The Client-Consultant Interaction in Professional Business Service Firms: Outline of the Interpretive Model and Implications for Consulting

Transcription

1 Client Consultant Interaction The Client-Consultant Interaction in Professional Business Service Firms: Outline of the Interpretive Model and Implications for Consulting Timothy Devinney* Natalia Nikolova** *Professor Centre for Corporate Change, AGSM ** PhD Student, University of Cologne and Researcher Centre for Corporate Change, AGSM The University of New South Wales, Gate 11, Botany Street, Randwick NSW 2031 Australia May 2004, Draft version Abstract In this paper, we analyse and seek to understand the complex relations that occur between clients and consultants during consulting projects. The paper uses ideas from Schön s model of the reflective practitioner to propose a new model conceptualising the client-consultant team as a community of interpretive communities. We show that client and consultant belong to different interpretive communities, and have, thus, difficulties to understand each other and share knowledge. An effective knowledge transfer and creation between client and consultant presupposes that the involved actors reflect on their own interpretations and perspectives on topics and share them with others. This is a process of translation between different interpretive communities that often enables new worldviews and simultaneously creates power positions within the client-consultant team. Thus, we extend Schön s model and embed it into a broader theoretical framework, proposing a more sophisticated picture of the client-consultant interaction. Keywords: management consulting, client-consultant interaction, interpretive communities, knowledge transfer and creation. The authors would like to thank Stewart Clegg, Markus Reihlen, John Stuckey and Ivo Zander for their valuable comments on this paper. The financial support by the German Ministry of Education and Research (research grant 01HW0168) is also acknowledged.

2 I. Introduction Professional business service firms are one of the most dynamic and advanced industries in the modern economy. Throughout the last several decades most professional business services, such as those provided by management and engineering consultants, lawyers, accountants, advertising agencies and market research firms, have gained increased economic significance in all advanced economies. But the post-dot.com economic slowdown hit them hard: they experienced a significant decrease in their growth and an increase in the number of potential clients expressing doubts about the value of their services (Prakash and Samwick, 2003). As Ashford (1998: xvi) suggests, increasingly clients are concerned that consultants are pushing standardised solutions rather than really listening to the issues and being guided accordingly. Some authors predict that, due to the gap between actual consulting solutions and increasing and unserviceable client s demands for innovative solutions, client s dissatisfaction will continue to rise (e.g., Czarniawska, 1999), making the clientconsultant relationship more difficult to sustain. As the client-consultant relationship is essential to professional service firms (Schön, 1983), it is important to discern the real nature of the client-consultant interaction, what makes the interaction successful, and how should it be designed. Prior research has focused on the client-consultant relationship as a client-expert interaction (Schein, 1987, 1988; Abbott, 1988; Kubr, 1996) and as a symbolic interaction (Clark, 1995; Clark and Salaman 1998a,b; Alvesson, 1993, 2001). We argue that both views stress single features of the client-consultant interaction without recognising its multidimensional and complex character. Reviewing contemporary research on consultancy projects, Engwall and Kipping (2002) conclude that the interaction process between consultants and their clients is still poorly understood, perhaps because, as Hislop (2002: 657) claims, the nature of the client-consultant relationship and the role of the client firm shaping this relationship has tended to remain neglected and unexplored. Fosstenlokken et al. (2003: 868) call for further investigation into the nature of client relationships and the role clients play in the knowledge creation process in professional service firms a suggestion that is all the more relevant as learning through project work becomes considered as the most important source of knowledge development in consulting firms. Recently, researchers have directed their attention to the interpretive and multi-layered character of consulting work. Fincham and Clark (2002) argue that interpretation about whether knowledge is needed (whether the client has a problem), what kind of problem the client has, how to apply and create knowledge with this particular client and how to make 2

3 sure that change occurred is the basis of consulting expertise. Kipping and Armbrüster (2002) stress that the otherness of consultants the fact that consultants are outsiders to the client organization and therefore their knowledge, their work methods, and language differs from the client s can be a considerable burden for a successful interaction with the client and may prevent consultants from being involved more intimately in the client s business. Early on Schön (1983) recognised that the otherness of professionals and the nature of their expertise required an analytic view on the client-consultant interaction as a reflective conversation with the client. Drawing upon Schön s model and using ideas from organisational learning approaches, actor-network theory, and Foucault, this paper addresses the client-consultant interaction in a new theoretical framework. It is argued that client s and consultant s knowledge are embedded in communities of interpretation, and knowledge creation and transfer takes place within such communities (Brown and Duguid, 1991). Because there has been little discussion in the literature about the interaction between separate communities and the difficulties of sharing knowledge across boundaries and reaching synthesis (Bechky, 2003), the concept of interpretive communities has to be expanded to allow for a better understanding of the interaction process between consultant and client. By applying the concept of interpretive communities, we show that clients and consultants may have problems in understanding each other s language (Carter et al., 2004). Consequently, within their interaction there is a problem of translation: knowledge transfer and creation will only occur if the individuals involved in the interaction process learn to share their understandings and to develop new meanings to familiar topics. Whose perspective will build the basis of the problem solution when there are several interpretations is dependent on the power or interpretive dominance of each community, which is constructed through discourse (Callon, 1986; Meindl et al., 1994). Thus, we argue that the translation of different interpretations is simultaneously always also a mechanism of power (Foucault, 1972; Clegg, 1989, 2001). We will focus on critical success factors and provide a more realistic portrait of the client-consultant interaction than current accounts. In addition, we will outline some practical implications for managing the interaction process and the customer-service provider interface, providing assistance to both consultants and managers to their attempt to improve cooperation. In section II, after a short remark to the two traditional models of clientconsultant interaction in the literature, Schön s model is introduced as a more promising foundation for a new understanding of the client-consultant relationship and main implica- 3

4 tions are shown. In section III, a new theoretical framework is introduced and its main characteristics and general implications will be discussed. Section IV draws the themes of the paper together and reiterates the main points. II. The Client-Consultant Interaction: Basic models Two basic models dominate the existing literature on the client-consultant interaction. Figure 1 summarises the main propositions of these models and presents Schön s model, which is the basis for the new framework that will be discussed later in Section III (see for a more comprehensive discussion of the existing models on client-consultant interaction Devinney and Nikolova, 2004). Insert Figure 1 Here Earlier research on the consulting industry regards the client-consultant relationship as a client-expert interaction. The consultant as an expert identifies a client s problem and transfers knowledge in the form of a problem solution while remaining an objective and neutral advocate of best practice (Alvesson and Johansson, 2002: 235). This literature is primarily concerned with the role of consultants as a vehicle for increasing the effectiveness of the client organization in a planned collaborative intervention process (Fincham and Clark, 2002; Kubr, 1996). However, there has been growing critic on this model from both researcher and practice. It is been agued, that the model presents consulting as a unidirectional gathering and processing of information rather than as a real interaction (McGivern and Fineman, 1983; Daft and Huber, 1987). The client-consultant communication is seen as an error-free process of message sending and message receiving through transmission channels (Boland and Tenkasi, 1995; Daft and Huber, 1987). Empirical research shows that especially in unfamiliar situations this model does not enable learning between clients and consultants (McGivern and Fineman, 1983; Kitay and Wright, 2003) because of the absence of understanding and reflection between the involved individuals (Schön, 1983). Thus, the generated solutions are often standardised and do not account sufficiently for the needs of the particular client (Ashford, 1988). Client s dissatisfaction with this kind of consulting is the reason why a growing number of researchers have doubts that the traditional research stream deals with the right problem. In the more critical literature on management consulting, it is argued that the real contribution of consultants is not in creating and transferring knowledge to the client but rather in making managers, who often are overstretched to respond to increasing 4

5 external pressures and demands, feel more secure in their role as a modern manager (Clark and Salaman, 1998a). It is argued that the nature of consulting knowledge is inherently ambiguous and slippery (e.g. Alvesson, 1993, 2001; Clark, 1995) and is subject to fashions (Jackson, 2002; Kieser, 1997) and that clients are sometimes left with the impression that they have paid for stories and images rather than real solutions (e.g. Clark and Salaman, 1996a, 1998a). Although the model pays attention to one important task of a successful consulting, the creation of image and positive client expectations, which is ignored in the expert model, it concentrates to much on the symbolic character of consulting work and results and ignores the fact that the cooperation of both parties generates tangible, valuable and rational outputs. It does not take much effort to question why, if there were no direct benefits for companies and management from engaging management consultants other than what amounted to mere rhetoric, how managers could justify the huge sums invest in contracting with consultants. Similarly, how is the large growth of consulting companies to be explained in a world with strong governance and demands for cost control? A recent empirical study on the long-term effects of hiring management consultants contradicts the critical model by showing that, over the long term, the investment in consulting projects yields positive financial and organizational returns to companies (Prakash and Samwick, 2003). To sum up, although the traditional research stream regards knowledge transfer as crucial for the success of the client-consultant interaction, the more heretical approach suppresses the question of knowledge creation and transfer and argues that the active management of client s expectations is more important. Both views stress single features of the clientconsultant relationship without recognizing its multidimensional and complex character. Thus, it can be concluded that both models do not ultimately explain the real nature of the client-consultant interaction. They both suggest a consulting centric view on the clientconsultant relationship rather that an interactive one. As every managerial decision requires some interaction between the involved individuals, these models do not tell us what makes the client-consultant cooperation successful. Table 1 summarises the main conclusions of the models regarding the nature of consulting knowledge and knowledge asymmetry between client and consultant, as well as the knowledge transfer process and power aspects in the client-consultant team. It further highlights major differences to the models we present in what follows. Insert Table 1 Here 5

6 The Reflective Practitioner Model The problems with the previous two models and their failing to give a satisfactory picture of the client-consultant cooperation implies that a different concept of the client-consultant interaction is required if we are to understand this industry adequately. Schön (1983, 1987), provide such a portrait, conceptualising the client-consultant interaction as a reflective conversation between clients and consultants. Schön s model of consulting practice is centred on the nature of consulting knowledge. He claims that there are two types of consulting expertise knowing-in-action and reflection-in-action and that both of them are embedded in the socially and institutionally structured context shared by a community of practitioners (1987: 33). This context includes the body of professional knowledge, the appreciative system, and frames of a particular community of practice, which serve as the source of rules and procedures routinely applicable when solving familiar problems. However, in unfamiliar situations, where there is no obvious fit between the characteristics of the situation and the available body of theories and techniques, new rules and new knowledge are created through reflection-in-action (Schön 1983, 1987).1 During this process, based on a repertoire of story types, interpretive explanations, examples, images and actions, consultants focus on certain details of the problem while leaving others in the background, thus framing the problem in a particular way (Schön 1983: 125). Through the framing of the problem they produce unintended changes, which give the situation new meanings. In this way, consultants yield new discoveries, which require new reflection-in-action. The process spirals through stages of appreciation, action, and reappreciation. (Schön 1983: 132). Consultants build gradually from their perception of the client s problem toward an interpretive synthesis congruent with their fundamental values and theories. Because both knowing-in-action and reflection-in-action are tied to the body of professional knowledge, as well as to the appreciative system and role frames of a particular community of practitioners, it follows, that consulting and professional knowledge in general represents just one way of looking at topics, one possible interpretation of problems. Furthermore, Schön (1987: 36) emphasizes that consulting knowledge is constructed as a result of individual actions embedded in a social context: through countless acts of attention and inattention, naming, sensemaking, boundary setting, and control community members make and maintain the world matched to their professional knowledge, appreciative systems 1 Reflection is the ability to reflect on the own intuitive knowing (knowing-in-action) in the midst of action and is concerned with double-loop and triple-loop forms of learning (Argyris and Schön 1996; Raelin 2001). 6

7 and role frames. They have, in short, a particular, professional way of seeing their world and a way of constructing and maintaining the world as they see it (Schön 1987: 36). Because both client and professional bring to their encounter a body of understandings which they can only very partially communicate to one another and much of which they cannot describe to themselves (Schön 1983: 296), the communication between parties to a consulting project is very difficult and requires intensive interaction. Schön sees the solution as open and minimally defensive interaction where both client and consultant commit to having their positions and interpretations confronted and tested and both are open to the reciprocal exploration of risky ideas. Thus, the reflective practitioner model emphasizes that clients need to function as reflective practitioners as well; they should develop a new kind of skill in asking questions and cultivate competence in reflective conversation with the professional (Schön, 1983: ). Without the exchange of perspectives, consultant s expertise will remain a black box for the client. Consequently, the relationship between client and consultant is more balanced and is based on the demystification of expertise and on mutual exploration. Schön views both consultants and clients as powerful and interdependent. Their power arises from the fact that both parties make valuable contributions to the problem solving process. Consultants are now expected to reflect, in the presence of clients, on their expertise, thus making themselves confrontable by their clients rather than keeping their expertise private and mysterious, which is the case in the expert model (Schön, 1983). By abdicating their unquestioned authority and the comfort of relative invulnerability they gain access to a new type of interaction, one of discovery and self-reflection. On the other side of the relationship, clients agree to join the consultant in the problem solving process and to work to make their knowledge and experience clear to themselves and to the consultants. The relationship becomes equilibrated as both parties give and receive help (Schön, 1983). The reflective practitioner model is a first attempt at creating a more realistic and thorough picture of consulting and professional practice in general. Compared to the expert and the critical models, it is a broader elaboration of knowledge creation in a professional practice taking into account that different individuals frame problematic situations in different ways (Schön, 1987: 4). People will see problems in the same way if they share the same appreciative system and body of explicit knowledge; e.g. if they belong to the same community. Thus, Schön realises that there are different communities with different views on topics and different interpretations on problematic situations. Hence, there are multiple ways of framing the practice role, each of which entrains a distinctive approach to problem setting 7

8 and solving (Schön, 1983: 41). This has several implications for the reflective concept of a professional practice: First, problems are not given as in the case of the expert model, but rather constructed in a process of reflection-in-action. Second, professional knowledge cannot be reduced to an objective commodity; it is rather a way of looking at topics, which was once constructed and can be reconstructed. Thus, this view of knowledge is contradictory both to the expert model, where knowledge is regarded as an objective reality, and to the critical model, where the existence of a professional body of knowledge is denied. Third, because clients belong to different communities, they will have a different view on the problem setting and solving than consultants. Hence, there is knowledge pluralism in the consulting company as well as in the client-consultant relationship, which requires both clients and consultants to reflect on their interpretations and frames and share them with the other if there are to work effectively together. Furthermore, only through reflection-in-action consultants can handle unfamiliar situations and create new knowledge. The reflective practitioner model is a first step to a more sophisticated view of the client-consultant interaction but one that is limited. For example, perhaps the most comprehensive limitation is that even Schön himself sees serious constraints in its application across the broad spectrum of client-consultant interactions. Because of the different demands on competence, and the different sources of satisfaction that are presented both to the professional and to the client (Schön, 1983: 298), he argues that it should be applied only in cases when the client s problem is of sufficient importance. In emergent or routine situations a consulting type as proposed in the expert model would be more appropriate. In addition, the model is underspecified. Schön mentions the existence of different communities of practitioners but does not discuss their nature and the process of their emergence, which is why he still regards the expert model as the appropriate consulting model in situations where clients problems are familiar. What is problematic is that both Schön s, and the expert model, are based on different assumptions about the nature of knowledge and knowledge transfer and creation. Further, the reflective practitioner model recognises that because consultants participate in different communities, as do clients, there are differences in their framing and naming of problems, which, because both perspectives are legitimate, may not be objectively resolvable. Schön does not offer a solution for how these differences can be overcome: the resolution of such differences depends on the little-understood ability of inquirers to enter into one another s appreciative systems and to make reciprocal translations from one to the other (Schön, 1983: 273). Thus, the model fails to explain the process of translation and to show exactly how the communication between client and con- 8

9 sultant can be improved. Furthermore, although Schön provides a very promising view of knowledge creation, he still describes it as a consulting centric process where clients play no role. He views consultants and clients as interdependent, and their roles as interchangeable, so that mutual learning takes place, but does so without an accounting of the role that power aspects play in the client-consultant relationship. To summarize, the reflective practitioner model presents a different view of the clientconsultant interaction, stressing the need for a reflective interaction between clients and consultants. It provides a promising starting point for a more sophisticated discussion of the client-consultant cooperation. However, his model remains too broad and unspecific and his suggestions do not pay sufficient attention to the interactive character of knowledge transfer and creation in consulting. Schön s model needs a further development, which we will provide in the following model. III. The Interpretive Model The interpretive model puts the client-consultant interaction in a new theoretical framework. Drawing on Schön s ideas it is argued that the origins of the different understandings of client and consultant and the difficulties with knowledge transfer can be found in the existence of different interpretive communities. Here we view the consultant as a reflective practitioner with a body of knowledge based on abstract concepts, theories (a technological body of knowledge) and experiences, and conceptualise knowledge transfer as a learning or reflective process within, and between, interpretive communities. The idea of interpretive communities is related to the concepts of communities of practice (Brown and Duguid, 1991; Wenger, 1998) and communities of knowing (Boland and Tenkasi, 1995). The interpretive model emphasises that stories and rhetoric, both as instruments for knowledge transfer and creation and as means of altering one s power position in the relationship, have an important role to play in the success of any client-consultant interaction. This model pays special attention to the power questions in the client-consultant relationship and shows that knowledge and power are eng related and enforce each other. Nature of knowledge and interpretive communities Similarly to Schön s notion that a professional knowing-in-action is embedded in the socially and institutionally structured context shared by a community of practitioners (1987: 33), our model is based on the assumption that knowledge resides in people s minds and is the result of individual s cognitive processes, which are embedded in a social context. As Tsoukas and Vladimirou (2003: 979) define it knowledge is the individual ability to draw 9

10 distinctions within a collective domain of action, based on an appreciation of context or theory, or both, where the context can be interpreted as a collectively generated and sustained domain of action (2003: 977) or a language-mediated domain of sustained interactions (2003: 978), and to engage in collective work is to engage in a discursive practice (2003: 978). It follows that knowledge is created and transferred within socio-culturally constituted interpretive communities or discursive practices through an ongoing process of interaction between individuals (Empson, 2001; Bechky, 2003; Reihlen, 2003). It is neither aggregated from individual cognitive processes nor does it originates entirely within social processes as argued in the critical model. In the following we discuss the nature of interpretive communities and the process of their emergence. Interpretive communities The starting point for the concept of interpretive communities is the view of organisations as dispersed or distributed knowledge systems. A firm s knowledge is distributed between different individuals; no single mind owns the totality of relevant knowledge (Boland and Tenkasi, 1995; Tsoukas, 1996). Moreover, all articulated knowledge is based on an unarticulated background, a set of subsidiary particulars which are tacitly integrated by individuals (Tsoukas, 1996: 17; see also Boland and Tenkasi, 1995). People construct their understanding2 out of a range of aspects like ambient social and physical circumstances, as well as histories and social relations to other people (Schön, 1983, 1987). In a process of interaction individuals with similar interpretive positions build a specific worldview or perspective. This is a learning process through which conceptual frameworks are aligned and shared meaning and understanding is formed. It is a spontaneous process emerging from the usually informal networking between individuals (Swan et al., 2002). Thus, an interpretive community reproduces itself through shaping individuals, yet individuals transform it as well (Hanks, 2002; Reihlen and Ringberg, 2003). As individuals become aware of how their thoughts are guided by such frameworks they can start to choose between them (Reihlen and Ringberg, 2003). In this way different interpretive communities evolve. Organizations are a plurality of intersecting and competing communities or communities-of-communities (Brown and Duguid, 1991; Fox, 2000). Interpretive communities are not stable; they are not a fixed system of positions (Clegg, 2001: 135). At any point in time, individuals access a number of different interpretive positions through which they make sense of an incidence or topic (Reihlen and Ringberg, 2003). 2 The terms perspective, meaning and understanding are used interchangeably throughout this paper. 10

11 Thus, when persons or groups switch from one interpretive position to another, their perspectives, preferences and dispositions will change (DiMaggio, 1997). Also, when newcomers access the community or the environment conditions within which it is embedded change, the shared meanings of the community can, and generally will, be transformed (Brown and Duguid, 1991). It follows that large-scale, more-or-less simultaneous frame switches by many independent actors may cause large-scale changes in interpretive communities (DiMaggio, 1997) and whole organisations. Thus, organizations should be seen in terms of the infinite play of differences that discourse constitutes (Clegg, 2001: 135). Accordingly, client s and consultant s organisation as well as the client-consultant team are communities of interpretive communities, which means that there will be always different perspectives and frames in regard to any particular situation. As Schön (1983) points out, only if client s and consultant s learn to reflect on their own understanding and to share it with other, will they be able to work effectively on problem solution. In the following, the process of sharing each other s interpretations is discussed. Knowledge transfer Because people participate in multiple interpretive communities they will have problems in understanding one another fully. For example, when looking at the same phenomenon members of different communities will not only see different solutions to the same problem but may see quite different problems from what appear to be similar conditions. In this sense, problems do not have an objective existence; they are defined and constructed in the interaction process (McGivern and Fineman, 1983; Behrens and Delfmann, 2002). Arguments that persuade their own community convincingly may have little or no weight in other communities (Boland and Tenkasi, 1995). Thus, communities may have problems in understanding each other s language, much of which might be perceived as pure noise (Clegg et al., 2004: 38). Schön (1983: 271) argues that members of interpretive communities develop a feel for the media and languages of their practices, which is why they cannot convey the art of [their] practice to a novice merely by describing [ ] or even demonstrating [their] ways of thinking. Similarly, Tushman (1977: 591) points out, the inherent conceptual and linguistic differences [between specialized communities] act as communication impedance or as a communication boundary hindering the free flow of information. Schein (1996: 18) concludes that organizations will not learn effectively until they recognize and confront the implications of the existence of different interpretive communities. Weick (1979) notes that managers will be more effective when they are able to generate several interpretations and understandings of organisational events so that the variety in their understanding is equiva- 11

12 lent to the variety in the situation. Thus, for knowledge transfer to take place between different interpretive communities, it is not sufficient to enable an open and minimally defensive interaction between members of different communities, as Schön (1983) suggests. Rather, it is necessary to span the boundary between interpretive communities and to establish and sustain connections amongst these communities to enable knowledge transfer (Wenger, 1998). Knowledge transfer is possible only between interpretive communities where some conceptual frameworks are shared. Sociologists talk in this case about aligning actions (e.g., Stokes and Hewitt, 1976), which occurs in situations where some minimal degree of shared understanding about the nature of the exchange and the rules governing it among different communities is necessary (Donnellon et al., 1986). Members of different communities will take organised action despite holding different interpretations of common experience if they share a repertoire of behavioural options that members of a given society recognize, respond to, and use to interact with one other (Donnellon et al., 1986: 44). The shared repertoire of communication behaviours leads to a development of so-called equifinal meanings dissimilar interpretations with similar behavioural implications. To create alignment amongst different communities is not a case of deciding which one has the right viewpoint, but of creating enough mutual understanding among them to evolve solutions that will be understood and implemented (Schein, 1996: 17). Fiol (1994) argues that meaning is not a one-dimensional construct. She discusses two meaning dimensions: content and frame. While content defines what is expressed, framing is reflected in how something is expressed. Thus meaning resides not only in the content of communications, but in the framing or language of communications as well. Framing refers to the way people construct their argument or viewpoint, regardless of its content (Fiol, 1994: 405). The way communications are framed (e.g., broad, flexibly-framed or narrow, rigidly-framed arguments) carries important meanings in collective decision processes. Fiol argues that learning can occur also in the absence of shared meaning (regarding the content) when people converge around a frame that is broad enough to encompass the differences in their interpretations. Figure 2 illustrates schematically the discussion about knowledge transfer between matching interpretive communities. We are presenting some concrete mechanisms for understanding and sharing of meanings later. Insert Figure 2 Here 12

13 According to these arguments, the following hypothesis [H1] is proposed: H1: The emergence of shared meanings, shared conceptual frameworks and a shared repertoire of communications behaviors will be positively associated with the transfer of knowledge within the client-consultant team. At the same time it is clear that if there is a high variety of perspectives in the clientconsultant team, and the members of the team are not able or not willing to find ways to share and make clear to each other their different interpretations, knowledge transfer will be inhibited. Malefyt (2003) emphasises another important aspects of a shared language: the way insights are communicated and presented to the client is vital for creating connections. For the case of advertising agencies he argues by presenting proprietary models that engage social interaction and by expressing ideas in a language that evokes shared emotion, the agency hopes to draw the client deeper into long-term relations of affinity (Malefyt, 2003: 144). Therefore, shared language is important not only for knowledge transfer between clients and consultants but it is also the precondition for the creation of long-term relationships with the client. He further argues that activities such as model-building and figurative stories ensure the right impression of competency while creating a closer affinity between service provider and client. Knowledge creation Existing knowledge structures the way people use information (DiMaggio, 1997). Thus, to create new knowledge, people need to critically reflect on their knowledge and reframe it, or as to use Schön s words, to engage in a reflective conversation (1983: 130). Psychologists note that such critically and reflexively thinking, called deliberative cognition, takes place when: (1) people face a problem, (2) they are motivated to search for new solutions because they are dissatisfied with the status quo, and (3) the existing schemata (or knowledge) fail to account adequately for the new problem (DiMaggio, 1997). Tsoukas (2003) argues as well that knowledge is created through stimulating the capability of individuals to draw new distinctions and to develop new meanings to familiar topics. Furthermore, the individual will re-arrange his/her knowledge while being located somewhere certain standpoint or tradition. Accordingly, the capacity to exercise judgement involves two things. First the ability of an individual to draw distinctions and, secondly, the location of an individual within a collectively generated and sustained domain of action (Tsoukas and Vladimirou, 2001: 977). It follows that knowledge creation takes place in a context of different discursive practices or 13

14 interpretive communities. Thus, we emphasise that the knowledge creation process in clientconsultant teams does not happen independent of the client. Rather, exactly through a dynamic interaction and an intensive sharing of perspectives between different interpretive communities it is possible for both client and consultant to reflect on their own knowledge and to reframe their perspectives. In this way new knowledge is created and established practices can be disrupted and transformed (Boland and Tenkasi, 1995). Malefyt (2003: 159) argues that the interaction process between clients and advertisers, for example in the form of a workshop, including making models and metaphors, and finding facts and figures is more of a process of invention than one of any actual representation of consumer culture. Innovation-related research shows as well that communication across boundaries supports innovations (e.g., Tushman, 1977). Thus, we share Clegg et al. s (2004) viewpoint that consultants serve as a source of noise in clients organizations that disrupt established ways of doing and being. Consulting practice produces and, then introduces, new language, deconstructing and disturbing established orders of discourse, translating and mediating between new and old languages and metaphors (Clegg et al., 2004: 36). Because of its very nature as a mix of interpretive communities, consulting companies can be effective by increasing the variety and complexity of interpretations within the client organization. Also, through the introduction of new language and perspectives, consultants can change the way thinking is conventionally organized (Otzel and Hinz, 2001). We argue here that clients can play a similar role in changing the way consultants work. Thus, the relationship becomes a process of mutual disruption and transformation. Consulting companies are aware that they need the exchange of perspectives with external partners if they want to remain successful. As John Stuckey, a director in McKinsey s Sydney office puts it into words We have to be beware of the trap that many large successful companies have fallen into by becoming too introverted, too satisfied with their own view of the world (cited in Bartlett, 2000: 9). Thus, consultancies actively search for innovative clients to work with in the hope that they will get insights into different and novel, emerging perspectives and share new knowledge. In this way they function as interpretive or knowledge brokers (Hargadon and Sutton, 1997; Hargadon, 1998) or cultural intermediary (Carter et al., 2004). Therefore, the clientconsultant interaction becomes a source for knowledge creation for both clients and consultants. This is how one consultant from the Boston Consulting Group expresses it: We have enormous opportunity to innovate in the consulting business because our clients represent an almost limitless laboratory (George Stalk Jr, BCG, cited in Biswas and Twitchell, 1999: 70). 14

15 To sum up, the dynamic interaction and an intensive sharing of perspectives between different interpretive communities will enhance their ability to reflect on their own knowledge and to reframe their perspectives. As a result, new knowledge is created. Hence, based on these arguments, the following hypothesis [H2] is proposed: H2: A variety of perspectives and interpretations within the client-consultant team will be positively associated with the creation of new knowledge and innovative problem solutions. It is clear that to be able to use this variety of perspectives in the problem solving process, clients and consultants need first to make sure that they share their interpretations. Joint knowledge creation without knowledge transfer is not possible. Furthermore, sharing of meaning involves translation, and translation always combines difference and repetition at the same time, and as such, it is a productive and creative process (Clegg, et al., 2004) enabling knowledge creation. If, however, in the process of sharing of perspectives and interpretations the differences between members of the client-consultant team start to blur, and the conformity of interpretations increases, the effect of such wide sharing and internalizing of others perspectives on knowledge creation could turn negative. This is usually the case when clients and consultants work for a longer time together. This question, however, is beyond the scope of this paper. The knowledge transfer and creation process Before some mechanisms for sharing of interpretations are discussed, the process of knowledge transfer and creation within the client-consultant team is illustrated in the following figure 3. The following steps are suggested. 1. Beginning with a discussion of the problem in a familiar way for the client s members shows them the legitimacy of their own perspective; 2. Using different visualization methods (or boundary objects) to point out different aspects/decompositions of the problem; and 3. Stimulating the client s members to generate their own solution paths. This makes them conscious, creative members of the problem-solving process and offers more information to the consultant. This enables the consultant team to offer different insights into the problem, and through this activity, both parties start to share meanings, reflect upon, evaluate, and validate community procedures in a collaborative process. Insert Figure 3 Here 15

16 Malefyt (2003) confirms studying the service delivery process in advertising agencies that the sequence of the knowledge creation process is anything but smooth and continuous. He found out that the process is one that progresses in jumps and starts, from inspirational meetings to setbacks. Brilliant strategies are created only to be altered later to accommodate the wishes of contentious clients, rushed deadlines or legal restrictions (Malefyt, 2003: 139). Not infrequently strategies are dropped altogether in mid development and new ones taken up. Mechanisms for understanding and sharing perspectives Schein (1996: 19) claims that in order to enable cross-cultural dialogues it is important to establish some communications that stimulates mutual understanding rather mutual blame. Furthermore the understanding of what it takes to create effective dialogues is itself coming to be better understood (Schein, 1996: 19). As we showed, Schön s suggestions to enable an open and minimally defensive interaction are not sufficient to explain, how is mutual understanding achieved. In the following, we seek to overcome this shortcoming by introducing some important mechanisms for enhancing sharing of perspectives and reciprocal translations. Thus, before introducing the different mechanisms, the following hypothesis [H3] is proposed: H3: The intensive use of different mechanisms for understanding and sharing of perspectives within the client-consultant team will be positively associated with the development of shared meanings and a common language, and thus with knowledge transfer. Construction of boundary objects. The ability to share a perspective with others presupposes that one s interpretation either in the case of an individual or as that of an interpretative community is made visible for self-reflection. Such a visible representation can become a boundary object and provides a basis for sharing knowledge with others. Star (quoted in Wenger, 1998: 106) defines a boundary object as one that serves to coordinate the perspectives of various communities for some purpose. Such objects could be artefacts (such as cause and narrative maps, labels, physical models, or diagrams), concepts, and encoded information and other forms of reification (Wenger, 1998; see also Raelin, 1997). Because a boundary object belongs, to some degree, to multiple communities, it is a nexus of perspectives. Thus, each community has only partial control over the interpretation of the object (Wenger, 1998). Moreover, boundary objects possess a kind of symbolic adequacy that enables conversation without recourse to commonly shared meanings (Boland and Tenkasi, 16

17 1995). Bechky (2003) shows that in a high technology manufacturing company tangible definitions, such as machine parts, were often necessary to overcome misunderstandings and enable knowledge transfer. In larger management consultancies there is specific training for the consultants where they learn different techniques to visualise and structure their arguments and thoughts according to prescribed models and frameworks. This facilitates their work during consulting projects but also serves to enforce a desired conformity, something viewed as particularly valuable when dealing with global clients. The main contribution of such broad methods and tools (different templates, process phases, general descriptions of a sequence of activities) used in consulting firms is their ability to provide a common framework and terminology for the consultant s work process (Werr and Stjernberg, 2003: 896). The authors state further that these tools are an important facilitator of interaction between consultants, which in turn was referred to by individual consultants as the most important vehicle for learning (p. 896). Creplet et al. (2001: 1527) argue as well that such tools that enable codification of knowledge allow, the spanning across cultural and managerial boundaries through the setting of a sort of meta-language or meta-code. Indeed, all large consultancies have instituted structured IT-based knowledge management systems that seek to provide standard toolkits and sharing of project experience (see, e.g., Weeks and Galunic, 2001; Barlett, 2000) aimed specifically at creating a standard look and feel even to the most exclusive firms. Computer-mediated knowledge transfer within consulting companies is effective, however, only in transferring more general knowledge (i.e. broad concepts, methods and tools). As Reihlen and Ringberg (2004) show, it cannot effectively capture and diffuse more specific expertise. The main reason for this failure is that consulting companies are communities of interpretive communities, and tools, designed during specific client s projects, enhance learning only within the interpretive community where they are created (Reihlen and Ringberg, 2004). Project groups working on specific client problems develop site-specific, local knowledge, and not involved individuals have difficulties in recognizing and applying such situated knowledge (Sole and Edmondson, 2002). Bechky (2003) points out as well that boundary objects are not effective in every situation and tangible definitions will not always be sufficient to create a common ground for understanding. For example, when clients do not have an access to the general methods and tools available to the consultants, the use of general consulting definitions and tools can make the client-consulting interaction more difficult. Thus, the difficulty associated with many consultancies in third world countries is sometimes related to the client s failure to understand rudimentary managerial 17

18 knowledge. Wong (2001) argues that this is the main reason for the serious problems of McKinsey with a client in China.3 Therefore, members of different interpretive communities can assess each other s interpretations or to understand specific artefacts created through others only through direct communication because it enables the sharing of mental models (Reihlen and Ringberg, 2004). Another negative is that prescribed methods and tools may have a standardizing effect on the consultant s interpretations by dictating the range and interpretation of experiences made (Werr and Stjernberg, 2003: 900). For example, many consultants feel that structured knowledge management systems have the impact of potentially creating homogenized foot solders as consultants fail to span newer and different interpretative communities. Both of these points highlight the importance of sharing and explaining of techniques and methods to members of other interpretive communities. Furthermore, it is important to reflect on existing methods and tools and to adapt them. Often, differences between interpretive communities can be overcome only through the involvement of individuals as broker of meanings. The role of brokering. Not all connections between interpretative communities evolve through boundary objects. People who transfer between different communities can understand differences in the interpretations and make them clear to the other members of these groups. Wenger calls this transfer of some elements of one practice into another brokering (Wenger, 1998). A common feature of the relationship of a community with the outside, brokering creates an indirect relation between people or communities where no direct relations exists.4 People who act as brokers or gatekeepers (e.g. Tushman, 1977) possess social capital (Burt, 1992, 1997) and social capital predicts that the return to an individual s human capital depends on that individual s location within, and between, communities. Burt sees the main value of brokering in connecting people who possess different information and in the entrepreneurial value that this bridging creates (Burt, 1997). Wenger (1998) emphasises the complex nature of brokering the new connections across communities that enable understanding and open new possibilities of meaning. This is not simply transfer of information but involves processes of translation, coordination and alignment between perspectives. Brokers also cause learning by introducing into a community elements of another. Tushman (1997: 591) sees in the development of special boundary roles 3 Start Computer Group executives, in a nationwide TV interview, linked the company s losses in two years in a row to advice from McKinsey (Wong, 2001). 4 Burt (1992) calls such disconnected social clusters structural holes and Granovetter (1973) weak ties. 18

19 one way to deal with the difficulties of communication across boundaries and emphasizes the need of specialized gatekeepers, as generally there will be several communities that have to be connected. Werr and Stjernberg (2003) argue that in consulting companies brokering is the most important way of transferring knowledge. Thus, brokering requires enough legitimacy to influence the development of a [community], mobilize attention, and address conflicting interests (Wenger, 1998: 109). It is especially difficult because brokering involves multi membership in communities while, at the same time, some isolation from the rest of the community, which Wenger calls uprootedness (Wenger, 1998: 110). The extensive literature on boundary spanners (e.g., Tushman, 1981) implies that certain individuals seem to be more able and successful as brokers than others. Tushman (1981) found out that brokers have more work-related competence and experience and higher formal status than non-brokers. Bartunek, et al. (1983) argue that people who have developed greater cognitive complexity are more capable than others of applying several different perspectives, thus acting as brokers. Cognitive complexity is characterised through differentiation the ability to perceive several dimensions in a stimulus array and integration the development of complex connections among the differentiated characteristics (Bartunek et al, 1983). Other important characteristics, Bartunek et al. (1983) discuss, are the capacity for understanding others, capacity of introspection and self-awareness, ability to build interpersonal relationships and increasingly broad views of society and social issues. All facets of cognitive complexity increase the ability to gain interpretive dominance over others, which can lead to advantages for the person in a broker position. The following table presents some important mechanisms for sharing of interpretations and their function. Insert Table 2 Here Power relations Acknowledging the existence of different interpretive communities and the importance of both consultant s and client s interpretations for the problem solving process, the interpretive model does not imply that consultants are more powerful than clients. Rather it advocates mutually dependency, which leads to natural questions about the balancing of the relationship. In this section, we will discuss the following questions: How is it decided whose interpretation is better regarding a specific problem? How do involved individuals win ac- 19

20 ceptance for their definition and representation of a problem and for particular solutions? Thus, we expand the reflective practitioner model to include some implications of power for the client-consultant interaction. Unlike the critical and expert models, where power is conceptualized as a possession of one person (or a group of persons) who dominate and constrain others, the interpretive model views power as a social construct that is both a product of collective activity and the medium by which it is developed and enhanced (Clegg, 1989; Blackler, 2000; Contu and Willmott, 2003b). Power is expressed in and through disciplinary practices and in and through struggles against or in resistance to such practices (Clegg, 1989: 109). Thus, following a Foucauldian view on power we argue that power relations and interests amongst groups and actors are dependent on knowledge, or in Clegg and Palmer s terms (1996: 4) on the recipes in use. These authors regard the formation and accumulation of knowledge that makes the translation of different interpretations possible as simultaneous mechanisms of power power makes knowledge as knowledge makes power (Clegg and Palmer, 1996: 5). Malefyt (2003: 153) shows that for the case of advertising agencies power resides not in the possession of raw information about consumers and brands, but in the way in which the meaning of that information is presented, understood and interpreted for the client. Thus, interpretive communities or discursive practices (Foucault, 1972; Clegg, 1989, 2001) effectively constitute power. Some of these practices will have greater power than other: If we treat organizations as communities of practice, we will find that some communities attempt to enrol, or betray, the others (Fox, 2000: 863). When partially fixed meanings or interpretations occur power is at work. Historically, power is the apparent order of taken for granted categories of existence fixed and represented in myriad discursive forms and practices (Clegg, 2001: 138; see also Foucault, 1972). So how do people sharing a specific interpretation achieve a partial fix of their view? Whereas the notion of interpretive dominance conceptualises a belief system as an active arena, where interest groups compete to impose their preferred psychological order onto nonbelievers (Meindl et al., 1994: 291), how do they gain interpretive dominance over others? How does the process of enforcing one s own perspective in a discourse with others evolve? Callon s (1986) four steps or moments of translation 5 are a useful concept to explain the mechanism of power emergence between interpretive communities. First, in the stage of problem formulation members of an interpretive community define a problem in a 5 Here we are not going to discuss actor network theory (ANT); rather we use some of its basic concepts to illustrate our model. See Callon and Latour (1981) and Callon (1986) for a more extensive discussion. 20