Paradox and the negotiation of tensions in globally distributed work

Tensions are a major source of communication problems, coordination issues, and conflict in globally distributed work. In this article, we argue that extant literature falls short of addressing tensions in globally distributed work at two levels. First, it fails to fully account for the intrinsic and entrenched nature of tensions in globally distributed work, suggesting instead that they can be resolved or made to disappear. Second, it does not examine the key interactions among different kinds of tensions. Drawing on qualitative data from a distributed finance organization and applying concepts from paradox theory, we show how globally distributed units negotiate knowledge, power, and identity tensions in collaborative work. The findings illuminate how a sequential enactment of both formal and informal solutions can better address tensions and generate collaborative opportunities in globally distributed work. Building on the findings, we develop a phasal model of tension evolution and management in globally distributed work which explains how tensions evolve from a phase of suppression through to a phase of attenuation. We demonstrate the interactions of knowledge–power–identity tensions against a background of defensive, interactive, and collaborative behaviors, and suggest several practical implications for globally distributed work practice.


Introduction
Organizations rely on globally distributed work (GDW) to take advantage of complementary objectives. In typical arrangements, highly skilled onshore teams focus on strategy and customer interactions, while offshore teams focus on repetitive back-office activities, reducing costs and improving efficiencies. Individually, each side can operate within the scope of their own tasks, priorities and values. However, when they are required to collaborate on distributed processes, tensions stemming from inherent contradictions in the working relationship cause strain and discomfort to both sides (Barney et al., 2014).
The extant literature recognizes three main sources of tensions in GDW: knowledge asymmetries (Vlaar et al., 2008;Hahl et al., 2016); power asymmetries (Levina and Vaast, brought to you by CORE View metadata, citation and similar papers at core.ac.uk provided by Loughborough University Institutional Repository Drawing on an in-depth qualitative case study of a distributed finance organization, we apply ideas from paradox theory (Lewis, 2000;Smith and Lewis, 2011) and offer an alternative approach to understanding tensions in GDW. The notion of paradox suggests that tensions are not permanently resolved in the traditional sense because leaning towards one polarity (such as cost reduction and efficiency) puts increasing pressure on the other polarity (such as quality and innovation) (Lewis, 2000;Smith et al., 2017). When applied to GDW settings, the paradox perspective helps us see that the successful implementation of collaborative GDW projects require managerial action and solutions different from the formal ones typically highlighted in the literature. Our findings suggest that defensive behaviors and attempts to force one side's preferred position may continue even after the implementation of formal solutions aimed at addressing tensions. While such behaviors strain interactions and result in a period of conflict, the findings show how subsequent informal managerial action can be central to improved mutual understanding and new collaborative ways of working.
Our study makes three contributions to research. First, we extend current debate in GDW, which mainly emphasizes the importance of formal solutions in managing tensions. Our findings point to the need for more sophisticated and nuanced solutions in the management of persistent and inter-related GDW tensions. Through the paradox lens, we demonstrate how a sequential enactment of both formal and informal solutions, rather than simple formal solutions (e.g. structural separation of the paradoxical units), can better address tensions and generate collaborative opportunities. Second, we develop a phasal model which explains how tensions evolve from a phase of suppression through to a phase of attenuation. It further describes and details the characteristics of specific behaviors associated with each phase. Third, we extend prior research by empirically illustrating the interactions of knowledge-power-identity tensions in GDW. In particular, we elucidate how these interactions unfold against a background of defensive and interactive-collaborative behaviors. These interactions suggest that structural changes in GDW can immediately amplify power and identity tensions. They also show how efforts to manage some tensions can impact a different type of tension in collaborative GDW.
In the following sections, we review the literature on GDW with a specific focus on tensions and introduce paradox as an appropriate theoretical perspective. Next, we provide a detailed account of the methodology and case analysis. Finally, we discuss the paper's main theoretical and practical implications.

Tensions in GDW
Globally distributed work (GDW) involves the demarcation and division of work into dispersed task specific work units in order to enable innovative ideas (Levina and Vaast, 2008;Eisenberg and Mattarelli, 2016) as well as cost savings (Dibbern et al., 2008). The success of GDW mainly depends on the ability and willingness of onshore and offshore units to collaborate on distributed processes. It follows that variations in such ability and willingness create tensions between the two sides. The extant GDW literature finds there are three prominent sources of tensions at play: knowledge asymmetries, power asymmetries, and identity threats (Table 1).
Knowledge asymmetries refer to the imbalance of knowledge and experience (Oshri et al., 2008;Vlaar et al., 2008) in performing specialized tasks (Tiwana and Keil, 2007;Leonardi and Bailey, 2013). Offshore units have been criticized for possessing only a limited knowledge of the onshore business and therefore contributing to poor organizational performance (Vlaar et al., 2008;Barney et al., 2014). For instance, Barney et al. (2014) found offshore Indian software engineers lacking in domain knowledge, leading to misunderstanding of requirements. On the other hand, onshore units have been criticized for information hiding (Connelly et al., 2012;Zimmerman and Ravishankar, 2014) and for not understanding offshore processes (Tiwana and Keil, 2007;Cha et al., 2008). Unmanaged knowledge asymmetries can hinder effective communication, reduce understanding (Levina and Vaast, 2008), limit positive knowledge spillover and reduce opportunities for joint problem solving (Jarvenpaa and Majchrzak, 2016). Firms invest in formal structures and systems to reduce knowledge tensions.
Rules, standard operating procedures, codified directories and common lexicon are used to overcome gaps in knowledge (Oshri et al., 2008) and carefully designed structural arrangements help restrict the scope of collaborations, limit the need for knowledge sharing, and mitigate the opportunity for conflict (Oxley and Sampson, 2004).
Power asymmetries in GDW stem from imbalanced access to resources and decision makers, allowing some parts of the organization to enjoy a higher status than others (Levina and Vaast, 2008;Ravishankar et al., 2013;Hahl et al., 2016). In general, research has found that power resides in onshore units that control structural resources (Levina and Kane, 2009;Ravishankar et al., 2013). Levina and Kane (2009) found that onshore teams are well connected to senior decision makers and have better access to budgets, training and technology. Onshore units also assume a higher status on account of their perceived advanced language skills and qualifications (Hinds et al., 2014), engagement with higher-value tasks and superior domain expertise (Levina and Vaast, 2008). More recent research suggests that in knowledge intensive collaborations, it is difficult to determine 'who owns the code', decipher where decision-making power should reside and evaluate how rewards should be distributed (Chen et al., 2017;Parker et al., 2017). Organizations draft unambiguous contracts (Dibbern et al., 2008;Parker et al., 2017), legalize ownership of intellectual property and formalize the distribution of power via clearer documentation of responsibilities (Parker et al., 2017) to mitigate and manage the harmful impact of power asymmetries. Studies also indicate that service level agreements (SLAs) help formalize expectations and reduce power-related task conflicts (Goo et al., 2009;Deloitte, 2016).  David et al., 2008;Zimmerman and Ravishankar, 2011;Eisenberg and Identity threats in the GDW context typically involve experiences that potentially violate an individual's or group's sense of professional self (Petriglieri, 2011;Koppman et al., 2016).
Onshore identities are prone to threats as service offshoring matures, leading to migration of complex tasks (Zimmerman and Ravishankar, 2014; and increasing need for crosscultural interactions to accomplish the said tasks (Petriglieri, 2011). On the other side, research has shown how offshore units' sense of identity and organizational membership are threatened when they are treated as subordinates by their onshore counterparts (Leonardi and Rodriguez-Lluesma, 2013;Koppman et al., 2016). In a study of German-Indian teams, Zimmerman and Ravishankar (2011) showed how offshore Indian engineers who considered themselves to be skilled professionals experienced significant identity threats when their onshore German colleagues assigned them repetitive and less interesting tasks.
Identity threats can give rise to a plethora of undesirable effects, including poor performance (Ravishankar, 2015), resistance to change (Zimmerman andRavishankar, 2011), stigmatization (Major andO'Brien, 2005) and reluctance to share knowledge (Zimmerman and Ravishankar, 2011;Eisenberg and Mattarelli, 2016). Organizations have sought to resolve identity threats using cultural training (Newell et al., 2007;David et al., 2008) and intermediaries (or 'inpatriates'), who transcend cultural differences between subgroups and work to create a strong sense of identification with the larger organization (Eisenberg and Mattarelli, 2016). A small number of studies also show how more informal approaches such as framing narratives can be used strategically to resolve identity threats and temporarily reduce conflict in GDW (Petriglieri, 2011;Ravishankar, 2015).

A paradox perspective
Often described as 'two sides of the same coin', paradox is a condition where two individual elements are interrelated and necessary to each other's existence (Smith and Lewis, 2011), and yet seem to be opposing or contradictory (Poole and Van de Ven, 1989;Smith and Lewis, 2011). The tensions created by the pull of two paradoxical elements are not easily resolved since leaning towards one puts increasing pressure on the other (Lewis, 2000;Smith et al., 2017). In business and management research, the notion of paradox has been used to examine identity is being threatened.
Mattarelli, 2016; Koppman et al., 2016 contradictory elements such as competition and cooperation in organizational alliances (Rai, 2016), exploitation of existing resources and exploration of new ideas in innovative projects (Andriopoulos and Lewis, 2009), and novelty and usefulness in creative endeavors (Miron-Spektor and Beenen, 2015). Within the IS literature, studies have considered paradoxical contradictions between control and autonomy in self-managing teams (Druskat and Wheeler, 2003), control and empowerment in hybrid teams (Cousins et al., 2007), flexibility and structure in client-supplier innovation projects (Kotlarsky et al., 2016), and trust and mistrust in distributed teams (Zolin et al., 2004). Paradox brings into sharp focus the role of behaviors in heightening or relieving tensions created by the opposing elements. Paradox posits that defensive behaviors such as avoiding differences or forcing one-sided solutions may only provide temporary relief (Smith and Lewis, 2011). In fact, they can create vicious reinforcing cycles that perpetuate and intensify tensions (Lewis, 2000;Smith and Lewis, 2011). For example, individuals may defensively retreat to their own divisions and stop engaging with partners, worsening the tensions in collaborative-competitive relationships (Jarzabkowski et al. 2013); offshore teams may defend contradictory elements in offshoring as 'cultural differences', further amplifying tensions (Ravishankar, 2015).
The literature suggests two broad formal strategies for managing paradoxes (Poole and Van de Ven, 1989;Lewis, 2000). The first, separation-based strategy, involves demarcating and honoring conflicting elements. For instance, distinctive formal structures can create much needed physical and temporal separation between tension-elevating elements (Bengtsson and Kock, 2000;Mayasandra et al., 2006;Stadtler and Wassenhove, 2016). Bengtsson and Kock's (2000) study of 'coopetition' illustrates the creation of departments that focus on either competitive or cooperative strategies. Further, information technologies make it possible to minimize communication and task interdependence between separated teams whilst at the same time connecting them and supporting task integration (Dubé and Robey, 2009). Other studies suggest 'time-bound' separations -teams focusing on one strategy for a period before returning to attend to the other (Poole and Van de Ven, 1989). Although separation strategies may avoid conflict and instability, they sometimes ignore the potential of blending for turning disagreements into value-creation (Stadtler and Wassenhove, 2016;Johansen, 2018). Overall, however, separation-based strategies dominate managerial efforts aimed at managing paradoxes The second, integration-based strategy, is less common and involves purposefully juxtaposing contradictory elements (Lewis, 2000;Lewis and Smith, 2011), thus seemingly intensifying tensions. However, if managed effectively, such tensions can be turned into good collaborative outcomes. For instance, Dubé and Robey (2009) found how employing individuals with experience of handling ambiguous situations, maintaining shared calendars for transparency and establishing a common focus while acknowledging differences helped overcome structure-flexibility tensions. In well-managed integration scenarios, actors break out of vicious cycles and move away from defensive responses by acknowledging tensions and reframing them as necessary or accepting them as natural conditions of work. The process of recognizing and accepting inherent contradictions can provide actors with a sense of 'freedom' in "learning to live with" the opposing elements (Lewis, 2000: 764). They may embrace tensions, making small cognitive and behavioral adjustments to find benefits within the otherwise tense environment (Jarzabkowski et al., 2013;Jarzabkowski and Le, 2016).
Integration-based strategies can encourage critical self-reflection, stimulate more ambivalent interpretations of surroundings (Lewis, 2000;Smith et al., 2017) and avoid oversimplification of problems (Plambeck and Weber, 2010).
In summary, our synthesis of the literature suggests a paradoxical mindset (i.e., a readiness to embrace inherent contradictions) can help organizations find more accommodating solutions (Smith 2014;Bengtsson et al., 2016), improve organizational learning (Lewis 2000;Sundaramurthy and Lewis, 2003), innovate more (Andriopoulos and Lewis, 2009), have higher problem solving capabilities and adapt better to change (Schuman et al., 2010), have successful inter-firm alliances and achieve overall superior performance (Rai, 2016). In terms of current research, we know a lot more about 'why' organizations strive for a paradoxical mind-set than 'how' it is achieved (Stadtler and Wassenhove, 2016). In the specific context of GDW, there is limited understanding of how tensions can be turned into collaborative ways of working.
Further, there is scarce research on the implications of interactions amongst different tensions for collaborative outcomes. In the following sections, we present the case of a distributed finance function, where work has been demarcated and distributed between two sides -onshore business units and an offshore shared service center. Using this case, we explain how the two sides negotiate tensions and explicate the role of a sequential enactment of both formal (integration-based) and informal managerial solutions in generating collaborative opportunities.

Research approach
We wanted our research approach to allow us to explore the potentially complex challenges onshore-offshore unit members face as they collaborate. In line with past research, we adopted the qualitative case study method (Walsham, 2006) which generates rich data and helps advance understanding of under-researched phenomenon (Eisenhardt, 1989;Kotlarsky et al., 2014). We drew predominantly on interview data from both onshore and offshore units to gain an in-depth understanding of the structural, as well as the social context of collaborative GDW.
The findings report data that emerged organically from our interpretivist approach to data analysis.

Research setting
Our research is set in the finance function of a large global logistics firm made up of four business divisions. Our research focuses on the paradox emerging from the migration of financial tasks from several onshore country level business units (BUs) within the firm's largest business division to an offshore shared service unit (SSU). Our empirical material provides retrospective accounts, triangulated with several data sources, to capture tensions across significant changes in organizational structure ( Figure 1). Throughout the paper we refer to employees working in the SSU as 'SSU members' and employees working in BUs as 'BU members'.
Finance Organization Structure 1: Centralization of financial operations into GBS A centralization strategy, which began in 2008, aimed to consolidate resources from across the multiple country level business units (BUs). Tasks were transferred out of the BUs into a dedicated financial shared services unit (SSU), which sat within a wider corporate center called Global Business Services (GBS). In addition to finance, GBS provided a range of support services to the BUs including procurement, legal, HR and IT support. The idea of centralizing finance was to help BUs improve financial reporting, better manage financial risk and improve customer focused activities. For instance, it was anticipated that moving a range of activities into the SSU would give BU members more time to visit customers, better understand their financial and shipment needs, offer more tailored customer service and increase sales by generating larger shipping orders. The SSU focused on standardizing and delivering a range of financial operations to the BUs such as the purchasing and paying for goods and services. The BUs paid for the SSU's services, which were governed via SLAs and performance was measured through key performance indicators (KPIs). During this period, BU members became increasingly dissatisfied with the service provided by the SSU. BU members were focused on being more responsive to customer queries and offering flexibility through shipment discounts and differentiated payment plans. SSU members, on the other hand, were focused on reducing the cost of finance and standardizing processes to gain efficiencies. In 2011, to achieve further cost savings, some transactional activities related to the invoicing provision were outsourced to a third-party provider. BU members saw this outsourcing arrangement as being at odds with their priorities and this view led to several BU-SSU disagreements. SSU members were also tasked with the long-term improvement of financial processes through standardization, integration of existing information systems and the implementation of new IT tools across multiple BUs. However, the BUs did not always have the funds to support these SSU objectives. Projects deemed to not have an immediate impact on BUs were often rejected by BU members who were more focused on their annual sales targets.

Finance Organization Structure 2: SSU integration into Business Division
In late 2014 the SSU's reporting lines were changed and it was integrated into the business division. The SSU did not physically change location, and tasks remained centralized, but SSU members now reported to the business division rather than to GBS. As the SSU became integrated into the business division, BU and SSU members were expected to collaborate more closely to perform and improve distributed processes. Our data describes the tensions between the BUs (onshore units) and the SSU (offshore unit), given their conflicting objectives (see Table 2), and explains the process of accommodating contradictory needs and creating collaborative opportunities.

Data sources
Data was collected in three rounds (see Table 3)   − Meet internal and external audit requirements − Support regional and global finance organizations − Improve customer experience of financial processes − Improve the financial position of BU: projects to speed up processing of invoices − Support sales team to increase revenue: e.g., work out discounts for customers to encourage them to commit to larger shipments or move from other logistics providers the SSU head office. Interviewees included managers from the SSU and the firm's finance organization (FO) headquarters. The third round of interviews was conducted in June 2016 with BU participants from four different onshore countries. To understand tensions the SSU members face in their own work and in collaborative projects.
To understand how SSU members cope with tensions.

3: June 2016
BU members 9 To understand how BU members experience and cope with tensions in their own work and in collaborative projects.
The interviews lasted between 60 and 140 minutes, averaging 77 minutes and amounting to about 33 hours of interview data. All but three (teleconference) interviews were conducted face-to-face. We collected several other documents, including annual reports, internal communications, performance dashboards, employee opinion surveys and employee development initiatives. These documents helped understand the chronology of structural and organizational changes and offered additional insights into the behaviors of onshore-offshore members. We also spent time in the offices, spoke informally with several participants, took photographs of the working environment and produced field notes.

Data analysis
We took the opportunity to organize and analyze data at each round of data collection. First, we transcribed all interviews from round one, studied the data, and made extensive notes to identify emerging themes of each interview individually. Second, we combined the interview data with documentation and field notes. This helped us to map the organizational structure, reporting lines, workflow and responsibilities, objectives and performance metrics. Third, we set about identifying common themes across the complete round one dataset. To this end, we wrote short guiding narratives and constructed a table of themes. The presence of opposing objectives, the underlying power and knowledge asymmetries, and accounts of conflict between onshore (BUs) and offshore (SSU) units emerged organically from this round of interview data. We repeated this process for the second and third rounds of data collection.
Data from the second round revealed in greater depth the conflicting objectives between the SSU and BUs and the apparent threats to participants' sense of professional self. FO informants discussed the application of formal tools and ICT to resolve issues, and the process of integrating the SSU into the business division. Data from the third round, consisting of BU interviews, also referred to different work, objectives, and highlighted the differences in the perceived value of work performed onshore -in the BUs, and offshore -in the SSU. At the completion of three rounds of interviews, we made extensive notes and created a thematic table of onshore-offshore conflicts and tensions (e.g., Corley and Gioia, 2004) and their ongoing management by the FO headquarters. We then went back to the literature and iterated between theoretical explanations and the data until we were confident that our interpretation reached a reasonable level of theory-data alignment.
Case Analysis Table 4 juxtaposes the perceptions of members from four onshore BUs with those from the offshore SSU to illuminate the specific paradoxical tensions in the relationship. In the analysis below, we show how members moved from defensive behaviors, which intensified tensions, to collaborative behaviors which attenuated them. We follow the analysis with a detailed discussion section.

Tensions and defensive behaviors
As summarized in Table 4, contrasting objectives, differing levels of knowledge and access to key decision makers, and issues of identity created tensions between the SSU and the BUs.
SSU members noted that they struggled to coordinate and control information because of the BUs' reluctance to share business knowledge. This restricted access meant that the SSU was not able to fully understand the BUs' requirements. SSU members also explained that they did know enough about the activities performed by the third-party provider, which further limited their ability to support the BUs. When explaining their weak knowledge position, SSU members noted that they were seen as a 'cost center' (i.e., the SSU did not directly generate revenue or make a profit and was instead perceived to cost the firm money). They reported that their position within the GBS, which was widely seen as a support service, had created a kind of status asymmetry between BU and SSU members. SSU members said that because they were support services, they were perceived as less capable, of less value, and as having different objectives to their BU counterparts. SSU members suggested they received limited financial investment and were unable to allocate resources needed to fully understand the activities performed in the BU, or to manage the third-party provider. In addition, despite being tasked with 'process improvement', they did not have the authority to implement new IT tools. Both BU and SSU members acknowledged the tensions as they transitioned towards a more centralized structure. Several members referred to the disparity between BU and SSU decision-making power. While the SSU was given the responsibility to centralize and improve financial processes, senior managers in the SSU suggested that they did not have the "stick" or "the backing" from the finance organization (FO) headquarters to force BU members to accept process changes and new IT tools. They highlighted problems of knowledge and control, referring to the relationship as "artificial" (P2.1), with "virtual" (P1.2) and "psychological barriers" (P2.3; P2.7). They also felt they were treated like "suppliers" (P2.4) despite being internal to the firm. A senior SSU manager recalled a humiliating experience where the BU had dismissed plans to develop a financial process and suggested that they "needed to have somebody from the business division looking into it" (P2.4).
From the BUs' perspective, they were legally accountable to external auditors for financial documents produced by SSU members. They were also accountable to the business division for the quality of financial reporting, and to their customers for service quality. BU members felt this accountability was at odds with handing over control of related activities to the SSU. They did not have enough understanding of the SSU's activities to be able to generate the necessary reports or rectify mistakes. BU members also felt that the SSU's emphasis on IT standardization was in sharp contrast to the BUs' focus on flexibility. For instance, a BU member explained that their team's daily plan, which they referred to as a "living thing", was devised on a whiteboard and adapted throughout the day. Others noted that they frequently had to respond to urgent calls from customers and requests for information from senior managers.
BU members saw the SSU as a "middleman" between themselves and the third-party provider who processed invoices. They felt this arrangement made accessing information and responding to customers more complex and time consuming. In addition, they questioned the SSU's ability to control and manage information and to improve financial processes. They highlighted the strategic importance of their own work and saw the SSU as the "poorer cousin…who got the bad stuff" (P3.5). On the other hand, SSU members felt unhappy at being perceived as merely a "cost saving entity" (P2.4) created to perform "transactional", "repetitive" and "back-office" work. They argued that such a view was not a true reflection of their ability to perform and improve processes. Many of them held master's degrees and formal accounting qualifications and argued that their low status in the relationship was unjustified. Defensive behaviors SSU members claimed that the BUs actively resisted their (SSU's) efforts to standardize IT processes. BU members explained their resistance in terms of the rigidity the SSU initiatives could impose on the BUs' everyday work and on the BUs' freedom to make local adaptations. SSU members explained that BUs invoked the contractual nature of their relationship with the SSU to reject initiatives which aimed at further standardizing financial processes. These projects tended to be rejected because of the perceived cost. Some SSU members explained that BU resistance was often supported by the business division, which did not enforce centralized process changes.
In the past if we challenged a process, we had nobody we could go to. If the BU didn't want to change, then they didn't change. The business division did not support our projects. We were there to deliver a service and that service could only be what the local BU 'needed'. We've never had that backing. (P2.1) To access information quickly, BUs often avoided standardized formal processes and contacted the third-party provider directly. Although this meant that issues were resolved locally, deeper and more fundamental issues with the process were not flagged or fixed.
I jump the gun and go to the [provider] to get a quick answer so I can quickly feedback to my manager. When it's to do with you, you're willing to jump any hoops to quickly resolve your issue. (P3.5) From the SSU's perspective, these BU behaviors restricted their ability to perform their own roles effectively. One SSU member explained: We were under continuous pressure to drive costs down. At the same time BUs were not allowing us to change and to optimise the process which would allow us to decrease costs! It's an interesting 'spagaat' [translation: splits] -cost reductions on one side and on the other side not being allowed to change the process. (P1.9) The SSU responded by trying to limit communication exchanges between the BUs and the thirdparty provider. For instance, SSU members improved formal documentation to reduce the need for direct communication. They kept 'log books' to record all mistakes, process delays and complaints from BU members. SSU members also tried to force BUs to use formal protocol and communication channels by boycotting deviations and not completing work that did not align with the standard format.

Someone will phone up or send an email saying 'can you add this' and we say 'no you'
re not going to do that. You are going to send the form'. Often, we have implemented an IT tool they need and we make it stricter -and god that's something we want. We want to improve the controls. (P1.5) The SSU also focused on the formal terms of their contracts with BUs in order to emphasize the obligations of all parties. SSU members invoked SLAs and statement of work documents, to reiterate the specific roles and objectives of the BUs, the third-party provider and the SSU.
However, both BU and SSU members referred to SLAs (and the focus on contractual obligations) as one of the biggest collaborative challenges. BU members pointed out that SLAs codified the misaligned objectives between themselves and the SSU -cost reduction for the SSU and service quality for the BUs. They noted that the SSU, BUs and the third-party provider were rewarded through financial bonuses for meeting conflicting objectives. The SSU's bonuses were linked to cost savings and efficiencies, the BUs' to improving financial outcomes for the unit, and the thirdparty provider's bonuses were based on the number of invoices processed. They used the terms set out in the SLAs to defend their own objectives and bonuses. BUs showed their discontent with the SSU's inflexibility by issuing 'red SLAs' to signal that obligations had not been met. SSU members countered with 'blue SLAs' to justify their own actions, noting that the BUs had not provided the information needed for the SSU to fulfill its obligations. BU members felt that the SSU had developed a "fighter's mentality" (P2.6) and that the two sides became stuck in blaming contests and "stalemate" conversations about SLAs. At the same time, some BU members admitted that they found it challenging to hand over control of tasks they had earlier performed themselves.

Negotiating tensions
In order to better align SSU's service provision with the needs of the BUs, the finance organization (FO) headquarters altered the organizational structure ( Figure 1: Finance Organization Structure 2). The SSU was structurally removed from GBS and integrated into the business division. Beyond the structural change, there was a conscious effort to break "psychological barriers" and integrate the SSU into the division. Senior managers overseeing the transition made it clear that they did not want the structural change to be seen as a "hostile takeover" of the SSU by the division, but only as a means for the BUs and the SSU to "come together". They actively emphasized "togetherness" in their communications and referred to "putting pieces of the puzzle back together" and "working as one" (P2.6; P2.1; P2.5). The business division held a "welcome" event as a starting point for their new unity. One BU member said: We held a small celebration with them (the SSU). You feel that people are more engaged to deliver the best they can, not just because we gave them a glass of prosecco and a couple of candies, but because we tell them: "Look, what you guys are doing is really important". We show them the value in what they're doing, and they see their full contribution to the network. Through the network you're part of the family. (P.3.9)

Interactive behaviors
The structural integration resulted in BU and SSU members being involved in more joint management meetings, joint service performance reviews (SPRs), and collaborative improvement projects. BU and SSU members reported that working together in this way was often more difficult than working independently. They found these meetings involved "uncomfortable" and "frank discussions". BUs also introduced a service 'health check' which rated service provision on a traffic light system (green -good, red -bad, and amber -requires improvement However, BU and SSU members noted that while meetings led to uncomfortable discussions and blame passing, they also provided opportunities to better understand their counterpart's needs, confront them on issues, and to defend their own positions. As an SSU member explained: Collaborative behaviors SSU members started to take a more collaborative view of their responsibilities and reflect on how their own actions impacted the larger organization. They spoke about picking their battles carefully and explained that they had become better aligned with the BUs' need for flexibility. As they increased their interactions, the focus on imposing rigid controls diminished significantly. One SSU member gave an example of a new online invoicing tool designed to require less manual input from the SSU when collecting payments. The idea was to speed up invoice processing times so that BUs received their customer payments quicker to improve cash flow. The implementation of the tool required the BUs to align their own systems and encourage their customers to adopt the online method of payment. The SSU's business plan aimed for 80% of invoices to be processed BU members also engaged in more collaborative behaviors. They referred to an internal "change in attitude" and began to encourage their SSU counterparts to make their own decisions and propose "improvement projects".
In the past they were treated as subordinates. I think they just kept their heads down because they were so scared of getting things wrong. Several BU and SSU members explained that while the integration process had been challenging, it had led them to engage in more collaborative decision-making. The "psychological divide" between the SSU and the BUs had reduced, and the decision-making process had become faster.
The SSU stopped complaining about the BUs' systematic resistance to change, and the BUs 24 stopped opposing SSU initiatives. The confrontational approach was replaced by collaborative discussions and negotiations. The SSU continued to be responsible for cost reduction and process standardization. The BUs continued to be "customer-centric" in their financial decisions but worked collaboratively with the SSU. The BUs acknowledged that the SSU's work on cost reductions and process efficiencies was enabling them (the BUs) to focus squarely on improving their customers' experience.
In the past a lot of the project change capacity was all focused on driving down cost. Now that capability is there for the BUs, but the CFO can easily say "Hey guys, let's park the cost project or the standardization in Asia for a little bit, and focus on this commercial one that has a revenue and customer satisfaction impact". This is now possible! In the past there would be reiterations and reiterations on the SLA impact and the budget. All that stuff has gone. (P2.6) While formal objectives did not change, the BU and SSU developed new ways of coping with tensions and working with one-another to find collaborative solutions. In accommodating one another, both BU and SSU members made practical trade-offs. For example, the SSU decided to reduce focus on cost reduction projects and increase focus on customer experience projects.
Similarly, the BUs volunteered to improve compliance with formal SSU processes. Both sides realized that not sharing knowledge was detrimental to achieving smooth distributed processes.
Rather than prioritizing their own objectives, they worked on understanding one-another's tasks as important components of a bigger picture. BUs empowered the SSU to make decisions in the larger interest of the organization even when such decisions made their own (BUs) work temporarily more challenging. SSU members engaged with BU members actively in the decisionmaking process. Overall, both sides explained that process performance had improved in terms of both efficiency (cost and speed) and quality (accuracy and customer service).

Discussion
The analysis above illustrates several inherent paradoxes in GDW. In particular, it shows how onshore-offshore collaborations involve opposing values and objectives and intense competition for resources and decision-making power. This analysis builds on extant literature which has identified three prominent sources of tensions in GDW -knowledge asymmetries (Oshri et al., 2008;Hahl et al., 2016), power asymmetries (Levina and Vaast, 2008;Ravishankar et al., 2013) and identity threats (Petriglieri, 2011;Koppman et al., 2016). In our case, the three tensions were most evident in the relationship between onshore BU units and the offshore SSU unit. While previous studies have acknowledged tensions and the conflict they create as a pertinent issue, thus far there has been very little understanding of how tensions can be turned into opportunities for effective collaborative practices. Further, only scant research has considered the interactions amongst tensions and how they play out in GDW. Drawing on our analysis and guided by ideas from paradox theory (Lewis, 2000;Smith and Lewis, 2011), we offer three contributions to the literature on GDW.
− First, we illuminate how a sequential enactment of formal and informal solutions may help address tensions and generate more effective onshore-offshore collaborative practices in GDW.
In the literature, tensions and their manifestations have been viewed as being either valuedestroying or value-creating. Symptoms of value-destroying tensions in GDW include a lack of Relinquished control of some processes

Attenuated
Tensions were less salient as both sides learned to live with them. They worked to align goals and bridge misunderstanding.
ability and willingness to share knowledge (Zimmerman and Ravishankar, 2014), poor integration across distributed process (Levina and Vaast, 2008) and growth of 'in' and 'out' groups (Hinds et al., 2014). Symptoms of value-creating tensions include better integrated activities and acceptance and accommodation of ambivalent situations and contradictions (Lewis, 2000;Sundaramurthy and Lewis, 2003;Johansen, 2018). Our findings go beyond viewing tensions in binary terms. They help explain how instability-generating tensions (i.e., value-destruction) in GDW can turn into an energizing catalyst and create opportunities for creativity and innovation (i.e., value-creation). As we elaborate below, in our case formal structural integration that seemingly intensified paradoxical tensions also led the SSU and the BUs to confront paradoxes directly and move towards more effective collaborative practices.
The SSU had adopted defensive behaviors such as 'working to rule', hiding information, and excessively emphasizing the need to follow processes. On the other side, the BUs leveraged their superior power position to resist change, access information directly from the third-party provider and force flexibility. These behaviors widened the relational gulfs, increased knowledge asymmetries and further reinforced the power imbalances and conflicting objectives. As prior research has shown, such defensive behaviors are frequently used to mute or disguise tensions and avoid 'paralyzing' conflict (Lewis, 2000;Ravishankar et al., 2013). They also ignore the practical effort required to make collaborations work, and can therefore lead to diminished opportunities for better understanding the GDW partner (Levina and Vaast, 2008). Interestingly, the structural separation of the SSU and the BUs -with the SSUs reporting to the GBS and the BUs reporting to the business divisions (see Figure 1) -and the systems used to make this uncoupling work, further exacerbated the paradoxical tensions. As our analysis shows, the two sides were unwilling to cooperate outside of formal obligations and share knowledge with one another and were prone to fiercely protecting their own positions. We would therefore suggest that simple formal solutions (e.g., structural separation of units) may not resolve the paradoxical tensions inherent in GDW.
They can be too formulaic to tackle the inherent paradoxes and additionally, may fail to recognize opportunities to generate new collaborative practices from the paradoxical tensions. Equally, as the literature points out, standalone informal coping strategies such as arriving at blatantly stereotypical 'cultural' understandings of the GDW partner's actions may not resolve tensions either. They may in fact hinder collaboration and create bigger relational challenges in the long run (Ravishankar, 2015).
However, our case suggests that when actors are given the chance to confront paradoxical tensions directly through a careful sequential enactment of both formal and informal solutions, it is conceivable that this will help address tensions and develop more effective collaborative practices. As our analysis shows, structural integration of the SSU and BUs -a formal solutioncreated proximity and forced both sides to share time and space, triggering a series of informal positive interactive behaviors such as frank discussions, empathy for offshore (onshore) counterparts, and demonstrations of better awareness of the paradoxical roots of their collaborative activities. In other words, both sides identified and negotiated relational challenges, and sought to address knowledge and power imbalances and align objectives. While these interactions were often described as uncomfortable, they energized both sides to develop new collaborative ways of working. For instance, our informants explained how 'new' joint meetings and frank discussions improved the speed of decision-making, led to greater knowledge sharing and empowered them to generate new IT tools for 'process improvement'. The key turning point for the emergence of collaborative behaviors was the actions taken by senior managers (i.e. BU CFOs) in overseeing the structural integration of the BUs and SSU 2 . The integration retained the physical separation of the SSU and BUs but eliminated the structural separation. BU CFOs paid close attention to reducing the power and identity tensions, explaining to the SSU members that they were highly valued, and that the integration was an opportunity to work with the BUs as same-status partners.
Thus, they provided both sides (esp. the SSU) a sense of psychological security to approach the integration with positive hope, rather than with worry and fear. This line of thought complements arguments in paradox theory which suggest that squarely facing up to opposing elements, albeit informally, within a safe and formally constructed space, helps mitigate tensions through better awareness and appreciation of mutual positions (Smith and Lewis, 2011;Smith, 2014). In this sense, confronting instead of avoiding may help turn potentially value-destroying tensions into opportunities for value creation (Johansen, 2018). There is a key insight here for the management of distributed teams, which often relies on distance to temper relationships and maintain stability: With an optimal combination of formal and informal solutions, tensions can be managed to improve relationships and encourage collaborate practices in GDW.

28
− Second, based on our analysis we have developed a phasal model which helps explain the evolution and management of tensions in GDW.
Tensions evolve with onshore-offshore exchanges. Specifically, our empirical data shows how tensions are suppressed, amplified, accommodated and attenuated in globally distributed collaborations (see Figure 2). Tensions evoke defensive behaviors and evolve in response to the behaviors. For instance, SSU's enforcement of formal communication channels compelled BUs to find alternative flexible ways to access information, thus further amplifying tensions. The structural integration of the SSU into the business division meant that both sides had to squarely face the tensions and negotiate differences. The SSU and BUs moved from defensive behaviors, which suppressed and then amplified tensions, to more positive interactive behaviors through which they learned (at least temporarily) to 'live with' and accommodate tensions (Smith and Lewis, 2011). As explained earlier, this process of coming together enabled both sides to better understand the other's perspectives and empathize with the other's motives. Informal meetings led to workarounds, compromises and collaborative behaviors, thereby attenuating tensions.
Our data also suggests that given the inherent and entrenched nature of the paradoxical elements, any resolution or attenuation of tensions may only be temporary. Our informants noted that as situations changed, for example when new activities migrated to the SSU or new members joined the team, tensions re-emerged, requiring further renegotiation in search of an amicable communication over thorny issues, monthly service performance review meetings and regular health checks were established. Thus, building on our phasal model, we suggest that at the heart of managing inherent tensions in GDW is the ongoing negotiation of amicable and 'best fit' solutions at that time.
− Third, we extend prior research by empirically illustrating the interactions of tensions.
Our analysis brought to light interactions amongst knowledge, power and identity tensions, which have been largely implicit in current debate. Only a small set of studies have hinted at interactions of knowledge-power-identity tensions in GDW. For instance, Hahl et al. (2016) imply such interactions in their study of knowledge asymmetries between teams and brokers when they highlight the 'esteem' of those brokering as an important influence on teams' willingness to share knowledge and reduce asymmetries. Similarly, Levina and Vaast (2008) refer to the status of teams as an important facet of knowledge sharing and innovation in GDW, while Zimmermann and Ravishankar (2011) argue that asymmetrical status positions influence the professional role identities of offshore-onshore units. Overall however, interactions between knowledge, power, and identity tensions have not been explicitly considered or illustrated in prior GDW research. Our analyses point to such interactions and their importance for effective collaboration.

Defensive behaviors and interaction of tensions
In the suppression-amplification phases (see Figure 2), the SSU tried to strengthen its weaker power-identity position in the relationship (being perceived as a 'transactional' 'poorer cousin' with limited access to decision makers, less investment and limited authority to enforce process changes) by asserting its stronger knowledge position and insisting that the BUs use formal systems for information exchanges. Such an emphasis on its knowledge position amplified the knowledge tensions between the two sides (

Interactive-collaborative behaviors and interaction of tensions
The accommodation-attenuation phases (see Figure 2)  The SSU also worked to improve their power position by offering more 'valued' services, which opened access to decision makers and financial resources. The SSU also started sharing knowledge with the BUs: …there has been a shift from a pure processing focus (how many invoices have we processed?) to a real end-to-end view (P2.4). The attenuation of power and knowledge tensions not only removed the hostility between teams, but (as described in the case analysis) but also had a positive impact on reducing identity tensions (e.g., the 'us' and 'them' barriers

Implications for paradox research and GDW practice
Thus far, studies of paradox have focused on a central actor who is challenged with accommodating two competing ideas (Lewis, 2000;Smith, 2014), such as novelty and usability (Miron-Spektor and Erez, 2017), innovation and efficiency (Andriopoulos and Lewis, 2009), and competition and cooperation (Rai, 2016). In all these cases, the actor is positioned as the focal point, facing two opposing forces. As a point of departure, here we have elaborated on how one side of the paradox (BUs) is at odds with the other side (the SSU). This viewpoint extends the somewhat traditional focus in the paradox literature on one central actor (or group) who can comprehend each opposing element individually, to include actors (e.g., onshore-offshore units) that are often unfamiliar with the specifics of one-another's tasks, objectives, and values. Our analysis suggests that while formal strategies create a common lexicon and standardized rules for performing different parts of the distributed process, the inherently contradictory elements of the two GDW sides and the added complexity of distance mean that knowledge, power and identity tensions are harder to overcome. Our case suggests that organizations may need to take additional steps (e.g., structural realignments) to ensure that actors comprehend the inherently contradictory elements before attempting to engage in meaningful collaborative practices. Managers need to be sensitive, first and foremost, to the perceived power and identity tensions, which may amplify with the implementation of structural changes. They need to address the source of these tensions, by giving all sides the reassurance and confidence that their interests will be protected and by explaining the honest efforts being made to generate strategic value through the collaboration. The empirical material suggests that once individuals become better aware of, and empathize with, each other's positions, the perception of power asymmetries and identity threats are likely to subside. In such a relatively benign environment, as we saw in our data, conflicting teams are likely to take concrete measures on their own to reduce knowledge tensions.
The structural integration meant that the business division's CFO became responsible for the BU and SSU teams and in many respects acted as the arbitrator. The success of this simplified hierarchical arrangement has two implications for practice. First, it indicates that even when teams are physically separate, having a common leader can help mitigate tensions. Second, it suggests that it may be crucial for such a leader, in the aftermath of the structural integration, to engage with the conflicting sides on a 'human' and 'informal' level rather than escalate the dependence on formal IT tools and systems. It is possible that many managers leading GDW relationship rely on formal IT tools to ease coordination and communication challenges, and to minimize or overcome conflict. They may find this paper useful for designing or re-designing their management of distributed teams. Our study suggests that over-reliance on formal IT tools (e.g., the creation of a common lexicon and standardized processes) can be counterproductive. Instead, our findings suggest that organizations need to create opportunities for GDW teams to informally discuss, challenge, and negotiate power, knowledge and identity tensions in an open and transparent manner. One way in which organizations can support this may be to invest in platforms (online or offline) which offer GDW partners a safe space to engage in informal, and potentially constructive, meetings. However, a big challenge organizations face (and one that is admittedly not fully explored in this study) is understanding whether defensive behaviors stem from attempts to overcome inadequacies in the work system, or whether they are driven by the self-interest of actors taking short-cuts, resisting change and protecting their jobs. If it is mostly the former, then the challenge for managers is to balance the inherent paradoxes in a stable and coherent work system that fosters constructive GDW partnerships. If it is the latter, managers need to re-empower teams by better communicating the value of new (or evolving roles) and helping actors develop positive evaluations of their professional selves.

Limitations and future research
The arguments in this paper are based on a single in-depth case study. While this approach is consistent with the underlying principles and aims of qualitative research, care must be taken in transferring insights to other contexts. Another limitation of our study is that the data for phases 0 and 1 (Figure 2) was captured retrospectively. We encourage scholars to adopt longitudinal methods to capture timely accounts of how multiple tensions evolve as GDW units collaborate over time. The collaborative behaviors we saw from both sides led to what some participants described as 'small wins'. There was a better handling of trade-offs as both sides began working with each other. However, we acknowledge that managing trade-offs is not the ultimate objective here. Collaborative behaviors and better appreciation of mutual positions lay a solid foundation for genuine value creation through the offshore (SSU) -onshore (BUs) relationship. More research is needed to explicate the process of generating strategic value through collaborative behaviors in paradoxical GDW relationships. Our case focuses on inherent contradictions in the relationship between onshore BUs and an offshore SSU. However, it is likely that the corresponding tensions could play out differently in different outsourcing and offshoring arrangements. Further research is needed to explore such variations. More work is also needed to understand how collaborative practices emerge as outsourcing/offshoring arrangements become more complex with the application of automation and artificial intelligence technologies.
We have shown how even when they cannot entirely resolve the inherent contradictions in GDW, a sequence of formal and informal solutions can address the corresponding tensions and improve collaborative practices. However, it is possible that in some cases, a sequence of similar formal and informal solutions may worsen rather than improve opportunities for collaboration.
Future research needs to closely investigate the range of conditions in GDW that invariably lead to negative manifestations of tensions and poor relationships. Our case featured a relationship between onshore BUs and an offshore SSU. However, we have not investigated in detail the extent to which geographic distance influenced the tensions. Indeed, some of the tensions may also appear in co-located and proximally located teams although distance obviously introduces a new dimension to the relationship. More research is needed to compare and contrast the influence of a wide array of possible geographic distances from onshore units (e.g., offshore -in the same continent; offshore -in a different continent) on the evolution of tensions in GDW. Finally, given that empirical research on paradox has so far largely restricted itself to contexts of equal power relations, there is a need for a more explicit focus on management of paradoxes in settings riddled with significant power asymmetries (Hargrave and Van de Ven, 2017;Smith et al., 2017).

Concluding remarks
In this paper, we have drawn on paradox theory in considering the intrinsic and entrenched nature of knowledge, power and identity tensions in GDW. While it may be hard to completely and conclusively reconcile the paradoxical elements underpinning GDW tensions, our study suggests that a sequential enactment of formal and informal solutions can help situated actors move towards better collaborative working practices with their GDW counterparts. Tensions in GDW evolve in phases and require a careful ongoing negotiation between diverse groups of stakeholders.
Crucially, improvements in collaborative practices cannot be taken for granted since new demands, on say, an offshore unit may trigger new interactions amongst tensions and significantly harm the offshore unit's relationship with its corresponding onshore unit. Thus, effective management of paradox is a difficult and often fragile achievement in GDW settings. Given the increasing proliferation of GDW projects and initiatives, we would encourage scholars to build on our study and look closely into how specific interactions between tensions may be exploited to generate value for organizations engaged in GDW.