As discussed in the cases, GSA relationships can be seen to pass through an evolutionary process towards maturity . We have interpreted these phases of the relationship based on amultiplicity of considerations, including the kind of work being done, the investments being made in infrastructure and the level of trust and understanding that exists between the partners . GSA relationships are far from being static, and tend to develop in scale, scope and form over time. A typical phase in a life-cycle from initiation, growth to maturity is schematically depicted in figure 11.1 (based on Murray and Mahon 1993). Following initiation, where most GSAs ˜start small with non-mission-critical projects, an increased proportion of the development tasks tend to be moved offshore where the costs of development are cheaper. A key question then relates to how much and what kind of development tasks should be moved offshore and when, thus making managing the balance between offshore and on-site work an ongoing challenge. The onshore “offshore mix is not static and shifts depend on peaks and troughs in workload and the kinds of work being done. If the relationship moves to maturity, higher-value activities may be undertaken, involving new forms of joint work. Alternatively, the partners may separate amicably temporarily or permanently (like a ˜divorce ). The time for evolution may be short or prolonged depending on events and circumstances, whether the relationship is a ˜one-off -type project or a long- term commitment.
As GSA relationships evolve over time, the contractual basis and organizational forms may also change. A trading relationship initiated with a number of ˜one-off projects based on time and material costing may over time evolve into an alliance or JV. JVs and other strategic alliances may subsequently end in the sale of one partner to the other. While not all GSA relationships follow this linear trajectory, it was relevant to the Sierra case; Sierra s GSA started with a trading relationship with offshore outsourcing companies and subsequently led to them opening up their own subsidiary. Gowing initiated its GSA strategy with ˜body-shopping on a time and material basis governed by a service-level agreement. The partners later moved to a new contract that covered the whole of the Cass group . In the new alliance relationship, some projects were outside of Gowing but within Cass and were valued on a turnkey or project-cost basis. In contrast, GlobTel s externalization programme started off as a ˜vendor “contract relationship and contractually remained the same after 10 years even though the content of the relationship had changed significantly over time.
The model presented in figure 11.1 does not indicate the possibility of non-linear movements in the GSA relationship, such as going in reverse, as was the case in Sierra, which had to shut down its India operations. There are various reasons for this ˜non- linear movement, including disruptive events in the industry and shifts in the balance of power and control between the partners and the consequent implications. As firms progress up the ˜trust curve to higher-value activity, the customer can become over- reliant on the outsourcer, creating a dependency situation that can be exploited through demands or refusals of price increases . At one point Gowing management carefully considered the reliance it had on Eron, and evaluated the option of appointing a portfolio of software suppliers (instead of just one) within a network of relationships that would also include Eron. Gowing eventually decided against this option because of the risks associated with new relationships, and instead entered into a closer alliance with Eron. GlobTel also considered and subsequently opened their own software development centre in India to guard against the possibilities of over-reliance on Indian firms and the consequences it might have on their activities in India.
Managing relationship evolution in GSAs is a key challenge, and one that is shaped significantly by issues of trust and its dialectical relation with power and control. Control and coordination of activity across time and space is a fundamental issue related to trust. Trust and control are not ˜either “or conditions, one presupposes and depends on the other. While control is crucial to successful completion of IS projects using the optimum resources, the need for control can be minimized in a trusting environment. Control in GSAs is concerned with the process by which one partner influences the behaviour of another and of the alliance itself (Faulkner 1999). Control mechanisms include traditional forms of equity ownership and board membership to mediate knowledge-transfer processes; training and personnel development; contracts with specified service-level agreements; and standardization of skills and processes. These structural controls are enforced through mechanisms such as rewards and punishments; penalty clauses; reporting procedures; client involvement in developing plans; and specifying methodologies for measuring outputs.
The relationship between trust and control can be examined from four perspectives:
Calculus-based trust (Faulkner and Child 1998; Sabherwal 1999).
While knowledge-based trust concerns shared experiences and personal relationships, identification-based trust is concerned with the extent of synchronizing of goals, objectives and expectations between both sides of the relationship. Performance-based trust is concerned with movement up the ˜trust curve through early project successes, interim deliverables, demonstrations and pilots. Calculus-based trust is related to the deterrence and rewards which are crucial to the successful completion of IS projects using optimum resources. We discuss in the following paragraphs the dialectical relation between trust and power and control.
While knowledge-based trust is concerned with shared experience and personal relationships, it is ˜grounded in the other s predictability “ knowing the other sufficiently well so that the other s behaviour is anticipatable (Lewicki and Bunker 1996: 121). At Sierra, developers in the UK lacked the assurance of knowing their Indian partners sufficiently enough so that their ˜behaviour was anticipatable . The security and comfort offered by a well- understood partner was undermined by communication problems and ˜misinterpretations , caused in part by differences in respective background knowledge and worsened by the problem of attrition. The lack of informal socialization in a ˜community of practice contributed to both sides not being able to understand the thinking or reasonably predict the actions of their partners in India or the UK. Indian companies are attempting to manage knowledge-based trust better by localization of their offices in North America or Europe to help present a local ˜front-end for direct liaison to enable more effective knowledge transfer. This ˜front-end is enabled through the opening of subsidiaries, acquiring or partnering with foreign companies to gain access to local domain experience and markets. Similarly, North American and European software houses have themselves opened subsidiary development centres or partnered with Indian companies to offer their customers an ˜indirect route to outsourcing through a local provider. Sierra attempted this route; another example is Arrk (www.arrk.net) which has an established JV with a Bombay software house. At the time of writing Arrk was planning to open their own Bombay subsidiary.
The subsidiary arrangements of the major companies such as Texas Instruments have contributed to knowledge of the offshore development process. Various middlemen, often experienced , well-connected ex- employees can help facilitate and foster relationships or set up and incubate offshore subsidiaries. Government and quasi-governmental organizations are tasked with facilitating GSA. The UK Department of Trade and Industry s ˜Enterprise India Initiative (www.tradepartners.gov.uk) claims considerably to smooth the GSA maturity process. The DTI provides a partner-matching service, expert advice and organized trips to visit India and suppliers. For some of these reasons, GlobTel proposed setting up it offshore development centre in India to deal with the knowledge gaps it felt existed in its understanding of India “ for example, how to deal with attrition. This move to come closer to India was, however, seen by the Indian firms as an exercise of power and control, leading to tighter micro-management. MCI s initiative to develop a proximity development centre in North America, with a view to developing a closer understanding of GlobTel s management practices “ that remained largely tacit despite the many years of the relationship “ was rejected by GlobTel on the grounds that it would add to development costs. We argue that while knowledge-based trust is useful in developing shared knowledge and expectations, it has continuously to contend with the implications for power and control.
Identification-based trust is concerned with the synchronizing of goals, objectives and expectations between both sides of the relationship and operates in different ways. At the strategic level as well as the operational levels of the relationship, identification is linked closely to the role of ˜ champions . Gowing and Eron s relationship was founded on the strong personal relationship and friendship between the Eron Managing Director based in the UK and Jones, Gowing s Managing Director. Jones became a powerful GSA ˜champion and strongly advocated Eron moving up the value chain partly through outsourcing and its extension to other parts of Cass. By virtue of their Indian roots, Ghosh and Paul similarly acted as ˜ shock absorbers in the GlobTel “India relationships. They served as conduits to develop local knowledge and to deal with crisis situations as and when they developed. However, both the Indian and North American management often criticized this strong identification as it was felt to impede developing stronger business oriented control structures. In Sierra, Mitra was seen to have a similar ˜champion role on account of his Indian origins and the identification that this implied . However, Mitra s understanding of the complexities in India was rather superficial and was reflected in his attempts to develop strong structures for power and control (to deal with the issue of ˜leakage , for example). The adverse effects might have been mitigated by development of a stronger sense of identification with his Indian colleagues. In Sierra, at the project level, developers at the two sites did not share the close ˜bonding friendships crucial for this type of trust to evolve. The physical and cultural distance between the sites made informal socialization impossible to achieve, and the staff could not genuinely ˜act for each other . In contrast, at the mature stage of the relationship the Gowing and Eron staff shared a higher level of identification with each other s needs, and this led to close bonding friendships.
Performance-based trust concerns building up confidence with respect to the ability of the partners to effectively carry out the tasks they are supposed to on time and on budget. In the GlobTel case, there was initially limited performance-based trust as the Indians had little prior experience in telecommunications and also in the conduct of GSW. But they had the willingness , aptitude and intellectual capacity to learn, and in the process they provided GlobTel with the potential for such performance-based trust to develop in the future. Efforts to build this performance-based trust had to confront a number of tensions along the way, for example the fact that telecommunications and software presented very different domains of expertise. These differences had serious implications for the relationship, such as in dealing with the question of attrition and in the implementation of and compliance with various control structures. As ComSoft gained in performance-based trust and obtained more ownership of GlobTel technology, it developed the confidence and need to forge its own growth path by exploring new Japanese markets. This led to a shifting of the power balance and a lessening of dependence on GlobTel for survival and growth. Sierra s ambitious goals and early failures with performance (reflected in the issue of ˜leakage ) led to an organization-wide dissonance. This contrasted with the Gowing case, where Eron quickly fulfilled the aspirations of Gowing management and moved along the GSA life-cycle and up the value chain. As a result of this performance-based trust, the Eron developers were used to create a ˜rock of discipline within the UK framework, attempts resisted by some groups in the Cass group (ex-RDC people) who felt that their own performance was being undermined.
Calculative trust is related to deterrence and the rewards crucial to successful completion of IS projects using the optimum resources. This form of trust tends to be strongly associated with the early stages of GSA relationship growth, before other forms of trust (performance and identification) grow. Calculative trust reflects a dialectical relationship that can be metaphorically viewed as ˜walking a tightrope . Project teams , for example, walk a tightrope between an exercise of freedom to organize their work and deliverables, while being subjected to exhaustive reporting and micro-management controls. At Gowing, in the early stages of the relationship, structural controls dominated through the use of structured methodologies and extensive reporting tools, including key performance indicators (KPIs) represented on a ˜dashboard on the Intranet. Strategic-level controls were exercised through varying the size of contracts and leaving open the potential for further contracts. During the later stages, with growing maturity and Eron s increasing integration within Gowing, the relationship depended more on knowledge and identity forms of trust. At GlobTel, standardization was a key force for long-distance control of Witech, supplemented through the use of expatriates and training programmes within the ˜global manager ˜template . Similar attempts at standardization in ComSoft met with resistance “ for example, their refusal to have an expatriate located on their premises. The ComSoft and Witech cases provide contrasting examples of how GSA strategies tread a thin line between control and standardization on the one hand and the need to nurture local identity and creativity on the other hand.
In summary then, the lens of a dialectic between control and trust through the different stages of a GSA relationship is useful for managers to develop the implications of managing the evolving process. This lens helps to formulate some key questions, summarized in box 11.4.
What is the role of GSA ˜champions in helping achieve the objectives of the relationship?
How can various forms of trust be developed and integrated for effective performance and the growth of maturity in the relationship?
What control mechanisms are appropriate at different stages of the relationship and how do they impinge upon issues of trust? How should control be balanced against such different forms of trust ?