A risk management framework for distributed agile projects

https://doi.org/10.1016/j.infsof.2016.12.005Get rights and content

Abstract

Context

Distributed agile development (DAD) approach has been adopted by the software companies for cost and time benefits. However, it causes significant challenges considering the contradicting nature of the agile and distributed development.

Objective

The objective of this study is to develop a risk management framework that comprises the perceived risks in DAD projects, their causes and the methods used in industry for managing those risks.

Method

This work is an extension of an exploratory study, wherein, DAD practitioners reported the risks they face in projects and the methods they use for managing those risks. The identified risks were further categorized based on their relevance to different aspects of DAD projects. In this extension, industry practitioners ranked the risks for their impact on DAD projects and rated the methods for the frequency of their use in projects. As the number of risks under each category was large for ranking, they were grouped under the risk areas within each category. The ranking of risk categories, risk areas and risk factors for their impact on DAD projects manifests their importance. The framework includes ranked risks, their causes and the risk management approaches. It was partially implemented in live projects in three different companies and was found to be beneficial.

Results

The perceived impact of the risk categories, ‘Group Awareness’, ‘External Stakeholder Collaboration’ and ‘Software Development Life Cycle’ on DAD projects has been found to be high and caused by the properties of Distributed Software Development (DSD). The partial validation of the framework in three companies reported the elimination of majority of risk factors and/or reduction in their impact.

Conclusion

DAD projects provide significant benefits but hold substantial risks due to the contradiction between distributed development and agile practices. The reported framework could effectively minimize the DAD risks in practice.

Introduction

Distributed Agile Development (DAD) is intended to attain the quality and speed benefits of agile along with cost benefits of distributed software development (DSD). However, it induces significant risks due to the differences in their key tenets [35]. Agile methods encourage frequent face-to-face communication and practices to build trust, while the distance in DSD signifies different organizational cultures, standards and policies, leading to reduced team cohesion [21]. It is observed that challenges in agile development and distributed development have been explored, but there has not been any comprehensive work on risks and challenges when these approaches are combined in practice [21]. Majority of research addressing the challenges in DAD are industrial reports [23,32,47], which is indicative of further need for investigation in this area.

This work is an extension of the exploratory research done by the authors on identification and classification of DAD risks and corresponding risk management methods in practice [40]. The exploratory study led to the formation of a list of risk factors that impact DAD project. A risk factor can be defined as a condition that can presents a serious threat to the successful completion of a software development project [41].

The classification of risk factors revealed five core risk categories, namely, ‘Software Development Life Cycle (SDLC)’, ‘Project Management’, ‘Group Awareness’, ‘External Stakeholder Collaboration’ and ‘Technology Setup’, which were then mapped to the components of the Leavitt's model of organizational change [28] The purpose of the mapping was to facilitate the identification of organizational aspects, which must be addressed for managing risks in DAD projects. The exploratory study highlighted the role of the intrinsic contradiction between the agile principles and the characteristics of distributed development in inducing risks in DAD projects. The exploratory work established the building blocks of risk management framework in DAD. These primitives have been deemed to be apt for providing deeper insights necessary for managing risks at the task/activity level by the practicing agile teams. However, there is a need for a systematic and structured approach for identification of the sources of risks and the methods used for managing them to facilitate DAD projects. The comprehensive framework must be able to manage risks in DAD projects at operational level. It is also expected to further the knowledge of the field for deriving new avenues for research and building dependable theories from experience. This study is the first step in this regard.

For the development of the framework, practitioners’ perception of the impact of the risk factors has been gathered using a self-administered questionnaire. The respondents had an experience of working at different levels in DAD project teams. At first, the practitioners ranked the risk categories for their impact on DAD projects. Then within each category, they ranked the risk areas and then, they ranked the risk factors within each risk area. The ranking revealed the most problematic aspects of DAD. Further, practitioners rated the extent to which they use specific techniques for managing risks in terms of the frequency of use. Based on the ranking and rating data, high impact risks and frequently used risk management techniques have been found. Further deliberation on the findings uncovered DSD properties that play significant role in inducing specific risks in DAD projects. Based on these findings, a DAD project risk management framework has been developed and subsequently, partially implemented in three different DAD projects. Substantial benefits have been achieved by the test DAD projects that validate the usefulness of the framework.

Following are the contributions of this study to the existing knowledgebase of the field of DAD:

  • 1.

    It has identified risk categories, risk areas under each risk category and risk factors under each risk area, which are perceived to have high impact on DAD projects.

  • 2.

    It provides the risk management methods frequently used in practice for managing each risk factor in DAD projects.

  • 3.

    Original exploration unveiled the DSD properties that conflict with agile principles and causes DAD specific risks. This extension study further established the role of those DSD properties with respect to the impact of DAD risks on projects and thus, equips the industry practitioners with the knowledge of the important aspects that must be addressed when certain DSD properties are more dominant than others.

  • 4.

    The study resulted into the development of a risk management framework for DAD projects, which was validated by implementing a part of it on real world DAD projects. The results obtained after implementing the framework showed that it helped the team to perform effective risk management and hence, was useful.

  • 5.

    Earlier research literature on the risks in DAD projects is either the studies based on literature reviews [8], [22] or the case studies [33,33,44]. The findings of these works were limited and did not provide a roadmap for operationalisation in the real world. This research provides a comprehensive framework for managing risks in DAD projects that can be implemented by the practitioners.

In this paper, Section 2 describes the background of the research that shaped the research objectives and the methodology (Section 3). Section 4 is on Data Analysis and Results. Section 5 holds a discussion on results and the risk management framework. Findings of the validation study have been reported in Section 6. Section 7 describes the Limitation and Threats to Validity of research. Finally, Section 8 holds the conclusion and future scope for the study.

Section snippets

Research background

Modern software development organizations prefer agile approach considering the benefits they achieve by offering better quality products quickly in market with improved customer satisfaction [32]. Due to the physical separation of the stakeholders in DSD, it becomes difficult to apply agile principles and practices effectively. There are studies which state that agile methods can be customized to suit to the needs of distributed development [10,42]. However, there is dearth of scientific

Research objectives and methodology

The objective of this extension of the exploratory work has been the creation of a DAD risk management framework that comprises risk factors, which are perceived to have high impact on DAD projects and the risk management methods, which are frequently used to control the identified risks in practice. The framework also relates the DSD properties such as spatial distance, temporal distance, work/development culture, language barrier and large project scope with the risks that have severe impact

Data analysis and results

The data analysis and results of ranking data and rating data are given in Sections 4.1 and 4.2, respectively.

Discussion

As distributed software development (DSD) and agile methods differ in their key tenets, significant risks arise when these two methods are combined for software development. Our exploratory work unveiled the properties of DSD which contrast with the principles of agile and hence cause risks [40]. We could relate each risk factor identified in DAD projects to one or more DSD properties, namely, spatial distance, work/development culture, large project scope, temporal distance, and language

Validation study of the risk management framework

In order to check the applicability and the feasibility of the risk management framework, we conducted a validation study. We provided the partial framework to three different companies for implementation in their real time DAD projects. The framework implementation benefited the companies as they were able to identify the critical risks in the projects and use the suggested risk management methods for controlling the same. The details of the methodology used, data analysis and results obtained

Limitations and threats to validity

The authors emphasize this work's significant contribution to the field through the risk management framework for DAD projects. However, there have been some unavoidable limitations, which must be stated for setting the relevance and significance of the framework. The large number of risk factors not only caused phased collection of data but also constrained online data collection as the respondents were not keen to fill the large questionnaire online. Majority of responses were collected

Conclusions and future scope

This study is an extension of an exploratory work on identification of the risk factors, which impact distributed agile development projects and risk management methods that are used in practice to control the risks [40]. The risk factors were grouped to form risk areas and high level risk categories. Risk categories includes, ‘Group Awareness’, ‘Software Development Life Cycle’, ‘Project Management’, ‘External Stakeholder Collaboration’, ‘Technology Setup’ [40] These risk categories were then

Acknowledgments

We would like to thank all the participants from the corporate who participated in the interview session and gave their valuable inputs for this research work. This work has been funded by Project Management Institute Educational Foundation, PMIEF (USA).

References (53)

  • B.G. Glaser

    The constant comparative method of qualitative analysis

    (1965)
  • J.D. Herbsleb et al.

    Global software development

    IEEE Softw.

    (2001)
  • M. Paasivaara et al.

    Using scrum in distributed agile development: a multiple case study

  • S. Ambler et al.

    Disciplined agile development

    A Practitioner's guide to Agile Delivery in the Enterprise

    (2012)
  • E Babinet et al.

    Dependency Management in a Large Agile Environment

  • B. Boehm et al.

    Management challenges to implement agile processes in traditional development organizations

    IEEE Softw.

    (2005)
  • M. Cohn

    Succeeding with Agile: Software Development Using Scrum

    (2010)
  • D.R. Cooper et al.

    Business Research Methods

    (2006)
  • G.M. Cottmeyer

    The goods and bad of Agile offshore development

  • M.A. Cusumano

    Managing software development in globally distributed teams

    Commun. ACM

    (2008)
  • F.Q.B. da Silva et al.

    Challenges and solutions in distributed software development project management: A systematic literature review

  • S. Dorairaj et al.

    Effective communication in distributed Agile software development teams

    Agile Processes in Software Engineering and Extreme Programming

    (2011)
  • B. Drummond et al.

    Yahoo! distributed agile: notes from the world over

  • T. Dybå et al.

    Empirical studies of agile software development: a systematic review

    Inform. Softw. Technol.

    (2008)
  • J.D. Gibbons
    (1993)
  • D.C. Gumm

    Distribution dimensions in software development projects: a taxonomy

    IEEE Softw.

    (2006)
  • J.M. Erickson et al.

    Risk factors in distributed projects

  • J.R. Evaristo et al.

    A dimensional analysis of geographically distributed project teams: a case study

    J. Eng. Tech. Manage.

    (2004)
  • R. Hoda et al.

    Self-organizing roles on agile software development teams

    IEEE Trans. Softw. Eng.

    (2013)
  • H. Holmström et al.

    Agile practices reduce distance in global software development

    Inform. Syst. Management

    (2006)
  • E. Hossain et al.

    Risk identification and mitigation processes for using scrum in global software development: a conceptual framework

  • S. Jalali et al.

    Agile practices in global software engineering-A systematic map

  • M. Jiménez et al.

    Challenges and improvements in distributed software development: a systematic review

    Adv. Softw. Eng.

    (2009)
  • M.K Mattsson et al.

    Classes of distributed Agile development problems

  • J.A. Maxwell

    Understanding and validity in qualitative research

    Harvard Educ. Rev.

    (1992)
  • P.H. Kvam et al.

    Nonparametric Statistics with Applications to Science and Engineering

    (2007)
  • Cited by (74)

    • Towards a Recommender System-based Process for Managing Risks in Scrum Projects

      2023, Proceedings of the ACM Symposium on Applied Computing
    View all citing articles on Scopus
    View full text