MCI is part of one of the largest industrial groups in India with companies in a diverse range of industries: steel , automobiles, hotel chains, etc. The group turnover is around $5 billion. MCI resulted from the outcome of efforts of some pioneering senior executives in this group. They were one of the earliest in India to recognize the opportunities for software work globally and to establish an organization to translate the opportunity into a major business. MCI, from small beginnings, has grown significantly and is currently one of the largest software service providers in Asia. It has a network of more than fifty offices in the USA, Europe and Japan, with more than 10,000 employees . It had a turnover of around $400 million in 1999.
GlobTel s relationship with MCI started in 1991, described as a ˜trialling exercise with ten developers working in Mumbai. There was an initial sense of uncertainty and excitement about how the relationship would develop since MCI was traditionally a software consultancy firm with limited experience in telecommunications. Also, MCI had historically specialized in on-site work (derogatively referred to as ˜body-shopping ) and had relatively limited experience in offshore long- term development relationships, of the kind that a GSA entailed. In 1993, after a visit to North America, Mani (MCI s account manager for GlobTel) announced to his staff in India ˜we need to become people in telecoms rather than in software . Mani speculated on the three stages that MCI could go through in such a relationship: (1) to be a vendor on a fixed-price quotation basis; (2) increasingly to enter the core areas of switching; and (3) to become GlobTel s ˜satellite lab by taking on ownership of some GlobTel products.
At this stage, MCI identified two key challenges:
How to manage the intricate relationship between technology and the issues that arise from working at a distance across time, space, and cultures.
How to mesh the project culture of MCI with the product culture of GlobTel. Mani said:
We are basically a project oriented company. We are not a product company, whereas GlobTel is a product company. So, I think there is a first challenge to try and mesh the people up, project culture versus the product culture. Project culture is typically a software consultancy or a matrix organization. We put a team together for a project. We do the project. After successfully completing the project, we hand over to the owner of the project and we disband the team till we start something else. Whereas in a product company like GlobTel, they expect a person to be there for a very very long time because it takes about 12 months to start appreciating the product.
Investments in technological infrastructure started to play an important material and symbolic role in the relationship. GlobTel set up a 9.6 kB baud link between India and North America without a video link, suggesting that GlobTel still wanted to ˜test the waters and keep MCI at arm s length. At this stage, GlobTel gave MCI mostly stand-alone work that was independent and involved minimum interaction with the company. As work started to evolve , GlobTel upgraded the telecommunications links to 128 KB including video lines, signalling to MCI an expectation of being ˜close . GlobTel established an expensive development environment in Mumbai that replicated their North American facilities. These infrastructural investments were reciprocated by MCI who established a GlobTel-dedicated lab in the suburbs of Mumbai where real estate was at a high premium. These mutual investments relied primarily on good faith, without a legally binding long-term commitment. They gave MCI the confidence of being ˜included in GlobTel s plans.
Expanded investment was matched with an increase in the number of developers in MCI s GlobTel group. They increased from 10 in 1991 to 70 in 1994 to about 300 in 1996, with an expectation of growth to 400 by the end of the year. GlobTel accounted for about $7 “8 million of MCI s revenue, nearly 5 per cent of their overall international business operations, it was one of MCI s top 10 customers. Initially, MCI did ˜lower-level bug-fixing kind of projects that were relatively independent from the need to interact with GlobTel. Over time, there was increasing interdependence as MCI became involved in GlobTel s core areas of switching which required them to acquire domain-specific knowledge of telecommunications, of GlobTel products and at the management level to incorporate GlobTel s practices such as those related to internal reviews and quality checks. This interdependence created an increased need for ˜proximity , and GlobTel upgraded their communication links from two to five “eight voice paths in 1994 “5 that gave them more bandwidth to work with.
With this evolution of work, GlobTel became increasingly concerned with the loss of programmers (referred to as ˜attrition ). After being trained and having acquired skills in the telecommunications domain, MCI programmers would seek more lucrative jobs, primarily in the USA. GlobTel tried to get MCI to institute measures to prevent this. MCI resisted what they called a ˜Berlin Wall Approach and instead saw attrition as something that could not be prevented but was a ˜way of life that needed to be managed. MCI desired autonomy to deal with the attrition problem according to norms existing in their organization. GlobTel and MCI subsequently both took a number of measures to deal with attrition. GlobTel located North American expatriates to MCI to try and ˜install the GlobTel culture and make MCI systems more ˜objective . MCI tried to create an internal ˜GlobTel island to prevent the movement of developers from GlobTel to other groups. MCI tightened their recruitment procedures and emphasized to potential recruits that they were being hired for telecommunications (not software) where the learning curve was much longer and more complex. Recruits were expected to give a verbal commitment to stay in the group for a minimum of 2 years , a requirement that did not extend to other parts of MCI.
Despite these initiatives, GlobTel s perception was that MCI s effort was inadequate. GlobTel was unhappy about MCI s resistance to their efforts to install new management systems, and viewed them as an ˜outlier in their international R&D relationships. Ram, a GlobTel expatriate of Indian origin, thought MCI was reluctant to take ownership of ongoing problems. GlobTel managers felt that MCI simultaneously relied on them for their vision, but also blamed them for ˜shoving things down their throats . MCI s perceived lack of cooperation and proactivity made Ram remark that GlobTel should critically re-examine whether the relationship was sustainable in the long run. Another GlobTel expatriate (of American origin) also was highly critical of MCI s management abilities and believed that the Indian managers were at least one or two levels lower in competence than their North American counterparts. She saw her primary mandate in India as trying to bridge this skills gap.
The period of growth was characterized by an increasing maturity in MCI and growing expertise in MCI both about GlobTel and telecommunications. In both India and North America, there was greater visibility to the relationship, which evolved from a ˜contractor to a ˜partner mode of working. Even with persisting concerns about attrition, micro-management and MCI s resistance to GlobTel s attempts to standardize, a number of key events took place in 1997 “8 that gave a strong impetus to the relationship. A number of very senior GlobTel staff visited MCI and made a public commitment to support them. The GlobTel CEO, inaugurated the new MCI lab in Mumbai, and promised to clear obstacles at the strategic level to enable more effective joint work between MCI and SMTP (a GlobTel lab in Texas, USA). There was also a visit by Clement, the head of GlobTel switching operations in North America that accounted for $1 billion annual business, who announced that MCI was designated as the ˜brother lab for SMTP.
These public events signified a serious business commitment, moving MCI employ - ees from a stage of ˜regional-level activities to an ˜Olympic level . Mani felt that ˜A brother-lab implied a move beyond the stage of technical content to a stage of true partnership involving ownership and collaboration . This heightened status to the relationship spurred both sides into action. While GlobTel was putting more money, people and executive time to make things happen, the MCI staff was actively debating the implications of being a ˜brother lab where they would not need to worry about project-to- project issues but on the overall strategic direction. They felt that the brother-lab status meant that GlobTel were affirming significant trust in them by inclusion in the ˜family .
Contributing to this significant growth was GlobTel s upgrading of the telecom link from 128 kB to 2 MB that would allow more interdependent work to take place. MCI felt that now GlobTel could not take decisions unilaterally by wearing their ˜customer hat , but mutually in consultation with them. Mani gave an example of such a consultation:
I can tell you what a 2 MB link really means to the relationship. A senior person [from SMTP] had collected a lot of the evidence to substantiate his concern [against a particular MCI contractor]. A simplistic thing is to replace him, OK. Now you have this as an opportunity to turn this relationship into what it means because in the contractor model you do that, ˜I don t like the contractor, please replace him . But if there is ownership and a 2 MB link which indicates the closeness of the relationship, you would go and say ˜I have the problem and how do we solve it? . I cannot insist on my solution to the problem, because it is supposed to be our problem and I respect that.
During this period of growth, the problem of attrition became more significant. While both GlobTel and MCI were in agreement that attrition was a problem, they had contrasting views on how it should be dealt with. While GlobTel believed that attrition should be prevented, MCI saw it as part of life that had to be accepted and its consequences managed more effectively. MCI felt that GlobTel s ˜Berlin Wall Approach to prevent developers from leaving was a paradox. It curtailed developers freedom while still providing them with the perception that they were free to move if they desired. Mani felt that attempts to prevent developers moving would be counterproductive and criticized GlobTel s standardized quantitative approach to solving even a human issue such as attrition:
The case in point is the standard problem of attrition and the GlobTel approach or the North American approach, I would not call it a GlobTel approach, is applied to it. I think the North American approach is to identify a need or a problem, collect data on that, put up a plan, put targets and track it to closure. So you take attrition, measure what it is, what is the attrition today, decide on a target. So supposing yousay attrition today is 20 or 25 percent, you say I want to bring it down to 15 percent. So you say in the next six months we would bring it down from 25 to 20 percent, and then from 20 to 15 percent. There are certain problems, like the problem of attrition, that are far more complex, more local, where such a project oriented planning will not work.
The spectacular growth in the work demands created issues to which both sides responded by trying to increase proximity to the other. With increasing stakes in the relationship, GlobTel s managers felt the need for greater control through increased physical proximity. That required them to have a stronger physical presence in India. As work became more interdependent, MCI simultaneously felt the need to understand GlobTel s ˜way of doing things and sought greater physical proximity to them in North America. In seeking proximity, both sides expressed the need for a place-based rather than the space-based environment in which they had started.
GlobTel considered establishing their own Software Development Centre (GSODC) in India, as it would increase their first-hand and intimate understanding of the Indian context and help develop a sense of ˜place . For the centre to work, expatriates would need to be moved from North America to India at high relocation costs and salaries. This expense would be in opposition to GlobTel s initial objective of cost reduction through offshore work in India. The GSODC could potentially also allow GlobTel to sidestep some of the IP concerns arising from development through an external contractor like MCI. MCI resisted the GSODC proposal as they saw it as a direct competition to themselves . Development of new and core technologies would be handled by GSODC, and they be given the ˜crumbs . In the long run, this would stunt their movement up the value chain and they would also be more closely controlled from GlobTel.
With increasing consolidation of MCI s work following the ˜brother lab announcement, anumber of projects were transferred to them, signifying a move from Level 1 to Level 3 work.A senior MCI manager described Level 1 work as ˜the easy, by-the-kilogram-kind of stuff (for example, bug-fixing and regression-testing projects), in which developers became productive and could deliver within 2 “3 months. Level 2 work involved adding new features to existing software. Level 3 work involved responsibilities for the entire product, including developing enhancements and responding to customer complaints at service levels similar to GlobTel. The move from Level 1 to Level 3 implied a massive leap of faith for both sides, which Mani described in the following manner:
The buck-stops-with-you kind of thing. That is what I call throwing your hat over the trapeze, Level 3 is the highest level. You are literally an equal partner of GlobTel.
The move to Level 3 implied increased expectations and responsibilities for MCI. It created the need for them to better understand GlobTel s ˜way of doing things . Mani theorized that GSO relationships involved the transfer of knowledge of products (for example, specifications of DMS switches), of processes (for example, of quality control) and of practices (referring to local ways of doing different kinds of tasks ). He felt that while GlobTel had done very well on the first two, they had generally overlooked practices, resulting in misunderstandings and often conflict between the two sides. MCI s managers were perplexed and frustrated as they felt that, despite meeting targets, GlobTel remained unsatisfied. While sitting in some GlobTel s internal meetings in North America, Mani realized the reason for this lack of satisfaction. He saw that the templates used by the North American managers to report results from projects were quite different from those used by the MCI managers. While the Indians presented only hard details of the project (deliverables, time deadlines, etc.) the GlobTel staff also gave details on intangibles such as the satisfaction of users, etc. The Indians did not report such intangibles since, Mani felt, they were rooted in situated practices that could not be transmitted in training programmes as easily as products and processes. To be able to do so in the future, Mani proposed to GlobTel that:
[We]...must have a management entity that is distinct from GlobTel, [our] own entity in North America. Let us concede the other way that if MCI is incapable of doing it even in North America they are definitely not going to do it in Mumbai; you might as well give up. If I cannot run it out of North America, I cannot run it out of India anyway. I have no excuses such as lack of knowledge or practices. I have no such excuses anymore...
MCI believed that the increased costs at GlobTel by higher on-site presence would be marginal, and the total would still be less than North American costs. More importantly, did the managers with this enhanced proximity gain an increased ˜ comfort level , especially in the prevailing context of churn and turbulence in the industry? GlobTel resisted MCI s proposal to increase their on-site presence to 30 per cent, and on the contrary, it was reduced from 15 to 5 and was later renegotiated to 10 per cent. Against the background of these ongoing tensions of proximity (place) and distance (space), we describe the phase of stabilization of the relationship.