Introduction

In an era where a strategic command of global resources is becoming a requisite for success, firms are struggling to successfully encourage collaboration across their onshore and offshore sites (Lipnack and Stamps, 2000). This challenge centers on how to distribute work, responsibilities and leadership across sites and then re-integrate them into a coherent whole, where decision making is well-coordinated and workers collaborate effectively to complete tasks. The literature on globally distributed teams generally frames the impediments to coordination and collaboration in terms of communication problems due to the divergent nationally-based cultural attributes of the sites, language barriers, and the limitations of information and communication technologies (ICTs) (Kankanhalli et al., 2006–2007; Mihhailava, 2007). As a consequence, firms are pursuing a dual strategy of improving the communication infrastructure (in terms of ICTs and formal global process standards) coupled with cultural trainings. However, in this paper we consider how this approach only addresses part of the equation for achieving integrated collaboration. Ultimately, globally integrated collaboration requires an approach that allows both the managers and the workers to examine the multitude of shifting factors that are rooted in the context of work.

In seeking to explore this context in more detail, we focus on how collaboration is influenced by enacted (rather than espoused) organizational strategy. More specifically, we demonstrate how an enacted organizational strategy can create structural impediments even while this strategy is focused on facilitating collaboration between sites. We do this by drawing on world-systems (WS) theory (Wallerstein, 1974, 2005) as a heuristic framework. This encourages a re-orientation of the unit of analysis from the attributes of individual sites (and their particular national cultural characteristics) to the social dynamics across sites (and the structural and interactional factors that influence them). By viewing the distributed sites through the WS lens, the problem onus shifts from the individual worksites or a grouping of sites according to national location to the relationships among the sites. In other words, this lens demonstrates how it is important to look at all of the distributed sites and their relationships with each other. This allows for a comprehensive view of the factors that contribute to the formation of actual and perceived relationships between sites, groups, and individuals within the context of the broad organizational strategy.

We apply this WS framework, coupled with a focus on the use of ICTs, to examine the attempt to build collaboration in the distributed software development department of a financial services company, referred to as GLOBALIS (for 'Global IS'). The case illustrates how tensions in social relationships across sites were influenced by the socio-politico-organizational context. Moreover, by focusing on emergent practices, the case also illustrates how some units in GLOBALIS were able to overcome these structural impediments to develop positive social relations that facilitated collaboration.

The rest of the paper is organized as follows. The next section outlines the traditional approaches to global collaboration and summarizes the theoretical perspectives used in the paper. This is followed by a description of the case study. In section four we explore the application of world-systems theory to global collaboration, bringing out the organizational factors that impact effective collaboration. This is followed by an analysis of a project where the proactive socialization by the project manager enables the team to overcome the impediments of global collaboration. The last section elaborates on the implications of the findings and future directions of research.

Traditional approaches to global collaboration

As companies are attempting to transform themselves from being 'multinational' to 'global,' they are facing the challenges of how to distribute, coordinate and integrate work across sites. Although global work is not new, the extent to which companies are attempting to achieve cross-site collaboration constitutes a new model of global work. In looking at this issue we can draw upon Thompson's (1967) seminal work on task interdependencies. This work identifies the processes through which task inputs are combined to complete a whole piece of work. Thompson identifies three forms of interdependence: sequential (where subtasks are completed in a specified sequence and the output of one task is the input of the next – as in assembly line manufacture); pooled (where subtasks are performed separately and outputs are pooled in an additive way); and reciprocal (where sub-tasks must continuously interact because the outputs and decisions from one will have a direct impact on the other). Many companies (like GLOBALIS) are attempting to distribute tasks across sites that need to be combined in a reciprocal way in order to successfully complete the project. This is because expertise is often distributed across sites in ways that do not align well with the various tasks that constitute a particular project and moreover, because reciprocal interdependence is more likely to lead to innovative solutions. However, as Thompson indicates, this form of interdependency requires more extensive collaboration than would a pooled or sequential approach to task interdependencies. In other words, there needs to be extensive interaction and knowledge flow between members at the distributed sites in order to complete the project tasks.

Thus, a key challenge of global collaboration is how to effectively accomplish highly interactive, project work virtually. Three key issues have been discussed in the literature as helping us understand such distributed collaboration: (1) overcoming cultural differences; (2) developing the ICT infrastructure; and (3) establishing standardized global processes. We next consider each of these issues in turn and illustrate how context is often under-theorized in relation to each area. We then turn to an examination of WS theory that provides a heuristic framework for thinking about this context which we then apply to our case.

Culture and distributed collaboration

National and local cultures are seen to impact distributed work (Krishna et al., 2004). The most often used conception of nationally-based culture is provided by Gert Hofstede (1981, 1991). His model includes a national culture ranking across five traits and has been a crucial part of research on global work and cultural diversity trainings. For example, Søndergaard (1994: 448) found over 1000 citations of Hoftede's 1981 book alone in journal articles during the period 1980 to September 1993. The traits surfaced by Hofstede are said to be implanted in the minds of culture members as a type of value programming, and they impact the mindsets, behaviours and decisions of the cultural member.

Despite the popularity of Hofstede's conceptualization of culture, there has been a growing chorus that questions the validity of Hofstede's findings as well as the applicability of his model. Westrup et al. (2003: 19–20) note that his approach 'promotes a static formulation of culture and can easily lead to treating culture as a causal agent.' Ford et al. (2003: 9) summarize three main shortcomings of Hofstede's approach, including that cultures are: (1) assumed to fall along national boundaries; (2) viewed as static; (3) are assumed to be homogeneous and devoid of subcultures. Given that Hofstede's research was conducted in the 1970s, there are concerns over whether the findings are generalizable to today. Avison and Myers (1995: 52) go so far to say, 'the prevailing taken-for-granted view of the culture concept within the IS research community needs to be abandoned.' Despite these shortcomings, this conception of culture is often at the base of organizational 'cultural training programs,' as it is in trainings hosted by GLOBALIS.

Beyond the limits of the mainstream national cultures framework, Huang et al. (2003) and Galliers (2003) remind us that there are organizational cultures and sub-cultures to consider. Knorr-Cetina (2000) speaks of epistemic cultures, which refer to the localized practices expressed by professional cohorts. Van Maanen and Barley (1984) discuss the presence of occupational communities in which a group of people share in the same kind of work and derive their identity from that work. As Liberman (1995: 119) observes, 'Analyses of intercultural communication too frequently read like they are rule-governed events; however, participants rarely perceive them that way.' These studies remind us that greater emphasis needs to be placed on studying how culture and communication manifest themselves in specific contexts.

ICTs, global processes and distributed collaboration

In relation to the ICT infrastructure, systems and databases provide the backbone for performing distributed work. ICT firstly enables workers to communicate and engage in collective problem solving, and research has considered how ICT can best support this communication. For example, information systems research on global IT has examined the effective and innovative use of communications tools such as email (Sproull and Kiesler, 1986), video conferencing (Meier, 2003), and IM (Hersleb et al., 2002) in distributed work. Secondly, since software projects require both application-problem-domain knowledge and technical software engineering expertise, it is also important for the ICT infrastructure to support knowledge management functions. Thus, the focus of knowledge management research in global IT work has been on knowledge transfer about the application problem domain from the client to the vendor organization (Robillard, 1999) and the knowledge of polices, processes and systems from the onshore group to the offshore groups (Tiwana, 2004). However, while ICTs are necessary for communication and knowledge sharing in distributed environments, they are by no means sufficient. It then becomes important to look at the social aspects of the context in which this knowledge sharing and interaction are taking place.

The ICT infrastructure will be combined with global processes to facilitate collaboration. That is there will be formally defined workflows, policies, procedures, metrics, skills and interfaces across the distributed worksites. Thus, research on process improvement models such as the Software Engineering Institute's Capability Maturity Model and ISO standards (Bamford and Deibler, 1993) and Spice (Dutta et al., 1998) demonstrate how adopting these standards can create improvements in requirements management, project management, configuration management, development methodologies and testing and validation (Damian et al., 2002). In other words, organizations with matured business processes have well-documented global processes in terms of policies, procedures, workflows and responsibilities (Hammer, 2007).

These standardized processes, along with the accompanying ICT infrastructure, become the environment in which global collaborative work is supposed to take place. However, research also reminds us that technological use, whether of ICT or global processes, in the pursuit of collaborative work is not something that can be planned a priori, but is something that is ad hoc and emergent. As Harper and Hughes (1993: 142) state, 'Controlling actions are not then to be looked at simply as the following of procedurally defined rules but as the contingent outcome of processes of interpretation as to how the rules fit the case to hand.' Thus, while rules and procedures may give the appearance of establishing order, that appearance often is illusionary. Bannon (1993: 8) echoes this sentiment, 'Information-flow diagrams of office activities do not, in any literal sense, specify how work actually is accomplished.' It is in the actual doing of the work that collaboration lies, and not the mandates of how work should be done: 'people fill gaps in technology, and construe their action together with their peers in ways that are more or less in line with official organizational policy' (Koskinen, 2000: 18).

This line of research indicates that while increased availability of communication technologies means that communication can take place; it does not mean that it will take place. Moreover, more communication does not automatically mean increased collaboration vs for instance increased conflict. Likewise, rules and production methodologies do not mean work will proceed in an orderly fashion. People are not 'cultural dopes' (Garfinkel, 1967) passively following rules and adhering to structural dictates. Rather, they are active participants in the creation of an emergent and situated social order. It then becomes important to understand the impact of organizational procedures and strategies, and how to facilitate collaboration in this global context. The key for organizations, thus, becomes how to provide the environment in which collaboration and cooperation can take place, using the tools and processes provided. In other words, context matters, and in order to understand how context matters, we next turn to a consideration of the WS framework.

WS theory

Viewing the world as composed of discrete entities, with national boundaries creating the basis for separation, was the dominant perspective in social sciences until a fundamental shift occurred with the advent of WS theory by Immanuel Wallerstein (1974, 2005). Skocpol (1977: 1075) notes, 'Immanuel Wallerstein's The Modern World-System aims to achieve a clean conceptual break with theories of "modernization" and thus provide a new theoretical paradigm.' This paradigm shift provides a lens through which global organizations can be viewed as continuous wholes rather than discrete entities. Doing so allows for attention to be paid to how the relationships between sites create the organizational 'reality' of everyday work. Thus, our focus becomes centered on workplace interactions and how these interactions build the structure of the global organization.

There are different versions of WS analysis (Skocpol, 1977; Evans, 1979) and they are exceedingly complex in their totality. We do not intend the use of WS theory in this paper to be support for or a refutation of WS theory as it has originally been developed or applied. Rather, we draw inspiration from it as a heuristic device for understanding globally distributed collaborative work, namely that in order to fully understand the dynamic of collaboration within a distributed organization (especially globally distributed organizations), you must treat the organization as a system rather than focusing on the sites as discrete entities.

There are two foundational ideas in WS theory. The first is to treat the entire world as a self-contained social system, with its own internal logic, mode of operation, unified and complete division of labor, and internal exchange. Or, as Janowitz (1977: 1091) observes in his analysis of Wallerstein, in WS theory 'it is necessary to analyze change in terms of the relations among nations, that is, the "social system" that is created by the linkages among nations.' In terms of what defines a world system, Chase-Dunn and Hall (1993: 856) state, 'We use the term world-system to refer to the actual social context in which people live and the material networks that are important for everyday life.' It is the emphasis on the social context and material networks that we use as a basis for our analysis of globally distributed work.

The second key idea is the system's single division of labor where the world economy is functionally and hierarchically separated into core, periphery, and semi-periphery. Since production processes change and shift as a result of technology, the concepts of core and periphery properly refer to the location of production processes in the world economy, and not nation-states per se. At the same time, nation-states are often associated with their position in the world economy whether it be as core, semi-periphery, or periphery nations. That is, some countries occupy a core-like production position, or periphery-like production position, etc. While this suggests that countries can move up or down the ladder and change positions over time, such movement is not common and will only occur as the result of a qualitative shift in the production that takes place in that country.

Since profitability and level of technological innovation define and determine the type of production, core countries monopolize high-tech, high profit enterprises, while the periphery countries are underdeveloped (McCormick, 1990: 126). In practice, the core maintains control of technology up until the point that the technology becomes routinized. Furthermore, since core-like productions are highly profitable, involve new technologies, innovation, etc. they are quasi-monopolies and they maintain tight grip on technological innovation.

However, over time the quasi-monopolies of the core become 'self-liquidating.' As such they exhaust themselves and their profitability drops. It is only after production becomes less profitable that production (technology) shifts to the periphery. It is at this point that existing technologies might be transferred from the core to the periphery in order to take advantage of the cheaper labor markets. Thus, while technology can move from the core to the periphery, the technology typically only moves to the periphery after it has become routinized, and savings are possible only from cheap labor. In summary, production moves to the semi-periphery, and eventually to the periphery in search of cost-savings from cheap labor.

As a heuristic device for framing our case analysis, WS theory thus focuses our attention on the system of social relationships across the distributed sites of an organization; it draws our attention away from a focus on deficiencies (or strengths) of individual sites. Moreover, within this social dynamic across sites, we need to identify whether sites operate as core or periphery, and then consider how this pattern of relationships influences collaboration. We turn to explore these dynamics through our case study, presented next.

Study site and data collection

GLOBALIS' headquarters is in Boston, MA with multiple solution centers in the New England region. Almost 15 years ago GLOBALIS established solution centers in Texas and Utah, and it has been operating two wholly owned solution centers in Ireland for the last 10 years. Three years ago it launched its first solution center in Gurgaon, India and a year and a half ago began a full-services solution center in Bangalore. Although GLOBALIS has considerable experience in establishing geographically distributed solution centers and is well aware of industry's best practices for dealing with issues and problems of geographic dispersion, for example, distance and time zone differences, cultural differences, loss of communication richness, etc., it is facing difficulty in forming a cohesive community among these distributed sites. The data used in this paper were collected over a span of 3 years through an ethnographic study of IT solution centers in the US (five sites), Ireland (two sites), and India (two sites). In our approach we used an interpretive epistemology (Walsham, 1993; Klein and Myers, 1999), using the narrative interview technique (Bauer, 1996) with the objective of collecting and collating 'thick descriptions' (Geertz, 1973). In our approach, then, the case was designed to seek 'validity...not [from] the representativeness of the case in a statistical sense, but on the plausibility and cogency of the logical reasoning used in describing the results and in drawing conclusions from them' (Walsham, 1993: 15).

A multi-layered data collection process has been employed by the research team to develop a multi-dimensional view of the projects and the subjects of the study. This allowed us to collect multiple interpretations and so helped us to improve the 'plausibility and cogency of our interpretive accounts' (Klein and Myers, 1999). To provide GLOBALIS management an understanding of how its global workforce collaborate, 3 years ago our team began observing the development of the GLOBALIS' global delivery process. For a period of 12 months (July 2005–June 2006) our research team tracked four IT projects using a workplace studies paradigm. This included interviews, site visits, observations of video conferences and conference calls, and frequent discussions with and presentations to GLOBALIS personnel regarding the team's findings. Thus, the project was carried out in a quasi-Participatory Action Research (see Whyte, 1990) framework in that it was hoped that the findings of the project would be beneficial to workers, management, and the organizations in terms of developing better relationships among distributed personnel resulting in higher quality work and a better global strategy.

Interviews were conducted with 40 employees of GLOBALIS that included six senior management personnel, six project managers and 28 workers associated with the four projects. These interviews occurred in both individual and group format and most of the interviews happened in person. Each interview lasted approximately 1 h in length. These interviews were semi-structured and conversational in nature, and they covered a range of topics related to GLOBALIS, such as:

• How project success is viewed and defined?
• Planned use vs actual use of standardized software development and project management processes.
• How communication and information technologies are used?
• What are the knowledge transfer and knowledge sharing processes?
• What are the key challenges and how those challenges are addressed?
• What, if any, are the processes to encourage the formation of social and personal relationship?

In order to get a more fully developed sense of the worksites, we made repeated visits to the onshore and offshore sites associated with the project. During these visits, we also observed the nature of the work associated with the project, especially meetings and other situations where people from the various sites interact with one another. The Ireland sites were visited in 2003 for 1 week, and again in 2006 for another week. The India sites were visited over 10 days in 2006. During the course of the project, the sites in the New England area were visited intermittently in order to attend meetings, conduct interviews, and deliver reports. No other sites in the US were visited. We also attended a cultural training program and collected documents from this program as well as other project documents, as they were made available to us.

Data analysis and findings

The espoused strategy: creating a unified work environment to support global collaboration

For a variety of reasons, GLOBALIS have explicitly shunned the idea of outsourcing because they want all work to be done in-house. Their stated rationale for going global is simply because they are seeking to find and use the best talent, at the best price. Thus, senior executives in the business have a very clearly stated offshoring strategy – to develop a unified business where work is undertaken collaboratively in globally distributed projects. In its attempts to establish collaboration, GLOBALIS is using approaches used by other organizations, which includes implementing and diffusing its organization culture across sites, standardizing both production and communications technologies, cultural training, and occasional travel. We describe these first, and then explore the problems that surfaced despite these attempts to unify.

To generate a feeling of a shared community and organization culture, GLOBALIS has created a similar 'feel' across sites through interior design, colors, and presence of company symbols. Company executives have repeatedly said, 'You can be in Bangalore and you could think that you are in Boston.' While an oversimplification, it speaks to the company's interest in developing a unified workplace.

The methodologies, software and systems used at the different sites of GLOBALIS are standardized as well. Three years ago when the Bangalore site was launched, GLOBALIS engineered a global delivery model built around an enterprise project planning and monitoring system to identify and allocate worldwide resources based on the time to market, quality and cost attributes of the project. This enterprise system allows authorized users to track and monitor projects from initial leads to completion with defined templates and procedures to support the intermediate activities. All the sites use the same set of tools for requirements management, configuration management, applications development, testing and reporting (although they may not use them in the same way). In addition, each site is equipped with worldwide telephone service, standardized teleconferencing, video conference capabilities, SameTime group ware, email, and mobile devices. Recently, GLOBALIS has made available a variety of 'social software,' which is intended to facilitate communication and information exchange across and within sites. In summary, in terms of tools, methods, systems and policies the GLOBALIS sites are essentially identical.

In terms of cultural training, GLOBALIS contracted with a consulting company to train their employees on how to do business with India. This training was a 1-day program where managers would attempt to learn about the traits and characteristics of Indian culture while developing an awareness of their own culture. The training program used the dichotomous cultural categories based on the work of Hofstede and found in much of the writing on national culture in global organizations. While efforts were made to caution attendees from over-generalizing the categories attributed to India, nevertheless these depictions were the basis of the training, and for calling attention to potential points of conflict between American and Indian workers. Even though there are GLOBALIS sites in Ireland, there was no training program on doing business with Ireland. Furthermore, Indian employees did not receive training on how to do business with the United States. Rather, the training experienced by Indian employees was more focused on how to work within the organization.

Finally, while travel does take place, it has become a more difficult proposition because of the entry restrictions for foreign nationals into the US, the cost of travel, the extensive geographic distribution of work, and also a general unwillingness of American workers to travel overseas (especially for a protracted period of time and to India in general). Thus, travel has become more limited than most senior managers would like. Also, when Indian workers do travel to the US for extended periods, the exposure to their American counterparts generally is limited to worksite interactions focused on training.

Despite the best intentions of senior management, the implementation of the approaches outlined thus far has not yielded the intended consequence with projects often not meeting deadlines or not fulfilling project goals. In many instances, as acknowledged by the managers and employees themselves, this is because collaboration is poor – communication is often slow, knowledge sharing is difficult, and tensions between sites are often high. Turnover is also very high at the Indian and Irish sites. In the next section we examine why this has occurred, using the WS framework to explore how core–periphery relationships between sites developed despite the rhetoric of global unity. This allows us to see how, in some ways, it is the very approaches that GLOBALIS management has used to try and create unity that have had the opposite effect to that intended, so making collaboration more, rather than less, challenging. Thus, we will examine how these approaches have contributed to creating the barriers they were trying to remove. The case of GLOBALIS demonstrates the potential shortcomings of these approaches.

The enacted strategy: creating core–periphery relationships between sites that undermined collaboration

We will present a series of observations that suggest that attempts to manage global relationships through a standardized, top-down strategy yielded limited positive outcomes, and resulted in social relationships of perceived inequality across sites, actually encouraging perceptions of a great social distance between sites. At the same time, there were positive examples of global relationship management that occurred in certain project teams. These successes had to overcome the core/periphery mentality that developed in GLOBALIS, and did so through the development of an interpersonal and collective (vs intergroup) orientation towards each other. In the final section we consider how this shift in awareness is key to facilitating globally distributed collaborative work.

Observation 1: asymmetrical interactions in technologically mediated communication

We observed monthly video conferences and regular conference calls between New England and other distributed GLOBALIS personnel. For the most part, these meetings were observed at the main corporate offices in Boston. During these meeting, we observed numerous asymmetries in the interactions. One asymmetry concerns the times of meetings. The meetings generally took place at 8:00 a.m. Eastern Standard Time (EST), which meant that it was 1:00 p.m. in Ireland, 7:00 p.m. in India, 7:00 a.m. in Texas and 6:00 a.m. in Utah. Thus, people in New England were beginning their workday with a meeting, while workers in Ireland were having their day interrupted and people in Texas, Utah and India had to work outside their normal work hours.

These monthly video conferences were meant to provide senior managers the opportunity to report on employee allocations, project status, and potential allocations of future work. However, more open discussions of future directions, organizational strategies, and general brainstorming regarding projects would take place. When a standard reporting format and agenda was followed, the allocation of speaking turns was based on who was next in the agenda. When the meetings took more of a free-form structure, turn allocation was centered in the Boston office, resulting in an asymmetry where people at the distributed sites were not involved (even with the technological capacity to do so). Gaze, which is an important component in next speaker selection (Goodwin, 1980; Atkinson and Heritage, 1984; Heath, 1984), would be directed at those sitting at the table in Boston. Given the way the video conference technology functioned, only the last person to utter an audible sound at a distributed site would appear on the monitor in Boston. If you did not make any sound during the meeting, you would never be visible to people elsewhere. This resulted in people muting their speaker to avoid being on camera.

Another asymmetry was rooted in who 'ran' the meeting, who was able to ask questions, and the nature of those questions. The meeting was 'run' by a manager in Boston, meaning time was kept from Boston, the meeting began from Boston, the agenda was followed or altered from Boston, etc. Questions asked from the distributed sites tended to be clarification questions, while the questions from Boston were accountability questions (why something was not done, project delays, employee allocations, etc.). We did not observe persons from the distributed sites asking accountability questions. Thus, accountability was a uni-directional practice originating from Boston.

Observation 2: cultural trainings emphasizing difference

We attended one full-day training program on Indian culture that was designed and delivered by an external consultant to GLOBALIS employees in the New England region, and were able to carefully review the training materials. The focus of the training was on how to manage the work of GLOBALIS employees at the Indian sites. Despite statements that you cannot generalize traits of any culture, the training material presented highly stereotypical descriptions of Indian culture, such as the one given below on the lack of initiative in Indian culture:

Because of the hierarchical nature of Indian society, the most senior or elderly male generally has the most authority in the workplace. There is little tradition in India of individual initiative or innovation. Employees generally wait for instructions and then do as they are told. They don't ask for or desire more responsibility. There is a strict adherence to the division of labor and acceptance of one's roles.

Beyond being stereotypical and condescending, it also is completely wrong. On our visit to India, we heard of the desire for higher-level work, to engage directly with customers, added responsibility, for greater research and development. This passage also completely contradicted statements made by management when marketing the services of the India sites, with these portrayals emphasizing the skills and initiative of 'Indian workers.' Nevertheless, the cultural training that was provided often framed the interactions between Indian workers and their American counterparts who have received this training, with the Americans explaining problems that were encountered during collaborations with workers in India through the simplistic discourse of cultural differences. In many ways, then, the cultural training was legitimating 'sophisticated stereotyping' (Osland et al., 2000).

During our visit to Bangalore and Gurgaon, we asked the Indian workers whether they took any culture training program on working with their US counterparts. Most of the Indian workers we interviewed barely remembered the training because it was a short 3 h training that focused on how to communicate in English. All they could remember was that they were told not to use the word 'but' in their emails. Apparently this advice was meant to counteract the stereotype 'Indians can't say no.' The way the two training programs are framed, namely US culture training focusing on how to manage work and Indian culture training focusing on how to communicate more precisely, actually establishes a hierarchical relationship between US and Indian sites.

Observation 3: unequal project and system ownership

Most projects and systems are owned by the Boston site. The Texas and Utah sites own systems that they manage in order to support their external customers. However, the sites in Ireland and India rarely own systems; during our research we only found two examples of temporary ownerships. The Bangalore site owns an Enterprise Problem and Change Management tool used by the various business units within the firm to manage changes and problems in their IT environments. This vendor product was customized by GLOBALIS to fit its business requirements. It is now an 8-year-old enterprise system that has undergone significant enhancements. Although the Bangalore site is responsible for upgrade-type development work and back-end support, the Texas site owns the product and performs requirements gathering and front line support.

The Ireland1 site owns a program that enables the management of access privileges and ensures security reporting compliance. Its users are all the global business units of GLOBALIS. The product is in 'keep the lights on' phase and there are only two business analysts at a New England site and 10 developers and support personnel at the Ireland1 site. Moreover, we were told by those at Ireland1 that the program was handed to them because of too many customer complaints; in other words, the Ireland1 site took over the program when it had become highly problematic. On taking over this program, the Ireland1 site assumed the ownership to develop a center of expertise in computer security in Ireland. However, recently, a decision was made to build an entirely new security system; and the decision has been made that this new product will be owned by another New England site, not the Ireland1 site, much to the consternation of those in Ireland1. Thus, once a program was moving from maintenance to development mode, it was transferred from outside the US back to Boston.

Regarding the implementation of technology, at one of the monthly video conference meetings, we observed that the manager at the Gurgaon site recommended that GLOBALIS adopt an Instant Messenger tool that the Indian site felt would make communications more effective. This recommendation was quickly rejected. Later we learned that decisions regarding which technology to use and standardize on are initiated and made at the Boston site. Thus, decision-making power lies primarily in the Boston site, with distributed sites retaining less authority regarding what technologies will be used.

These three observations all demonstrate how approaches that were supposed to unify the sites, in fact had the opposite effect to that intended. Applying the WS framework, we can see how these different approaches helped to generate core–periphery relationships across the sites – the Boston HQ site was able to control interactions through the use of ICT, emphasizing how this site is central or core relative to other sites; the cultural trainings emphasized differences and reinforced a client–vendor (or core–periphery) mentality; and project ownership was very difficult for remote sites to secure, emphasizing their periphery status. It is hardly surprising that these core–periphery relationships impeded genuine collaboration across sites. Before we turn to our discussion, we introduce two further observations that provide examples of how these core–periphery relationships can be overcome where more personal relationships are given the opportunity to develop.

Observation 4: rule following vs ad hoc'ing

In one GLOBALIS project, we witnessed how rule following disrupted global relations, while ad hoc'ing the process resulted in building relationships. A team in an Indian development center was assigned to work with a team in an Irish development center. The Irish workers had prior experience with the customer for the project, while the Indian workers were new to the project and the customer. Because of time constraints, the Irish team wanted to build a prototype based on their knowledge and experience with the customer. They asked the Indian team to start coding from a general set of requirements that the Irish team provided. The Indian team did not want to start work until the requirements were formally specified. While the Irish team felt that Indians were being too process oriented, the Indian team claimed they were following the protocols laid out by the organization. The Indian team was following the organizational processes; the Irish team wanted to ad hoc the process.

The team manager of the Indian team happened to travel to Ireland for unrelated training. The manager of the Irish team found out about this visit, and was annoyed that he was not told of this visit to arrange a meeting. The Irish manager invited the Indian visitor to stay for three more days to meet with him and other members of the team. This face-to-face meeting allowed the manager of the Indian team to appreciate the situation of the Irish team and understand that uncertainties needed to be resolved before the system requirements could be thoroughly specified. By the third day, the Indian manager and Irish team was making jokes and developing more of a personal relationship. This resulted in the team members in India and Ireland starting to converse and joke through email, signifying a breakthrough in the team relationship. The building of the personal relationship among the team members, together with an increased understanding of the project context by the Indian team manager, resolved the conflict and work began on the project. While following the rules disrupting collaboration, ad hoc'ing the process allowed the sides to build a relationship.

Observation 5: 'social engineering' and building relations

The last example also demonstrates the importance of personal relationships in facilitating distributed collaboration. One of the projects we tracked involved a 9-month long human resources and payroll application project, which was the largest attempted by GLOBALIS. The business analysts and systems testers were located in New England, the project management and half of the development team resided in Ireland, and the other half of the development team was in India. The Irish team, who had prior domain knowledge on the project and requirements, were to recruit and work with the Indian team. The Indian employees, many of whom were new to the organization, had no domain knowledge. Thus, the issue of transferring knowledge and building up the Indian team's capacities became a paramount issue.

A focus of recruitment was the interpersonal skills of the team members. The Irish management team used conference calls to ascertain the Indian applicant's ability to communicate through technology. In fact, interactional competence was given greater importance than technical ability, which was believed to be more easily taught. The project started with a team kick-off, where members were asked to post their photographs, and say something novel or funny about themselves. The managers initiated a buddy system so workers could get to know one another and form a virtual workplace community. Also, some members of the Irish team were sent to India, during which time there was opportunity to exchange personal information as might occur during a normal workday. Upon return, the Irish workers were able to facilitate relationship development between those who never met face to face, as with the Indian manager in the previous example. By facilitating relationship transfer, the workers ultimately were able to facilitate collaboration.

DISCUSSION AND CONCLUSIONS

Our case demonstrates how the strategic decisions made regarding how work should be done and who should do it can create a global hierarchical structure, even where managers espouse the ideal of a unified company (see Figure 1). Since all the strategic decisions concerning the selection of the production technologies, ownership of core programs, and IT alignment with the business are made at the headquarters site in Boston or the neighboring New England sites, from a WS framework we can locate the Boston and other New England sites as part of the core. Next, since the sites in Texas and Utah are allowed to own systems to better serve their local customers, these sites possess the full-lifecycle development technology. As a consequence, the Texas and Utah sites can be positioned as belonging to the semi-periphery. The two sites in India with no system or product ownership role belong to the periphery. The remaining two sites in Ireland are somewhere between the semi-periphery and the periphery. The extent to which process standardization solidifies this structure makes integrative or reciprocal collaboration (Thompson, 1967) exceedingly difficult to achieve.

Figure 1.

GLOBALIS decision-making hierarchy.

Full figure and legend (78K)

Adopting a WS perspective to globally distributed organizations allows one to see the impact of core/semi-periphery/periphery relationships among the sites. First, such arrangements can impede the creation of trust among the sites. As von Krogh (1998: 136) has found, 'Good relations purge a knowledge-creation process of distrust, fear, and dissatisfaction.' Inversely, poor relations make it difficult to achieve knowledge sharing. If the relationship is marked by social distance, collaboration becomes very difficult. Second, since periphery sites typically do not get to own systems and products, there is little opportunity for workers at these sites to interact with the customers and develop domain knowledge. Since customer contact and domain knowledge is held by core sites, it is much more difficult for periphery sites to move up the production value chain. This lack of opportunity to develop domain knowledge increases attrition and thus reduces the ability to develop good relationships and collaborative work. Third, since the sites on the periphery are not included in direction setting and new technology selection, organizations miss out on the contributions of its global staff and the full benefits of creativity that diversity brings. Fourth, the WS perspective clarifies a variety of behaviors that are incorrectly attributed to national culture. For example, behaviors such as reluctance to push back, doing what one is told, showing little initiative and leadership, lacking sensitivity to deadlines, etc. were often associated with 'the culture of India' by their US colleagues. However, the cause for these behaviors can also be traced to the core/periphery relationships (and client/vendor mentality) that frame global relations in GLOBALIS. Moreover, in GLOBALIS we found that workers in Bangalore and Gurgaon work more easily with the Texas and Utah sites compared with the Boston and New England sites. Furthermore, they were most comfortable working with the Ireland sites. This suggests that viewing the US as one unified culture (as Hofstede, 1981) for studying collaboration is inadequate.

Our case also shows, however, that there were examples where this core/periphery relationship was overcome. In these cases it was the establishment of personal relationships that was key. In WS theory, power is postulated as being rooted in the global structure of production relations in which certain locations maintain control over the means of production and other locations are exploited for their possession of the raw materials for production. What we have found in our case study is that power is rooted in the everyday interactions that make up the basis for organizational structure. Even though the strategy of the organization was to have a 'flat' organization, the actions of managers created a hierarchical organization. At the same time, certain managers and workers were able to overcome these structural barriers to achieve the organizational goal. These emergent personal relationships helped to facilitate a sense of commitment to shared practices, which became the fundamental element in establishing a 'mutually intelligible social order' (Rawls, 2005). However, our results also demonstrate that facilitating travel so that people can meet face to face does not necessarily create the social relationships that can overcome structural impediments. For example, when the Indian workers traveled to the US, we observed that there was little interaction after work hours or even during lunch breaks, unlike when the Indian manager traveled to Ireland, or when the Irish workers traveled to India. This meant that social relationship development was limited during these visits to the US, and actually creating feelings of alienation and isolation, rather than shared commitment.

These findings lead us to postulate a new global collaboration model that is shown in Figure 2. The figure illustrates the situated nature of distributed relationships, where nationality is not the deciding factor in the amount of social distance between sites. For example, despite being 'US' sites, the type of decision-making power held by the US sites can differ significantly. The 'core' site possesses the ability to make 'strategic' decisions regarding the ultimate direction of the organization and how its resources will be allocated and used. Distributed sites can possess tactical and operational decision-making power (in terms of system and product ownership, work allocation, etc.). Thus, being a US site does not mean the site has power to control the direction of the organization. Furthermore, distance is not the determining factor in where a site resides in the global system. In GLOBALIS, the Irish and some US sites share a similar position in the organization in terms of decision-making and evaluative authority. Likewise, the Indian sites typically are allowed to make operational decisions only. It also should be mentioned that this distribution can change based on the specific project, although the core site retains central authority.

Figure 2.

Integrated model of distributed work.

Full figure and legend (124K)

We found that when sites occupied a similar stratum of the organization, they were able to use this position to develop social relationships. This was based on a feeling of 'being in the same boat' vis-à-vis the core site. At the same time, occupying the same position could also create a feeling of competition. Thus, it became incumbent on the managers to facilitate an environment in which rapport could be developed. This was done through purposively creating opportunities to learn about one another and generally 'humanize' co-workers at the distributed sites. It also was the case that relationships across strata of the organization could be problematic because of feelings of entitlement and resentment created by the structural arrangement of sites. Even though the organization was attempting to create integration through their global strategy, the impact often was quite different.

Figure 2 focuses on the interaction between workers as the foundation of workplace culture. In this milieu, national culture is but one of many variables that can impact these interactions. Too often, 'misunderstandings between members of different cultures are assumed to be comprised of consciously remembered and constructed differences rather than relating to situational factors' (Miller, 1995: 144). Our model reintroduces the importance of the situational context while retaining the impact of demographic, occupational and individual factors. It also demonstrates that despite 'obvious' differences, it is possible for people to create a shared identity and culture. As Seely Brown and Duguid (2002: 140) observe, 'People with similar practices and similar resources develop similar identities.' The goal of global managers is to create the space in which these shared practices can develop.

The points presented in this paper have theoretical and managerial implications. Theoretically, we expand the pool of concepts in examining distributed collaborative work by including WS theory. This provides a linkage between social (political) science and management and IS literatures. While WS theory is more nuanced than presented here, it provides a heuristic to explore how social distance and organizational structure matter more than national boundary as well as indicating the importance of looking at the system-context, rather than isolated units. In terms of managerial implications, this paper calls on decision makers to carefully consider how a hierarchical distribution impedes collaboration. As sites develop competencies, there will be a need for them to move up an organization's value chain and possess greater decision-making power. At the very least, organizations need to be aware of how these dynamics operate as they attempt to achieve distributed collaboration and virtual workplace communities.

For a global organization to establish a 'flat,' integrated and collaborative workforce, they need to establish a collective orientation based on reciprocal relationships developed through joint work. While standardized processes can provide the parameters for this work, 'human behavior (is) not based on plans or on cultural universals but on the situatedness that characterizes human acting' (Spinelli and Taylor, 2003: 1; also Suchman, 1987; Salamon, 1993). This emergent ad hoc adherence to organizational protocols facilitates mutual reciprocity, establishes trust, and promotes the development of positive global relations. For organizations to adopt this approach, it would mean giving up the sense of control that comes from standardization. At the same time, organizations stand to gain from developing a workforce that is flexible, agile, and committed. Ultimately, it is up to organizations to determine whether the risk is worth the reward.

This paper is limited in that it only examines how WS analysis can be applied to one organization. More work needs to be done with a broader representation of organizations to see its ultimate utility. However, we can say that the model has resonated with organizational managers when presented during corporate trainings on how to encourage more reciprocal collaboration across distributed teams. This provides some anecdotal evidence that, while organizations are proclaiming to be developing integrated sites, the actual enactment of this strategy can interfere with this goal. Ideally, a longitudinal analysis could be done of an organization or project team that implemented changes based on the model. This would provide a better indication of the how a change in policy can bring about the formation of integrated distribution in collaborative teams.

References

1. Atkinson, J.M and Heritage, J (eds.) (1984). Structures of Social Action: Studies in Conversation Analysis, Cambridge: Cambridge University Press.
2. Avison, D.E. and Myers, M.D. (1995). Information Systems and Anthropology: An anthropological perspective on IT and organizational culture, Information Technology and People 8(3): 43–56. | Article |
3. Bamford, R.C. and Deibler II, W.J. (1993). Comparing, Contrasting ISO 9001 and the SEI Capability Maturity Model, IEEE Computer 26(10): 68–70.
4. Bannon, L. (1993). CSCW: An initial exploration, Scandinavian Journal of Information Systems 5: 3–24.
5. Bauer, M. (1996). The Narrative Interview: Comments on a Technique for Qualitative Data Collection, London: School of Economics and Political Science.
6. Chase-Dunn, C. and Hall, T.D. (1993). Comparing World-Systems: Concepts and working hypotheses, Social Forces 71(4): 851–866. | Article |
7. Damian, D., Maurer, F. and Sridhar, N. (2002). International Workshop on Global Software Development, in Proceedings of the International Conference on Software Engineering (Orlando, Florida, 2002).
8. Dutta, S., Van Wassenhove, L.N. and Kulandaiswamy, S. (1998). Benchmarking European Software Management Practices, Communications of the ACM 41(6): 77–86. | Article |
9. Evans, P. (1979). Dependent Development: The Alliance of Multinational, State, and Local Capital in Brazil, Princeton, NJ: Princeton Press.
10. Ford, D.P., Connelly, C.E. and Meister, D.B. (2003). Information Systems Research and Hostede's Culture's Consequences: An uneasy and incomplete partnership, IEEE Transaction on Engineering Management 50(1): 8–25. | Article |
11. Galliers, R.D. (2003). Information Systems in Global Organizations: Unpacking culture, in S. Krishna and S. Madon (eds.) The Digital Challenge: Information Technology in the Development Context, Aldershot: Ashgate.
12. Garfinkel, H. (1967). Studies in Ethnomethodology, Englewood Cliffs, NJ: Prentice-Hall.
13. Geertz, C. (1973). The Interpretation of Culture, London: Hutchinson.
14. Goodwin, C. (1980). Restarts, Pauses and the Achievement of a State of Mutual Gaze at Turn-beginning, Sociological Inquiry 50: 272–301. | Article |
15. Hammer, M. (2007). The Process Audit, Harvard Business Review (April): 111–123.
16. Harper, R. and Hughes, J.A. (1993). What a F-ing System! Send em all to the Same Place and then Expect us to Stop em Hitting: Making technology work in air traffic control, in G. Button (ed.) Technology and Working Order: Studies of Work, Interaction and Technology, London: Routledge, pp. 127–144.
17. Heath, C. (1984). Talk and Recipiency: Sequential organization in speech and body movement, in J.M. Atkinson and J. Heritage (eds.) Structures of Social Action: Studies in Conversation Analysis, Cambridge, UK: Cambridge University Press.
18. Hersleb, J.D., Atkins, D.I., Boywe, D.G., Handel, M. and Finholt, T.A. (2002). Introducing Instant Messaging and Chat into the Workplace, in Proceedings of CHI (Minneapolis, MN, 2002).
19. Hofstede, G. (1981). Culture's Consequences: International Differences in Work Related Values, London: Sage Publications.
20. Hofstede, G. (1991). Cultures and Organizations: Software of the Mind, London: Sage Publications.
21. Huang, J., Newell, S., Galliers, R.D. and Pan, S.L. (2003). Dangerous Liaisons? Component-based Development and Organizational Subcultures, IEEE Transactions on Engineering Management 50(1): 89–99. | Article | ISI |
22. Janowitz, M. (1977). A Sociological Perspective of Wallerstein, The American Journal of Sociology 82(5): 1090–1097. | Article |
23. Kankanhalli, A., Tan, B. and Wei, K.-K. (2006–7). Conflict and Performance in Global Virtual Teams, Journal of Management Information Systems 23(3): 237–274.
24. Klein, H.K. and Myers, M.D. (1999). A Set of Principles for Conducting and Evaluating Interpretive Field Studies in Information Systems, MIS Quarterly 23(1): 67–92. | Article | ISI |
25. Knorr-Cetina, K. (2000). Epistemic Cultures: How the Sciences Make Knowledge, Cambridge, MA: Harvard University Press.
26. Koskinen, I. (2000). Workplace Studies: An ethnomethodological approach to CSCW, Nordic Interactive Meeting, 31 March, http://www.ncrc.fi.
27. Krishna, S., Sahay, S. and Walsham, G. (2004). Managing Cross-Cultural Issues in Global Software Outsourcing, Communications of the ACM 47(4): 62–66. | Article |
28. Liberman, K. (1995). The Natural History of Some Intercultural Communication, Research on Language and Social Interaction 28(2): 117–146. | Article |
29. Lipnack, J. and Stamps, J. (2000). People Working Across Boundaries with Technology, 2nd edn, Chichester: John Wiley and Sons.
30. McCormick, T. (1990). World Systems, The Journal of American History 77(1): 125–132. | Article |
31. Meier, C. (2003). Doing 'groupness' in a Spatially Distributed Work Group: The case of videoconferences at Technics, in L.R. Frey (ed.) Group Communications in Context: Studies of Bona fide Groups, Mahwah, NJ: Lawrence Erlbaum, pp. 367–397.
32. Mihhailava, G. (2007). Virtual Teams: Just a theoretical concept or a widely used practice? The Business Review, Cambridge 7(1): 186–193.
33. Miller, L. (1995). Two Aspects of Japanese and American Co-worker Interaction: Giving instructions and creating rapport, The Journal of Applied Behavioral Science 31(2): 141–161. | Article |
34. Osland, J.S., Bird, A., Delano, J. and Jacob, M. (2000). Beyond Sophisticated Stereotyping: Cultural sensemaking in context, The Academy of Management Executive 14(1): 65–79.
35. Rawls, A.W. (2005). Epistemology and Practice, Cambridge, UK: Cambridge University Press.
36. Robillard, P. (1999). The Role of Knowledge in Software Development, Communications of the ACM 42(10): 87–92. | Article |
37. Salamon, G (ed.) (1993). Distributed Cognition: Psychological and Educational Considerations, Cambridge, UK: Cambridge University Press.
38. Seely Brown, J. and Duguid, P. (2002). The Social Life of Information, Cambridge, MA: Harvard Business School Press.
39. Skocpol, T. (1977). Wallerstein's World Capitalist System, American Journal of Sociology 82(5): 1075–1090. | Article |
40. Søndergaard, M. (1994). Hofstede's Consequences: A study of reviews, citations and replications, Organization Studies 15(3): 447–456. | Article |
41. Spinelli, G. and Taylor, S. (2003). Ethnomethodological Reflections on Collaborative Work Practices in Simulation Modeling: A short journey to elsewhere, Proceedings 15th European Simulation Symposium, http://www.scs-europe.net/services/
    ess2003/PDF/TOOLS16.pdf.
42. Sproull, L. and Kiesler, S. (1986). Reducing Social Context Cues: Electronic mail in organizational communication, Management Science 32(11): 1492–1512.
43. Suchman, L.A. (1987). Plans and Situated Actions: The Problem of Human-Machine Communication, Cambridge, UK: Cambridge University Press.
44. Thompson, J.D. (1967). Organizations in Action, New York: McGraw-Hill.
45. Tiwana, A. (2004). Beyond the Black Box: Knowledge overlaps in software outsourcing, IEEE Software (September/October): 51–58.
46. van Maanen, J. and Barley, S.R. (1984). Occupational Communities: Culture and control in organizations, Research in Organizational Behavior 6: 287–365.
47. von Krogh, G. (1998). Care in Knowledge Creation, California Management Review 40(3): 133–153. | ISI |
48. Wallerstein, I. (1974). The Modern World System, New York: Academic Press.
49. Wallerstein, I. (2005). World-System Analysis: An Introduction, Durham, NC: Duke University Press.
50. Walsham, G. (1993). Interpreting Information Systems in Organization, Chichester: Wiley.
51. Westrup, C., Al Jaghoub, S., El Sayed, H. and Liu, W. (2003). Taking Culture Seriously: ICTs cultures and development, in S. Krishna and S. Madon (eds.) The Digital Challenge: Information Technology in the Development Context, Burlington, VT: Ashgate Publishing Limited, pp. 13–27.
52. Whyte, W.F. (ed.) (1990). Participatory Action Research, vol. 123, Newbury Park, CA: Sage Publications.

About the authors

Dr. Gary David is an associate professor of Sociology at Bentley College. Dr. David's research examines the role of intergroup relations and intercultural communication in the formation of social relationships among distributed teams through a workplace studies perspective. Current projects include a study of the medical transcription industry, cross-boundary information sharing between government agencies, and the development of virtual workplace communities in distributed organizations. Dr. David has developed training programs on the use of 'everyday ethnography' as a device to decreasing social distance and increasing rapport in the pursuit of collaborative work. He has published numerous articles and chapters internationally across a range of disciplines.

Dr. Donald Chand is a Professor of Information and Process Management at Bentley College. His current teaching and research interests are in the areas of design methodologies, performance metrics and balanced scorecard implementation, software process improvement, offshore software development, distance project management and ERP systems. Dr. Chand has published articles in such flagship journals as the Communications of the ACM, Journal of ACM, IEEE Software, and the Journal of Management Systems.

Dr. Sue Newell is Cammarata Professor of Management at Bentley College, and part-time Professor of Information Management at Warwick University (UK). Her research focuses on understanding the relationships between innovation, knowledge and organizational networking. She was one of the founding members of 'ikon,' a research center based at Warwick University. Her research emphasizes a critical, practice-based understanding of the social aspects of innovation, change, knowledge management and inter-firm networked relations. Professor Newell has published over 70 journal articles in the areas of organization studies, management, and information systems, as well as numerous books and book chapters.

Dr. João Resende-Santos is an associate professor of International Studies at Bentley College. His areas of specialization include international relations, international political economy, and Latin American studies. Research and publications span three areas: international relations theory and security studies; military organizational behavior and civil–military relations; and Brazilian foreign relations. Most recent publications include: 'Neorealism, States, and the Modern Mass Army' (2007: Cambridge University Press), and 'Origins of Security Cooperation in South America,' in Journal Latin American Politics and Society (W'2002). He is also involved in business and technology incubator development in Cape Verde and Brazil.