As we all see that global software development happens everywhere. Software projects are dispersing globally to search for cost reduction and time to market shorten, proximity to customer. In our study we consider the term global team in a broad range. It includes at least the offshore, outsource, virtual team and open source projects. In these projects the team members are working in different geographical dispersion. Here are some figures illustrate for the increase of global teams.30% of US IT jobs are expected to be offshored by 2015 160.000 projects registered in Source Forge in the end of 2011, which is doubled compared with that in 2004.Working in global team is a challenge not only for manager but also developer. When the projects disperse in multiple locations, the dependencies among the task and task carieer become more complicated. And it also more difficult for task management and coordination.
In Information System literature,Team coordination is considered as the important factors that influence the global project success. In this context, team coordination is defined as extra activity to o manage dependencies between task and task carrier.There are several previous works highlight the importance of team coordination in global projects. The inconsistency among the papers may be caused the different aspects of dispersion are investigated, the different team coordination context and different way of define project outcomes.
From this observation, conducted a systematic review to systemize the knowledge on this area and to identify the gaps for the future research. Basically I have four research questions, which fit into the input=process-outcome framework
From this slide I would like to present the review process we have gone through. We have 6 main steps .. We follow the review guideline written by Kitchenham
In the next step, we identify search terms, data sources, inclusion and exclusion criteria.
When conducting the paper selection, at first the result from the search engine is 11222 papers
The form is constructed, tested and modified in the adhoc review step
when team members locate in different physical places. . Temporal dispersion occurs when team members are separated through different working hours, time zones, and/or working schedules. Temporal dispersion also accounts for situations in which team members probably located at the same site but work different shifts or even flexible hours.Organizational dispersion occurs when team members are from different organizations. The team collaboration may face with organizational level issues, such as difference or even conflict of orragnizational strategy and objective, difference in organizational structure of different part of the team and formal issues like obligation or contract. Work process dispersion happens when team members work in different working environment and engineering process. The difference can be from adopted development methodologies, work infrastructure like computer, network or set of development and communication tool. Also some studies address the difference in distribution of experience and expertise among locations.Cultural dispersion is found as team members’ difference in communication language, cultural background, custom and national context.These dimensions are not clearly seperated and there are some overlaps and correaltion among them.
The picture shows the influence of geographical dispersion on team performance. We try to connect this relationship with the concept of team coordination in the between.The studies are categorized by the measure of the dispersion dimension. Red part shows the negative impact, green part show the positive impact and the grey part show the neutral result.All papers that compare distributed works with collocated works show that distribution is associated with lower team performance. The main reason is that the team need to invest more time and effort on team coordination.Papers quantify geographical dispersion shows a mixed result. Four studies reports … Three studies shows no association … and One study shows a
1 Dispersion, coordination and performance in global software teams: a systematic review Anh Nguyen-Duc, Daniela S. Cruzes, Reidar Conradi Department of Computer and Information Science, Norwegian University of Science and Technology 2/21/2013
2 Agenda • Motivation • Research questions • Review process • Review result • Conclusion 2/21/201
3 Motivation • Global software development are becoming more and more popular: – offshore team, outsourcing, virtual team and open source projects – 30% of US IT jobs are expected to be offshored by 2015  – 160.000 projects registered in Source Forge in the end of 2011 • Challenges in globally dispersed projects – More complicated task dependencies – More difficult team coordination  ACM Job Migration Task Force, “ Globalization and Offshoring of Software”, Association for Computing Machinery, 2006
4 Motivation • An existing model on team input, process and IS project outcome : Team coordination Dispersion Project outcome (Social- emotional (Input) (Output) Process) • Existing empirical studies is inconclusive about the impact of dispersion on team coordination and project outcomes – Dispersion dimensions – Team coordination context – Project outcomes measure – Influence direction J. A. Espinosa, W. DeLone, and G. Lee, "Global boundaries, task processes and IS project success: A field study," InformationTechnology and People, vol. 19, pp. 345-370, 2006
5 Research questions RQ1: Which dimensions of dispersion are explored? RQ2: How is team RQ1 Context factors RQ4 coordination influenced by these dispersion dimensions? Dispersion Performance RQ3: How is team dimension performance influenced by these dispersion dimensions? RQ2 Team RQ3 Coordination RQ4: Which context factors could explain the heterogeneity among empirical findings on the influence directions? 2/21/201
6 Review process Literature review Purpose: Provide knowledge background Protocol Collect key words to build the search string development Construct data extraction form Result: Paper selection 27 seed studies Data extraction Quality assessment Data analysis 2/21/201
7 Review process Literature review Search string: (coordinati* or collaborativ* or cooperati* ) AND Protocol (distributed or offshor* or "open source" or outsourc* or development  global or dispers*) AND (software or project or team) Data source: Paper selection Scopus, ISI Web of Science, Reference list Exclusion criteria: Data extraction 1. Short papers 2. Not in SE or IS area 3. Not about dispersed context Quality 4. No empirical report or validation assessment 5. Study team coordination without relationships with project outcomes Data analysis B. A. Kitchenham, “Guidelines for performing Systematic Literature Reviews in Software Engineering”, EBSETechnical Report, 2007 2/21/201
8 Review process Literature review Search result ……………… 11222 unduplicated papers Protocol development Selected by reading titles and abstracts ….. 470 papers Paper selection  Selected by reading full text …….…………….48 papers Data extraction Extra papers by reference scan ….…………….8 papers Quality assessment Total papers to be extracted …...……………..56 papers Data analysis B. A. Kitchenham, “Guidelines for performing Systematic Literature Reviews in Software Engineering”, EBSE 2/21/201
9 Review process Meta data Literature review Study design, Protocol background concept development Context setting Paper selection Independent factors Dependent factors Data extraction Control factors Quality assessment Findings, threats to validity Data analysis 2/21/201
10 Review process CHECKLIST Literature review Problem statement 56 extracted papers 1. Is the aim of the research sufficiently explained and well motivated? Research design Protocol 2. Is the context of study clearly stated? 3. Is the research design sufficiently prepared development beforehand? Data collection 4. Are the data collection and measures adequately Remove 8 papers: described? 1. Poor research design Paper selection 2. Insufficient data 5. Are the measures used in the study relevant for answering the research question? 3. Poor/ No data analysis Data analysis conducted 6. Is the data analysis used in the study adequately Data extraction described? 7a. Qualitative study: Are the interpretation of result clearly described? 7b. Quantitative study: Are the effect size reported Quality with assessed statistical significance? Studies on team assessment  8. Are potential confounders adequately controlled performance: 28 papers or discussed? Conclusion 9. Are the findings of study clearly stated and Data analysis supported by the results? 10. Does the paper discuss limitations or validity? T. Dybå and T. Dingsøyr, “Empirical studies of agile software development: A systematic review”, Information and 2/21/201Soft- ware Technology, vol 50, pp. 833-859, 2008
11 Review process Tailored thematic analysis  (RQ1, RQ2) Literature review Extract code …...……………………………………….. RQ1 ……………………………….....53 codes Protocol RQ2 ………………………………...137 codes development Identify common themes……………………..……….. RQ1 ………………………………….5 themes Paper selection RQ2 ………………………………….8 themes Vote counting (RQ3, RQ4)  .….……………………….. Data extraction Geographical dispersion ………… 14 studies Temporal dispersion ………………. 8 studies Quality assessment  D. S. Cruzes and T. Dybå, “Recommended Steps for Thematic Synthesis in Software Data analysis Engineering”, pp. 275–284, ESEM, Calgary, Canada, 2011  L.M. Pickard, B.A. Kitchenham, and P.W. Jones, “Combining empirical results in software engineering,” Journal on Information and Software Technology, vol. 40, Dec. 1998, pp 811-821 2/21/201
12 Demographics Publication by year 7 6 5 4 3 No of study 2 1 0 2003 2005 2006 2007 2008 2009 2010 2011 2/21/201
13 Demographics Publication by research methods 14 12 10 8 6 4 2 Data collection 0 2/21/201
14 Demographics Global dispersion type Global branch Outsourcing Open source No. of studies Laboratory 0 2 4 6 8 10 12 14 16 18 20 2/21/201
15 Result RQ1: Which dimensions of dispersion are explored? Dispersion dimensions Geographical Temporal Organizational Work process Cultural dispersion dispersion dispersion dispersion dispersion (16) (8) (8) (7) (5) 2/21/201
16 Result RQ2: How is team coordination influenced by these dispersion dimensions? Coordination problem Frequency of communication and feedback Choice of communication mean Coordination delay Perception and attitudes toward collaboration Misinterpretation Coordination requirement gaps Team structure configuration Task scheduling complexity 2/21/201
17 Result RQ2: How is team coordination influenced by these dispersion dimensions? Coordination problem Geo. Tem. Org. Wor. Cul. Frequency of communication and X X X X feedback Choice of communication mean X X X X Coordination delay X X X X Perception and attitudes toward X X X X collaboration Misinterpretation X X X X Coordination requirement gaps X X Team structure configuration X X Task scheduling complexity X 2/21/201
18 Result RQ2: How is team coordination influenced by these dispersion dimensions? Coordination problem Geo. Tem. Org. Wor. Cul. Frequency of communication and X X X X feedback Choice of communication mean X X X X Coordination delay X X X X Perception and attitudes toward X X X X collaboration Misinterpretation X X X X Coordination requirement gaps X X Team structure configuration X X Task scheduling complexity X 2/21/201
19 RQ3: How does the team performance Result influenced by these dispersion dimensions? Dispersions is associated with lower team performance Negative impact on team performance on team & task level Positive impact on team performance on project level No association with team performance 2/21/201
20 RQ3: How does the team performance Result influenced by these dispersion dimensions? Perception about team performance Direct measure of team performance at project level 2/21/201
21 RQ3: How does the team performance Result influenced by these dispersion dimensions? No consistent picture from empirical studies on the influence of dispersions on team performance
22 Result RQ4: Which context factors could explain the heterogeneity among empirical findings on the influence directions? Variables: – Lack of data: dispersion type, number of sites, level of communication technology and practices – Study subject, sample size, quality of study – Team performance measure type, Unit of analysis Unit of Geographical Temporal analysis Pos. Neg. Neu. Pos. Neg. Neu. Task 0 5 1 0 0 0 Team 0 4 1 0 3 1 Project 1 1 1 3 1 0 2/21/201
23 Implication for future research Research Include and distinguish among different type of dispersions Report dispersion context and level of communication technology and practices Further research on how work process & cultural dispersion impact team performance Further research on impact of dispersion on mechanistic coordination Further research on dispersion on open source projects 2/21/201
24 Implication for practice Practice Understand that impact of dispersion is context-specific Promote technology and working style that support effective informal communication Configure team structure that addresses coordination requirement Be aware of positive effect of temporal dispersion on team performance Look for evidence at team and work unit level to decide the cost-benefit of being distance.
25 Q&A Contact • Anh Nguyen-Duc: firstname.lastname@example.org • Daniela S. Cruzes: email@example.com • Reidar Conradi: Reidar.Conradi@idi.ntnu 2/21/201