making senior management control problematic. In our context, the main political bargaining goes on between top management and groups which derive power from expert knowledge. One such group is that of the information system employees, who exert power due to their position within an organization and their professional allegiance (Scarbrough 1993):
' ...a distinctive feature of expert groups is the extent to which their control of certain tasks derives from access to a specialist knowledge base.' (ibid.: 941)
Whilst management typically has to 'pay' in order to access external knowledge, experts can normally do so as part of an established ongoing nonfinancial exchange relationship with outside experts. Information ghettos might thus become jealously guarded by human resources. marked by their knowledgeable expertise. Persuading such expert staff to remain in these ghettos whilst at the same time orienting their actions towards organizational goals, may then become a matter of carefully crafting career and promotion paths. This reflects that the IAP has both a collective and an individual dimension that has to be addressed by management. The EAP is partially circumscribed by the structural characteristics of the firm's assets. The main rationale governing the relationship between the firm and its `human assets' can be described (in simplified form) as follows: if management successfully raises the specificity of employee A's knowledge, then, ceteris paribus, A is more dependent upon that particular firm in order to appropriate personal added value (see Becker 1964; Doeringer and Piore 1971: 84). The appropriation conditions of the individual can be further constrained if management installs a team structure, where individual withdrawal of skill will not adversely affect the effective functioning of the team (Aoki 1984). In contrast, where the purely `contextual' element of knowledge is relatively unimportant, such as for engineers and financial specialists, it will `not (be) difficult for other firms to identify their true productivity' (Aoki 1988: 83), and mobility or exit will thus become a practical option for the employee. The EAP has been analyzed mostly in the context of high-tech firms, a sector where the loss of crucial knowledge to a competitor can mean the 'failure' of the organization. In the face of the EAP, the challenge therefore is to retain valuable organization-specific knowledge. Harrison and Carrol (1991), for example, demonstrated that persistence in building a knowledge base can be a greater problem in an organization than in society because members can come and go more quickly. A firm's boundaries have become increasingly permeable inasmuch a firm may hire experts from competitors, and clients or competitors may lose experts to them (Starbuck 1992). Furthermore, organizational learning systems vary from being highly dependent upon a single person to being highly participative - e.g. from processes of knowledge sharing revolving around key managers to facilitating learning through work teams or organizational culture (Shrivastava 1983). However, even team-based embedding of knowledge does not eliminate the EAP, because sometimes a firm's entire work or design team may join a competitor this represents the collective dimension of the EAP. Mobile employees, especially experts, pose a particular threat to the organization because they are the ones who are most likely to join competing firms, or set up firms of their own. There is some empirical work to suggest that this is an especially relevant consideration for high-tech firms (Kleingartner and Anderson 1987; Jelinek and Schonhoven 1990).
Clearly, the IAP and EAP are not independent from each other. However, there is little explicit guidance in the literature as to interdependence between the two. Measures that make effective internal appropriation difficult (e.g. a reward system that is perceived as unfair, thus motivating employees to hold back valuable knowledge) can simultaneously create an acute EAP (if employees decide to exit). Furthermore, the IAP and EAP are not always neatly distinguishable. Stata (1989), for example, pointed out that if an organization relies exclusively on individuals for learning, then, as people migrate between jobs, the organization risks loosing much of what it has learned. Thus, employees moving within the organization might jeopardize appropriation as much as employees actually exiting the organization. With the creation of mixed teams composed of both internal and external experts, the Inland Revenue attempted to address questions of IAP and EAP simultaneously and, as we will discuss later, was quite successful in this. Attempts to overcome the three main APs by codifying and trying to capture expert knowledge through hardware and software systems as well as databases (Quinn 1992) is questionable, because the fact remains that it is employees who are the most important carriers of knowledge. The discussion so far can be summarized in the following table: In this second section, we discussed a variety of appropriability problems facing companies undergoing technological change. In the third section, below, we discuss a variety of design options based on the literature such as building relational teams, raising asset specificity of expert knowledge, creating cross-functional or even fractal organization design. We focus on whether management have adopted any of these design solutions, or appropriation strategies, and also on which factors can help explain the actual adoption of one rather than another design solution.
Practical Implications and Appropriation Strategies
To secure the continuity and modernity of organizational knowledge and tackle the appropriation challenges described earlier, management can employ certain strategies or mechanisms.[Table]
Although Williamson (1985) treats asset specificity as part of an organization's objective external environment, Lazonick (1991: 200) rightly reminds us that organizations design asset specificity. These are decisions that can be analyzed in terms of a managerial strategy for appropriation. However, the empirically grounded studies that one could build upon in this context are very limited indeed (Teece 1987). Williamson recommends adopting a `protective governance structure' (Williamson 1985: 243), for example, in the shape of an enhanced career and reward system, or through the construction of `relational teams'. The defining characteristics of a relational team are that members are prepared to place the interest of the group above their own interest and that of other individuals; the existence of trust between the members; high continuity among members and a restraint on their opportunism. Organizational theorists have argued much the same when advocating `committees that cut across operating and strategic business units, technology programs common to many units, and incentives to share information across organizational boundaries' (Doz 1989: 204). Other lateral integrating mechanisms are idea generators, boundary spanners, informal communication networks or task forces, underpinned by employee development programmes (Schein 1983).
One recurring argument suggests that by creating suitable career and reward structures with privileged employment and career prospects, the contributions of selected employees will be assured, and opportunistic self-interested behaviour will be largely pre-empted. In fact, both among organizational psychologists and economists, the expectation is that employment security and clear internal routes of promotion discourages self-serving behaviour, and allows the firm to retain and access its most valued knowledge. The empirical reality, however, is more like the picture presented by Prietula and Simon (1989) who pointed out that `few organizations structure their compensation and reward systems to permit superior performers to become more valuable as they stay put' (p. 124). However, there are differences: organizational theorists tend to look more at processes rather than structures, and view employment as a bargaining process where the parties have their own interests and try to obtain the most for the least effort (Rousseau and Greller 1994). With some justification, one could argue that these two positions are not fully compatible. The first presents a static solution to problems in the labour market, whereas the latter posits a dynamic on-going bargaining relationship over the appropriation of knowledge (or more generally -- added value) (Mueller 1995). To put it differently: the formal contract of employment is essentially ambiguous and cannot cover the rich mutual expectations of employees and corporations. This implies the existence of a psychological contract but one which needs to be continuously maintained (Townley 1994). One major instrument for maintaining the psychological contract is the reward system which, as some argue, should be designed so as to encourage employees towards information-sharing with colleagues (Davenport 1994). As discussed earlier, such prescriptions are all but meaningless if they ignore the particular appropriation interests involved in any employment contracting relationship. Evans et al. (1989) pointed out that there is always a restricted number of managerial options which can work within the available design space, because '(T)he company's basic reward system has to be aligned with the types of innovative behaviour that management requires' (p.187). Specifically, this hints at a carefully defined 'fair' bonus scheme (Raelin 1985) or motivation enhancing promotion policy specifically tailored to 'experts' (Allen and Katz 1989) or installing firm-specific career paths (Hendry 1990).
The need to achieve integration between different organizational units has remained central for much of the research on organizational change. More recently, the desire for integration of- or consistency between -- different managerial policies has become more prominent. In practice, however, managerial policies often conflict. Collinson (1993) found in an insurance firm that introduced on-line processing that corporate policies were partly conflicting: a regrading scheme introduced to retain skilled employees was opposed by a corporate staffing policy aimed at stopping any further recruitment. In the Department of Social Security an incremental policy of IT adoption conflicted with a policy of radical change within the work organization. Both illustrate the practical difficulty that organizations encounter in seeking to integrate an increasingly fragmented knowledge base. In this section our aim has been to discuss some practical design strategies that management can employ in order to tackle appropriability difficulties. A clear and agreed analysis of actual business problems, for example, can facilitate cross-functional awareness and the appropriation of knowledge. However, the literature appears to suggest that structural solutions to the appropriation problem might be wishful thinking, and that management cannot escape the ongoing challenge of having to secure appropriation. This also means that continuous ongoing effort is required for the building of competencies as, among others, the Halifax case demonstrates.
Data and Research Methods
Our focus is the appropriation of knowledge to the organization rather than to individuals within the firm, or to consultants drawn from outside the organization. To illustrate our analysis, we will draw upon case material that has been generated as a result of London Business School's (LBS) `Technological Change Processes and Outcomes Within the Firm' study and funded by the ESRC/DTI `New Technology and the Firm' research initiative. The LBS project, conducted over a five-year period, brought together academics from two differing disciplines, economics and sociology, to study the interplay between organizational processes of technological change and their economic consequences. For the purposes of this paper, we concentrate upon the management of large-scale computerization in three organizations: one drawn from the financial services sector, the Halifax Building Society; and two drawn from the central government sector, the Department of Social Security (DSS) and the Inland Revenue.
Longitudinal field research was conducted in the target organizations over a four-year time period. The linking theme of the research framework was the relationship between organizational systems and corporate performance. Primary data was collected by using semi-structured interviews held with a number of representatives at all levels across the organizational hierarchy, from the Project Director to user groups. Thus the case material was generated and informed by the responses recorded from IT project staff, senior management, end-users and consultants. Two researchers, an economist and a sociologist were assigned to each target organization. A checklist of questions was drawn up for the interviews and this included key areas such as the cost and benefits of the IT project, both anticipated and actual; IT and firm strategy; implementation experience; staff development and training; and corporate performance.
All three cases were essentially engaged in implementing large-scale computerization across a national branch network. The computerization projects were known, respectively, as `Branch 2000' (Halifax), `Computerization of Pay As You Earn Tax' (COP - Inland Revenue) and `The Operational Strategy' (OS - The Department of Social Security). The Halifax successfully computerized and up-dated its branch network by essentially taking manual processing out of the branch network and into newly created regional administration centres. The Halifax has a lower budget per staff for information technology than its closest rivals, measured according to size, national coverage and functionality. The Inland Revenue computerized a range of formally manual processes in the collection, recording and processing of personal tax liabilities. This computerization took place on a national basis involving about 600 local offices, and was implemented to time, according to budget and with additional functionality. In contrast, the DSS undertook a similar task of national large-scale computerization, which came in on time but at greatly increased cost and initially reduced functionality.
In the long term, the goal achievement of our cases relates principally to the extent to which IT was effectively integrated into organizational processes. This was very much a post implementation focus. Here, through the interviews and end-user groups, an indication of the smoothness with which IT systems had been integrated was established by gauging the level of user satisfaction and the need for system rewrites. Again, we concluded that both the Halifax and Inland Revenue were generally more successful in the long term than the DSS. More precisely, both the Halifax and Inland Revenue implemented IT systems that appeared to be integrated smoothly into the operational business of the organizations; user dissatisfaction, system rewrites appeared to be missing, following major periods of computerization. In contrast, the DSS appeared to have an IT system that was poorly integrated into the operational business of the organization, marked by high levels of user dissatisfaction and costly system rewrites. Two measures of long-term success were deployed: the retention of IT staff and the recorded use of external consultants.
The project adopted an exploratory research design that encouraged contributions from a number of diverse disciplines. To us, it appeared that one of the main problems with the existing literature was that different strands in the literature dealt with issues of external appropriability and internal appropriability. A related problem is that research often has to rely on simplified models, whilst analyzing complex interactions. Mindful of Van De Ven and Astley's (1981: 455) warning against `conceptually vacuous and empirically incorrect' unidirectional causal models of organization, the project focused on the extent to which appropriability questions resulting from the implementation of a new IT system enhanced or hindered the organization's goal achievement. Empirically, a distinguishing factor between the organizations appeared to be related to the effective management or otherwise of the technology-organizational interface. In particular, a strong appropriability of technological expertise by senior managers rather than by individuals within the organization was associated with goal achievement.
What type of conclusions can be drawn from such a limited sample? The methodological role of cases depends partly upon the nature of the predominant argument in the relevant literature(s). As was argued earlier, there seems to be a disjointed literature making strong propositions about lAPs on the one hand, EAPs on the other hand. What even a small sample can do is reject a proposition. This is, in fact, what we see as the primary methodological status of the case studies used.
Case Background: Strategic, Institutional and Technological Changes
The past two decades have seen a gradual change in the way financial services and the public sector are administered. Branch based, manual operations have given way to highly automated information processing. Large-scale computerization has liberated branch office staff to concentrate on offering service and client care rather than manual processing and data recording. Nevertheless, across such service-orientated organizations, IT has been adopted with patchy success. For various reasons the 'return' on the implementation of technology in both the financial and public sectors seems to have been disappointing. Underpinning these differences has been a long ranging debate concerning the appropriate organizational strategy for the adoption of new technology in financial services. Very much an `outsider, in the sixties and seventies, the IT department during the eighties has been drawn into the strategic core of organizational functions (Scott Morton 1991) At its crudest, this debate can be effectively summarized as to whether to contract out systems development or `do it yourself internally. In the 1970s and early 1980s, when the over-riding objective of computerization was to reduce costs, the balance of opinion favoured the use of consultants as a quick fix. However, since the more disillusioned latter half of the 1980s, fuelled by a more competitive environment, the emphasis has tended to swing from short-term cost reduction to the more longterm service oriented innovation.
Over much of the Post War period, the organizational challenge for largescale service orientated organizations, such as in financial services, has been the maximization of internal efficiency at a time of low levels of environmental (technology, market, regulation) uncertainty. This environment became more turbulent in the 1980s when structural appropriability problems in the form of changing regulatory and technological factors pressured organizations to guard against obsolescence. Whereas in 1980 it was still possible to characterize the British domestic finance sector as two distinct entities -- building societies concentrating on the provision of mortgages and banks focusing on the provision of cheque accounts and loans - by 1988, banks and building societies were directly competing against each other in the provision of mortgages, cheque accounts and loans. Central Government departments have encountered similarly turbulent environments. Although this has not been directed at introducing competition per se, moves towards greater accountability, the hiving off of hitherto large and unified departments into smaller, semi-autonomous, agencies and the introduction of market testing has emphasized the necessity for strict cost control and enhanced client awareness.
At the same time, technical change in the form of information technology and evolving computer technology, the use of on-line systems and the introduction of networks, has also changed the environment from one of staid certainty to that of uncertainty and threat. Yet while technology has contributed to the more unstable environment, technology per se has not been the basis of competition between organizations in the financial sector. The hardware side of technology cannot provide a sustainable competitive advantage because similar computer hardware and software can be bought off the shelf as it were, or from consultants. Any competitive edge derived from having the latest computers would be rapidly competed away. Further, networked services, implying commonly agreed standards in hardware and software to ensure equality of provision, conceivably dissipates the opportunities for lasting competitive advantage through technology (Scott Morton 1992: 270; Teece 1987).' As Capello et al. (1990) have noted, it is not the hardware of the network that provides competitive advantage, but rather the way in which information is used. Ettlie and Reza (1992) argued that, given that the technology underlying process innovations is generally available for purchase on the open market, organizations have to employ firmspecific integrating mechanisms in order to capture the value from process innovations. How information and skills are used by an organization is largely determined by its ability to tackle problems of appropriability as outlined above.
Barras (1990) has suggested that the adoption of IT in financial services most closely resembles a reverse product cycle, in which process technology innovation precedes product technology innovation. This sequence is largely a function of time and growing user experience. The potential of IT for the development of new services is not fully exploited at first; rather the more constrained objective of reducing the provision cost of existing services is sought. In our cases, the move towards large-scale computerization improved efficiency by reducing staff numbers. As organizations gained competence on the first generation of computer systems, they began to identify the scope for further, more customer-oriented services. In effect, such organizations were crucially aware of the structural appropriability problems discussed earlier and took steps to guard against obsolescence. Changes in branch design to facilitate sales - the designation of branch staff purely as service specialists and the elimination of back office functions - have all followed computerization as organizations have added to their knowledge base.
However, the creation of firm specific competencies can create problems for managers within the organization. Simply put, the tacit and team-embedded attributes that make such competencies problematic for insiders or outsiders to appropriate can also lead to the creation of internal ghettos of competencies, from which senior managers are jealously excluded. Under such circumstances, these ghettos can enjoy a strong bargaining position within the organization that can raise costs through higher wage settlements and impede strategic development. This threat is not wholly internal. IT consultants can also establish a strong bargaining position which may be translatable into higher contract settlement costs and strategic lock-in. These issues will be looked at in more detail in the following two sections.
Managing Expertise Internally: The Halifax Case
One of the most successful organizations in our empirical portfolio is the Halifax Building Society, the largest and most consistently profitable building society in the post-war period. Created in 1852 in the Oak Room of the Old Cock Inn in Halifax, the Society has grown to become Britain's largest building society, with total assets of some 58,710m and nearly 1.8m borrowers by the early 1990s.2 Unusually for the sector, the Halifax has grown to its present size almost entirely from internal resources. In contrast to most other building societies, the Halifax has never expanded via the merger or takeover of a rival. Importantly, the Halifax's organizational routines have never had to respond to submerged internal challenges as a result of a merger or takeover. This has meant that the Halifax has enjoyed an internal consistency that few of its immediate competitors can begin to match.
Interestingly, such competitive distinction has not meant being ahead of the pack as far as newness of technology is concerned. If the price for high consistency was relative obsolescence it has clearly not resulted in complacency. The Halifax's IT architecture has ev,lI (1 adapted to market and technological challenges. Twenty years ago, for example, only a handful of staff would have been concerned with computer technology, now the function attracts several hundred staff members. Equally, twenty years ago, money transactions would have been organized and processed manually at branch level, now most of the process of money transmission has been automated and computerized. Indeed, most of this process, the clearing of cheques for example, has recently been taken out of the branch office and relocated in out-of-town computer 'factories'. Cressey and Scott (1992) describe a similar process of concentration of back office functions into regional centres for similar organizations such as the TSB, Midland and the Cooperative Bank (ibid.: 89).
Quicker than most building societies, the Halifax recognized that changes in the regulatory environment threatened the obsolescence of its organizational knowledge base. To this end, in 1988, the Halifax announced the development of the Branch 2000 concept, replacing old style branch offices with an open plan, inviting layout, complete with reception points within which customers could browse and be counselled in appropriate services by trained staff. Changes such as these reversed the traditional 'iceberg' structure of the branch office, where most of the physical space was dominated by hidden, back office functions. Large-scale computerization ushered in an administrative restructuring that moved paper and back office processing from branch to regional level. The success of the exercise reflects Halifax's position as one of the most successful adopters of IT in the financial services sector. Here, IT has been treated as part of a more encompassing architecture which requires gradual and incremental development, but one that can be irreparably damaged by the externally-motivated' imposition of fashion-driven fads.
The Halifax's IT strategy has been to add to its existing technical capabilities based on business need, rather than haphazardly redefining itself according to the latest technological trend. In our terms, their approach has been to give priority to addressing the distributive over the structural appropriability problems confronting the organization. The antecedents to this approach was a strategic rethink in the Seventies, which culminated in a direction that gave clear priority to the delivery of services to the customer whilst minimizing technology costs. Lower costs were partly achieved by encouraging competition in hardware supply through the early adoption of a multi-vendor policy. More subtly, the incremental approach adopted by the Halifax was also designed to contain the distributive appropriability problem presented by technical staff, who were at that time in short supply. This was to be solved through human resource policy and computer technology forming compatible parts of` an encompassing consistent architecture.
As Figure I indicates, all of the Halifax's IT development work, and responsibility for the day-to-day running of existing systems, is organized within the computer division known as the Business Information Systems (BIS). Technological expertise chiefly resides in this division having developed from its original role as the Halifax's computer department in the early 1970s. Two main groups largely make up BIS: programmers drawn from school leavers and the local population, and business analysts drawn from the Society's branches because of their practical experience. Since the strategic rethink that took place in the 1970s, almost all system development work in the 1980s was constructed through the use of an IBM Assembler. The internal, three-year, training course that programmers undergo is unusual for its concentration on this low-level programming language.3 Curiously archaic by the standards of modern third and fourth generation languages, the Halifax has deliberately chosen to retain this effectively obsolete language as the mainstay for its development work and since the mid seventies all of the Halifax's large-scale computerization applications have been developed using this language. Compared to modern methods, writing programmes using the IBM assembler is time consuming, and relatively expensive in staff costs. However, these potential disadvantages have been mitigated by the Society's extensive code library, built up over the years, which can be redeployed in a modular fashion to future development work. Internal consistency and skill-base continuity has thus received priority over market-driven updates of-hardware and software - aspects of technology. Thus, whilst the technology might be archaic in a marketing sense it is, nevertheless, an element within an architecture that appears to be functioning very well.
The decision to retain the IBM Assembler as a development tool - a decision periodically reviewed by the IT Department - has not hampered the Halifax's market appeal in the eyes of the consumer. The pace of systems development at the Halifax has been neither slow nor costly. One example of this is in the development of the SWITCH network, a network that the Halifax successfully had up and running some six months before its consortium partners of banks and building societies. Similarly, the Halifax has been one of the first in the market in the UK to use a computerized mortgage service. Helped by the `Branch 2000' Project, the Halifax has been successful in projecting to its customers an up-to-date image, paradoxically underpinned by a programming language that its competitors have long since abandoned. However, aside from the structural appropriablity problem, the organization has also had to formulate an appropriate response to the distributional appropriability problem, both internal and external. The Halifax has approached this by utilizing three main mechanisms: technology, reward systems, and location.[Chart]
Technically, the significance of the decision to retain the IBM assembler as the core around which development is organized has helped to produce a high level of staff continuity within BIS. Average staff turnover is less than 5 percent, lower than the average at Halifax itself (see Figure 2 for the situation in the second half of the 1980s). Requiring BIS personnel to train and use this, by today's standards, archaic development language effectively renders Halifax's technical staff less marketable to other institutions, thereby `locking in' these programmers. This helps to prevent both internal and external appropriability problems from arising: internal staff find it difficult to leave, and would-be external poachers find it difficult to quickly assimilate the accumulated experience housed in the Halifax's library of computer routines and sub-systems. Technical decisions to adopt a specific programming language as a way of retaining staff and so avoid the external transference of organizational knowledge have thus been supported by deliberate human resource practises.
Of course this process of lock-in can work both ways. Having invested three years in training staff in its own idiosyncratic ways, qualified personnel become more important to the Society than to other employers. From the Society's point of view there are high opportunity costs attached to its trained technical staff. In these circumstances, staff can have an incentive to act strategically, by under-performing for example, because of a perception that they have 'a job for life'. These inclinations can be reinforced by personnel policies based on routine increments in pay and position largely divorced from actual performance. BIS successfully won control for their own personnel over selection, finance and training to introduce performance-related pay and promotion policies, formulated on the longterm success of the division.4 Compared to the rest of the Society, BIS became a hot house of `tremendous internal competition' geared towards the performance of the systems.[Graph]
The insular approach results in promotion taking place almost wholly internally at Halifax. Almost all senior positions within the computer division are filled by managers who originally started either as programmers or system analysts. BIS's project `Technical and Training' managers for example, can all trace their positions back to the setting up of the original computer department in the 1960s. Continuity is also reflected at executive level, where, for example, one executive can trace his career from mortgages to programmer to project manager, from BIS director to operations executive, over a thirty-year period. Such continuity and longevity among IT professionals is unusual in a labour market that experienced a shortfall in skilled technologists over much of the eighties. Barclays for instance lost twenty of its senior IT managers, including the two top positions in 1990.
Further, the Halifax's incremental approach built around a consistent software base also provides the organization with scope to control for opportunistic action. Since the computer systems in development today use `the same building block' as the computer systems of the past, senior managers have little difficulty in keeping up-to-date with current development projects. Unlike their rivals in competitor organizations, managers are not at a knowledge disadvantage with respect to their staff. This helps to constrain the development of appropriation problems within the IT department. Establishing in-house skills around a technology that is somewhat archaic helps to curb the technological enthusiasms of IT specialists and so helps to avoid the pursuit of sub-goals within the organization. As the Technical Controller explained:
`There are no prizes for pioneering, unless you need that facility.... I think there's always a business reason to go down that route, and then we will implement what we know will work... It might not be the latest technology. We pride ourselves on the functions that we deliver - how we deliver it might not be the latest technology. but we deliver it.'
The retention of technical staff has also been helped by the geographical location of the Halifax Building Society in Halifax itself. Opportunities for alternative technical employment in an area in which the Halifax so dominates the local market are comparatively poor. This limited scope for alternative employment in the area helps to check opportunistic actions by technical staff because there are few alternative sources of employment. Although this was largely fortuitous in the Halifax case, geographical location is not a factor that should be lightly overlooked: both the Inland Revenue and the DSS deliberately located their IT development staff in the fairly remote locations of, respectively, Telford and Lytham Saint Anne for precisely the same reasons. The Inland Revenue's technical director has commented that: `in the early days it was easy to retain staff because the only thing in Telford town centre.... was the shopping precinct'.5 Location as a device to retain scarce skills was also considered in the DSS:
`There was the lack of civil servants with development skills. Those that did have the skills moved on too quickly ... high turnover because it's the M4 corridor ... a decision was taken to move the project lock, stock and barrel to Lytham, because here we have obviously a less fluid job market. There are only a couple of other ADP outlets here; although they tend to be specialist computing rather than our kind of commercial computing.' (Technical Manager).
In-house development also reduces the likelihood of post hoc appropriability problems arising from consultants. As one BIS Project Manager commented, imagining himself in the shoes of an outsourcing competitor:
`If you get a problem then you go and talk to a programmer. That programmer fixes it, and you're back up and running within a matter of hours. If I have a problem then I have to write to the supplier of the software, who will then tell me "it's not my fault...." I then have to write out a request for a change; they have to come back and quote me on that change. Having quoted me, they will then tell me how long it's going to take, and they'll deliver it in their time. When I get it back, I've got to test it, accept it, (and) if I get it wrong, I've got to go through the whole rigmarole again.'
Within the computer division, technology is treated as the means to an end, rather than the end per se. In other words, BIS is driven by the needs of the organization rather than the enthusiasm of its technicians. This is quite different from the attitude of the computer division in the past, as one Manager explained:
`If you go back to the time we were doing ATMs, the requests from outside BIS were very few and far between. Fifteen years ago when I started, I suppose we decided what we were doing. Nobody else did.... The Computer Department was left to get on with it, and did basically what we wanted.'
So much so, the computer division now appears to act as an internal policing mechanism, guarding against the pursuit of sub-goals by other departments, as is apparent in the comments of the Business Systems Project Manager:
`Nowadays the demands are not internal within BIS, but are actually hitting us from outside. We will still generate ideas from inside .... Anybody out there, whether it be a programmer, business analyst or whatever. If they feel there's something can be done to improve a system, there's nothing to stop them coming forward and saying "can we do this to improve it?." But again, I would cost out those particular jobs and say "do you realise it will cost us this much to do it, and really, is it saving us that much out there in the branches or in the other areas of the Head Office?".'
Simply put, the Halifax approach is to critically evaluate the importance of the SAP and how this affects their main business and then to concentrate on DAPs, in particular the EAP. Somewhat speculatively, one could argue that the two are connected: trying to be up-to-date with all the latest technological fads clearly increases the danger of DAPs arising. For example, in small high technology firms, where modernity is a matter of survival, DAPs are notorious. The Halifax experience suggests that familiarity with the organization, its requirements, and the needs of the business may be far more important than instances of technical wizardry. Of course, the danger of the Halifax approach to system programming is that it can become parochial, too far removed from the technical actions of its competitors. To counteract this potential problem, BIS are active in monitoring the technical environment through conferences, industrial meetings and workshops.6
Safeguarding Knowledge from External Threats: The Public Sector
Within the public sector, our cases indicate a differential response by two central government departments in designing appropriation strategies to safeguard organizational knowledge from external threat. Our more successful case. the Inland Revenue, placed considerable emphasis on the generation of in-house technical skills. In our relatively unsuccessful case, the DSS, this build-up of experienced staff came late into the project, and there was too much dependence upon external knowledge providers.
A key ability for any organization is the co-ordination and application of potentially diverse sources of knowledge within a business context; combining technical knowledge of the computer systems held by consultants, for example, with knowledge about the customer base held within the organization (Dyerson and Roper 1991). As was argued earlier. for effective introduction of new information technology, it is necessary for a firm's professional staff not only to receive training, but also to have an incentive to share information with colleagues. Our cases suggest that the success with which this can be achieved depends in part upon previous experience and staff continuity. It is the skills base of the firm's employees, their stability, and the effectiveness with which those skills can be integrated which provide the base conditions for the nurturing of organizational as opposed to purely individual knowledge. Institutional change and technological challenges requiring the development of new routines and new competencies, led to a managerial response which, in effect, changed the conditions for knowledge retention and integration or, to put it differently, the respective role played by SAPs, lAPs and EAPs.
Arguably one can analyze HRM activities with regard to their implications for appropriability problems. For example, apparently similar employment policies might have very different effects for questions of knowledge management, depending upon the context in which they are applied. Desiring more information sharing, the technical division at the Inland Revenue developed an ethos based on teamworking. Encouraging individuals to act co-operatively and thereby forgo short-term gains turns upon rewarding or encouraging individuals to identify with the organization's long- term performance. In doing so, the direct link between effort and reward was broken, extra motivation was mobilized and IAP incentives reduced. Loyalty is encouraged in individuals by drawing them into the organization and giving them a stake in the company's long-term future through, for example, a carefully defined bonus scheme or promotion policy. Co-operative behaviour arising from mutual self interest is a characteristic of this type of teamworking.
Tushman et al. (1986) showed that organizations build upon an existing orientation incrementally, a fact which can lead to extreme `organizational conservatism' (Child et al. 1987). This has led researchers to argue that effective change will often come from outside the organization (Ginsberg and Abrahamson 1991). However, the flip side to this is the danger that skills brought in from outside will sit uneasily with the existing orientation, and/or will not contribute to cumulative learning. In this scenario, the organization has attempted to address the SAP, but because of an acute IAP, little actual benefit has been gained. The reason for the lack of tangible results is that bringing in expertise from outside is not the same as deploying internally nurtured skills. Crucially, what is lacking in the employment of external skills is the customization of those skills to the organization's tacit routines. This can lead to an undue emphasis upon the technological aspects at the expense of architectural aspects (Kay and Willman 1991), the latter referring to the effectiveness of organizational arrangements in solving DAPs. In one of our cases (DSS), this was indicated by a largely technically driven implementation of their computer system, coupled with a high dependence upon external 'experts', indicating loose management.
Malerba (1992) indicated that each learning process (e.g. learning by doing, learning by using, learning from supplier) is linked to a specific type of (already existing) productive and technological knowledge within the firm. This points to a firm-specific knowledge base which is the prerequisite for the purposeful and tailored assimilation of external knowledge held by consultants or suppliers. In both the Inland Revenue and the DSS, technical skills in large-scale computerization were missing at first, and so had to be brought in through external consultants. However, whereas the Inland Revenue deployed a careful strategy to nurture and appropriate this `seed bed', the DSS floundered with a haphazard approach that placed the organization at a knowledge disadvantage and provided the conditions for appropriation problems to emerge.
At the Inland Revenue, development work and responsibility for the dayto-day running of existing systems was organized within a single division. The division itself was largely made up of programmers drawn from the local population and business analysts drawn from branch offices. Incorporating the practical experience of the business analysts helped to curb the technological enthusiasms of the technologists within the division; as the Director commented:
`The users, the good users, have got the bug and have translated into IT people. That's important because we need to have real business analysis and appreciation integrated into an IT team, rather than simply IT analysis.'
The average turnover of technical staff was less than 5 percent, lower than the average for the public sector. Just as was found earlier, in the Halifax case, almost all senior positions within this division were filled by managers who had originally started as programmers or system analysts. The position of project manager, for example, was filled by someone who had joined the department during the 1960s. This unprecedented level of stability at senior levels was useful in meshing conventional routines with new routines more appropriate for the faster pace of technological change. In practical terms, this meant the replacement of the accepted committee style of decision making by project management techniques. Direct control over project staffing was introduced, facilitating integration between in-house and external expertise. Responsibility was streamlined through the Project Director with ministers, trade unions, suppliers and all the development teams. Normal financial restraints were eased, allowing the project to have control over its budget. For the first time in the Civil Service, a bonus scheme was introduced that rewarded staff in post for their performance.
Direct control allowed Inland Revenue managers to institute several devices designed to constrain the formation of information ghettos under the external control of consultants brought in to help with the computerization process. In particular, consultants were fully integrated into a line-management system, directly answerable to Inland Revenue managers. A policy of 'tasking' was also imposed in which Inland Revenue staff were actively encouraged to shadow the activities of consultants. This was a deliberate attempt to ensure that the knowledge and experience of the consultants would be appropriated by the organization. As a senior project manager commented:
`We deliberately set out to have mixed teams of consultants and in-house staff working together. We do not put the consultants all in one place and ask them to deliver a part of the project and put our own staff in another place and ask them to deliver another part of the project. We put the two together with firm deadlines, integrated teams, and in that situation all our in-house staff will be learning from the special expertise of the consultants working with them.' (Public Accounts Committee 1987, para. 1783)
Such mixed teams composed of both internal and external experts were seen as the best solution to simultaneously addressing questions of IAP and EAP. They also go beyond the recommendations found in the literature described in the second section of this paper, which has largely focused on the creation of inter-departmental teams.
Longevity and commitment played an important role in the gestation of internal expertise, but success in this endeavour was not typical of the public sector in general, where promotion has traditionally required frequent changes in position. In the DSS, for example, the stewardship of the department's operational strategy changed three times in as little as six years. Not only did this hamper the effective build up of organizational knowledge, such fluidity also enabled decision takers to avoid responsibilities for their actions. This provided scope for the formation of information ghettos within the organization. In the DSS, established routines were not appropriate for the task: a hierarchical, top-down, centrally planned management structure was put in place to organize the design, development and implementation of the computers into the department. Traditional personnel management methods were also adopted. Yet this was inimical to establishing continuity, commitment and effective communication within the organization. As one interviewee (a project manager) tellingly put it:
`Some of our consultants have been around for seven years. There is hardly anybody, any civil servant who will have been around for more than three or four. The continuity is actually provided in the main by consultants.'
Responsibility was split between three layers of management, with little integration between them. This was also apparent at the human resource level where technicians and users enjoyed few incentives for the development of co-ordinated work practises. At a practical level, the inappropriate human resource approach yielded computer systems that fell far short of users' needs having largely been driven by technicians. Then again, the literature suggests that a lack of integration between organizational and technological aspects seems to be the rule rather than exception in the innovation process (Bilderbeek and Buitelaar 1992; Curry 1994). As one implementation manager commented:
`The user community ... is not ... skilled at doing the user testing or representing the user, because they don't have sufficient knowledge. We've lacked skilled users at times, and forceful users, who would insist the user spec. was met, who would insist that the design was as per user spec., and would then insist that completely independent testing was done.'
What is being expressed here, in terms of our conceptual framework, are IAP problems where different organizational groups pursue differing agendas. Managerial policies designed to encourage a stable skills base are important in helping the firm to achieve a close fit between technology and organization through the generation of context-specific highly idiosyncratic knowledge. But our cases also suggest that organizational routines have to be carefully tuned to effectively capture this expertise. The DSS, in effect, employed both internal and external developers to work on the project without a clear strategy for the appropriation of their knowledge. The DSS, unlike the Inland Revenue, appear never to have thought critically about the SAP confronting the organization. Consequently, knowledge creation lacked focus and was at best only weakly developed by the DSS. Faced with this void, external consultants developed a consistent strategy that exploited organizational weaknesses to safeguard the retention of their expertise, as one project member commented:
`You found that after about three or four months you were meeting the same people in the corridor again, because of problems identifying the system: "Who designed this?". "Oh, that guy was from X, or the Y Corporation". "So what did they do?" . "Oh I don't know". "Well, we'd better bring him back then and ask him what he did"... And before you knew it they were all back. And they must have known that when they did it. We were the stupid people who didn't actually appreciate the long-term implication of what we were doing.'
In summary, mechanisms designed to capture relevant knowledge, such as the choice of technology, were not implemented successfully in the DSS case. At a very early stage in the DSS project, attempts to build an internal group of developers who would stay with the project foundered following trade disputes. The lack of effective communications, and the lack of stability amongst senior personnel, precluded the development of a consistent strategy for development and implementation. Both IAPs and EAPs became serious obstacles without always being clearly recognized as such. By way of contrast, the Inland Revenue achieved both better staff continuity and knowledge integration. This enabled the Inland Revenue to deal more successfully vis-a-vis external knowledge providers.
Conclusion and Summary
Separate literatures have been devoted to studying different aspects of appropriation of organizational knowledge. What is missing however, is a theoretically guided consideration of how managers decisions can affect the stock of organizational knowledge. Nonetheless, useful theoretical insights can be drawn from a number of diverse disciplines and approaches. Indeed, the case studies seem to have vindicated our theoretical framework that these dimensions are not in isolation from each other, but have to be considered simultaneously. Keeping technology up-to-date is not a function that should be optimized under any circumstances; rather, decisions regarding SAPs need to be seen together with a firm's decisions on DAPs. These in turn necessitate a simultaneous consideration of IAPs and EAPs: a decision taken with regard to the latter can vindicate a decision taken with regard to the former. In the existing literature, IAPs should typically not be analyzed in conjunction with EAPs but, rather, allocated to separate domains in the literature.
Specifically, failure in achieving staff continuity can give rise to both IAPs and EAPs, and the DSS case on the one hand, and the Inland Revenue case on the other, provided different degrees of failure and success, respectively. At Halifax opportunism was restrained by establishing in-house skills around a technology that is somewhat archaic, which -- although unlikely to be totally anticipated -- helped to curb the technological enthusiasms of IT specialists and so helped to avoid the pursuit of sub-goals within the organization. In simplified terms, the Halifax approach is to downplay the importance of the SAP and concentrate instead on DAPs, in particular the EAP. Thus, in-depth analysis of our cases suggests that managerial choices are important in the appropriation of knowledge, both with regard to the idiosyncrasy of knowledge, and the control of individual behaviour. How and in what way technological knowledge is built up as an organizational asset is essentially the cumulative result of strategic choices over time. Pursuing a strategy of knowledge appropriation is a key requirement. From our case studies, establishing a successful environment meant installing careful control over personnel, avoiding dependence on external knowledge suppliers as well as high discontinuity within the organization's knowledge base. Our conclusion suggests that much of the existing literature that focuses on the consequences of technological change for the individual, such as deskilling, and how they can be countered or mitigated through training and/or enhanced employment security, is too narrowly drawn. Strategic choices about human resource management, location decisions, promotion and reward systems can all work together to incrementally build an organization-specific knowledge base. However, much more research has to be conducted to specify the exact nature of the interplay between managerial decisions, technology and the creation of a superior and distinctive base of organizational knowledge.[Footnote]