Managing global software development teams : technology and policy proposals for knowledge sharing
Author(s)Seshasai, Satwiksai, 1980-
Massachusetts Institute of Technology. Technology and Policy Program.
MetadataShow full item record
This thesis uses an in-depth case study, with integrated data analysis, to compare and contrast globally distributed and co-located software teams within the IBM Corporation. Important differences in information sharing, collaboration and other behaviors were observed, along with a range of innovations in work operations. Technology and policy implications that draw on the benefits of each model are identified. The development of this thesis began with a seminar that was conducted at MIT to invite stakeholders in various areas of knowledge-based offshore outsourcing to discuss strategic, economic, organizational and technical issues raised in various environments. Large and small firms, various industries, and various business models were covered. The context provided by these stakeholders was used to design an in-depth case study at IBM, with the focus on a matched pair of software teams, which were studied for a period of one year. Both software teams were identical in aspects such as product scope, team size and domain; however, they differed in the key aspect that one team's members all work on the same hallway while the other team's members are geographically dispersed among multiple locations in the United States and India Quantitative technical data from the source control system of each team, the software problem report database, frequency and content of group emails, weekly meetings, and individual interviews were combined with qualitative data from stakeholder interviews to distinguish key benefits and challenges of each model. The quantitative measures gauged data such as frequency and methods of collaboration, social and technical networks, and differences in handling strategic and tactical decisions.(cont.) The qualitative interviews discussed stakeholder perceptions of the quantitative data, and their motivations for decisions related to knowledge sharing. Key findings from the data include a number of observations about specific forms of knowledge sharing which differentiate the two teams. The distributed team used electronic mail as a forum for discussion which peaked around project deadlines, while the collocated team relied on e-mail as an announcement mechanism. Team meetings for the distributed team were much more tactical and task oriented in nature than meetings of the collocated team. With regard to the technical project itself, developers on the collocated team shared source code to a much greater extent, however status input to the software problem report database was much more interactive on the distributed team. This thesis is also important for pioneering highly precise indicators of team interactions based on the coding of archival data derived from e-mail, telephone, meeting and other interactions. The methods developed hold great promise for further studies of design teams, as well as a feedback tool that could be highly valuable for these teams. A number of emerging themes were found in the data analysis, which suggest that lessons from this study need not only apply to cases where geographic distribution is a factor. The teams consistently showed that the same technologies, processes and stages of the project lifecycle can be handled very differently based upon context. Also, social relationships and dominant individuals on a team can have an impact on technical productivity.(cont.) Finally, the evidence in this case suggested that geographic structure need not define destiny, and in some cases geographic structure can be used as an asset. The data analysis points to preliminary technology policy implications at the individual, team, organization, and national levels. At the individual level, it is recommended that workers in distributed teams alter work hours to devote a few minutes after-hours to synchronous communication with team members in different time zones - something that happened more often among the members of the co-located team. On the other hand, in a collocated team, it is recommended that the team use software tools to discover technical expertise that is more formally recorded among the members of the distributed team. At the team level, specific added value gained unintentionally from one geographic structure - such as greater documentation of decisions on a distributed team - can be achieved in co-located teams. At the organizational level, this thesis provides methods for assessing an organization's tacit knowledge capital at a much more granular level than tabulating patents or licenses. A number of institutions such as corporate training and development departments, labor unions, professional associations and government education and training initiatives may be impacted by the changes to workforce training and work methods suggested by this thesis.(cont.) At the national level, lessons from these teams demonstrate that the drivers for policy decisions related to offshore outsourcing need to be adapted in knowledge-based industries which have the potential for globally shared tasks, and export regulations dealing with intellectual property exchange in global software teams need to account for the daily trade of IP in geographically distributed teams. As the thesis focused on one in-depth case study, a significant effort is made to propose future research directions which can validate the proposals with broader data collection.
Thesis (S.M.)--Massachusetts Institute of Technology, Engineering Systems Division, Technology and Policy Program, 2005.Includes bibliographical references (leaves 94-97).
DepartmentMassachusetts Institute of Technology. Technology and Policy Program.
Massachusetts Institute of Technology
Technology and Policy Program.