Abstract
Context
Previous research in the Technical Debt (TD) field has mainly focused on the technical and economic aspects, while its human aspect has received minimal attention.
Objective
This paper aims to understand how software developers’ morale is influenced by TD and how their morale is influenced by TD management activities. Furthermore, this study correlates the morale with the amount of wastage of time due to TD.
Method
Firstly, we conducted 15 interviews with professionals, and, secondly, these data were complemented with a survey. Thirdly, we collected 473 data points from 43 developers reporting their amount of wasted time. The collected data were analyzed using both quantitative and qualitative techniques, including thematic and statistical analysis.
Results
Our results show that the occurrence of TD is associated with a lack of progress and waste of time. This might have a negative influence on developers’ morale. Further, management of TD seems to have a positive influence on developers’ morale.
Conclusions
The results highlight the effects TD has on practitioners’ software work. This study presents results indicating that software suffering from TD reduces developers’ morale and thereby also their productivity. However, our results also indicate that TD management increases developers’ morale and developer productivity.
Introduction
In recent years, there has been an increasing interest in technical debt (TD) within the software engineering discipline. The majority of previous studies focused on investigating the technical, financial, and organizational aspects of TD (Ampatzoglou et al., 2016), (Li et al., 2015), (Besker et al., 2018). However, until now, the literature has paid minimal attention to the human, and social aspects of TD, including the role of developers in the occurrence of TD as well as its consequences for them. More recently, several studies have suggested that TD has a negative influence on developers’ emotions and affects (Yli-Huumo et al., 2014), (Codabux and Williams, 2013), (Spinola et al., 2013) and their morale (Tom et al., 2013), (Fernández-Sánchez et al., 2015), where morale can be explained as a multidimensional concept that “subsumes confidence, optimism, enthusiasm, and loyalty as well as a sense of common purpose” (Peterson et al., 2008). However, none of these previous studies specifically focuses on empirically investigating and explaining the relationship between TD, the developers’ morale, and software developer productivity. Since no known research has focused on exploring these relationships empirically, this study seeks to obtain data from practitioners who experience TD on a daily basis, in order to expand the empirical findings in this area.
In a study by Tom et al., (2013), the authors suggest that morale, alongside quality, productivity, and project risk, are the four main areas that are negatively influenced by the occurrence of TD. The authors claim that, since the occurrence of TD reduces software quality, developers must spend more time and effort to address quality issues in the future. This, in turn, will decrease developers’ productivity and maintenance costs in the long term (Tom et al., 2013), (Fernández-Sánchez et al., 2015). On the other hand, some studies suggest that developers do not feel comfortable about taking on TD (Yli-Huumo et al., 2014), (Codabux and Williams, 2013), and it lowers their motivation (Spinola et al., 2013).
Software development is a sociotechnical phenomenon (McLeod and Doolin, 2012), and therefore, its success depends on both its social and technical aspects (Feldt et al., 2010). The ways in which today's software developers work require a comprehensive understanding of their feelings, perceptions, motivations, and identification with their tasks in their respective project environments (Fagerholm et al., 2015). Recent studies have shown that positive affective states, such as happiness, satisfaction, and motivation, increase software developers’ productivity and software quality (Feldt et al., 2010), (Graziotin and Abrahamsson, 2014). On the other hand, based on the results of previous studies, mainly from management science (Abbott, 2003), (Stowe, 2009) and more recently in software engineering (Damian and Chisan, 2006), (Fairley and Willshire, 2005), (Foulds and West, 2007), (Hall et al., 2007), developers’ morale and their productivity seem to correlate. However, to the best of our knowledge, there is a lack of studies focusing on investigating the relationship between TD and morale and productivity using empirical data.
Considering the large number of previous studies which argue that software firms take on TD to boost their productivity and obtain added business value (Tom et al., 2013), it becomes apparent that minimal attention has been paid to the human and social aspects of TD (Becker et al., 2017), and that there is a need for empirical studies (Cunningham, 1992) to explore and explain the effects of TD on developers’ morale and to correlate it with developer productivity.
The goal of this study is, therefore, to understand how software developers’ morale is influenced by TD present in the software they are developing and also to understand how their morale is influenced by TD management activities. Furthermore, in order to understand if their morale affects their work productivity, this study explores associations between morale with the amount of wastage of working time due to experiencing TD.
Note that the occurrence of TD and its produced time waste are two separate concepts: companies can have TD with a low “interest rate” (the term used to characterize the negative impact of TD). In such case, the presence of TD does not produce time waste but can still influence morale. On the other hand, TD can produce time waste, which may be the reason why TD affects morale. For these reasons, we study the concepts separately and we have two separate research questions (RQ1 and RQ3).
The term developer morale will be examined by surveying developers to indicate on a Likert scale their opinion about the impact of TD on three different dimensions of morale. The three different dimensions are: a) Affective antecedents such as focusing on developers’ moods, feelings, emotions, and attitudes, and b) Future/goal antecedents with a focus on the developers’ goals for the future and, finally, c) Interpersonal antecedents addressing the relationships and communication between developers. These dimensions are described in detail in Section 2.2.
Based on this goal, this study will examine the following research questions (RQ):
- RQ1: Does the occurrence of TD influence developers’ morale?
- RQ2: Does the management of TD influence developers’ morale?
- RQ3: Does wasted time (due to TD) correlate with morale?
- ○
- RQ3.2: Does wasted time correlate with TD Management dimensions of morale?
RQ3.1: Does wasted time correlate with TD occurrence dimensions of morale?
Our study has several novel contributions to software engineering research and practice. First, our study specifically concentrates on investigating the influence of TD on developers’ morale. Our findings are encouraging since they clarify the impacts of TD on different dimensions of morale and illustrates its relationship with developer productivity. Especially by indicating that developers consider TD and its management as important factors influencing progress and future development activities, this study encourages software firms to consider the human and organizational consequences of TD more seriously. Additionally, since previous studies have indicated the link between morale and developers’ productivity, our study investigates this relation empirically. In future research, this approach will enable enhanced exploration and measurement of software developers’ morale in different contexts.
An earlier paper at the 11th International Symposium on Empirical Engineering and Measurement (ESEM) (Ghanbari et al., 2017) presented part of the results reported in this study. However, this manuscript extends that previously published study significantly by triangulating the previous data collection and adding more data and, moreover, by adding more analysis. The novelty of this study's approach compared to the previous study lies in the morale focus of this paper, where we correlate the waste of time with three different dimensions of morale together with seven additional interviews and additional survey questions related to the morale perspective of TD.
The data representing the wasted time are based on reported data from developers who were asked to individually keep track of the time they wasted due to experiencing TD, on a daily basis. They were thereafter asked to themselves report this time to us, twice a week in a survey. Meaning that the data are based on the participants' own tracking and calculations.
The original study was exploratory, and based on our initial results, we proposed a set of propositions. In this study, using those propositions as a starting point, we conducted an additional cycle of research to investigate further the validity of those propositions and also to explain the logic behind the correlation of TD, morale, and productivity. In order to investigate if developers' morale affects their productivity, we have added one additional research question (RQ3), where we correlate the reported amount of working time wasted due to experiencing TD (as a proxy of productivity) with three different dimensions of morale and discuss them in detail. The related Research section has been extended to be broader and more carefully cover additional related research publications. We have extended the data collection by adding a longitudinal study collecting additional quantitative data in order to understand how much time developers report as being wasted due to experiencing TD over a longer period of time. We also augment our previous study with seven additional interviews. We believe that this additional context extends and strengthens the results derived in the original study.
This paper is structured as follows: Section 2 presents the theoretical background of the study. In Section 3, we describe the research methodology. Section 4 presents the research results. Sections 5 and 6 discuss the findings and threats to the validity of the study, respectively. Finally, Section 7 concludes the study.
Section snippets
Theoretical background
In this section, we discuss the background of the study in terms of TD, morale, the relationship between TD and morale, and the relationship between TD and productivity, and finally, the relationship between morale and productivity.
Methodology
Since our goal is to study TD and morale in its natural context, we have adopted a mixed-method approach where we use both qualitative and quantitative data collection and analysis methods, and part of the study also collects data longitudinally.
This study is based on data from 15 face-to-face interviews, and a survey, together with a longitudinal study, in order to examine the negative impact TD has on software developer morale and developer productivity. The study was conducted between
Results
The following subsections present the results for the research questions presented in Section 1. First, we present the results from the analysis of the qualitative data collection for each of the three identified dimensions of morale. Secondly, we present the results from the quantitative data collection.
Discussion
In this study, we aim to answer and discuss the earlier stated research questions. In response to RQ1, regarding the influence of TD on morale, it can be said that the occurrence of TD may reduce developers’ morale mainly because the developers perceive it to hinder their progress and makes it challenging to perform their tasks. In response to RQ2, it can be said that proper management of TD appears to increase developers’ morale since it is associated with positive personal and interpersonal
Limitations and threats to validity
As empirical research, this study is subject to different threats. We discuss these limitations according to four types of threats to validity suggested by Runeson and Höst (2009).
Concerning the construct validity (Runeson and Höst, 2009), there is a concern about using the right operational measures for studying the concept of morale. To mitigate this threat, we decided to follow an approach that relies on capturing different levels of morale by investigating its antecedent factors (Foulds and
Conclusions
The presented research aimed to examine the potential impact of technical debt and its management on developers’ morale and productivity. To the best of our knowledge, this study is the first study to assess these relationships empirically using a mixed-method approach.
The findings from this study make several contributions to the present TD research. First, the results from this study indicate that the occurrence of TD reduces developers’ morale since the presence of TD hinders the developers’
CRediT authorship contribution statement
Terese Besker: Conceptualization, Methodology, Validation, Formal analysis, Investigation, Data curation, Writing - original draft, Writing - review & editing, Visualization, Project administration. Hadi Ghanbari: Conceptualization, Methodology, Validation, Formal analysis, Investigation, Writing - original draft, Visualization. Antonio Martini: Conceptualization, Methodology, Validation, Formal analysis, Investigation, Writing - original draft, Visualization. Jan Bosch: Conceptualization,
Declaration of Competing Interests
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
We thank all the anonymous interviewees and survey respondents for their contribution to this work.
Terese Besker is a Ph.D. candidate in the Software Engineering at Chalmers University of Technology, Sweden. She is working in the research fields of technical debt management. Before becoming a Ph.D. student, she had worked as a senior software engineer in the software industry for more than fifteen years. She also has a bachelor’s degree in software engineering and a master’s degree in applied IT from Chalmers. She has published several peer-reviewed articles in journals, conference and
References (69)
- T. Besker et al.(2019)
- A. Ampatzoglou et al.(2015)
- D. Graziotin et al.(2018)
- H. Alahyari et al.(2019)
- G. Suryanarayana et al.
- B. Vogel-Heuser et al.(2017)
- J.M. Verner et al.(2014)
- S. Beecham et al.(2008)
- F. Fagerholm(2015)
- A. Martini et al.(2015)
Software developer productivity loss due to technical debt—A replication and extension study examining developers’ development work
J. Syst. Softw.
The financial aspect of managing technical debt: a systematic literature review
Inf. Softw. Technol.
What happens when software developers are (un)happy
J. Syst. Softw.
An exploratory study of waste in software development organizations using agile or lean approaches: a multiple case study at 14 organizations
Inf. Softw. Technol.
Chapter 1 - Technical Debt
Adapting the concept of technical debt to software of automated production systems focusing on fault handling, mode of operation and safety aspects
IFAC-Papers OnLine
Factors that motivate software engineering teams: a four country empirical study
J. Syst. Softw.
Motivation in software engineering: a systematic literature review
Inf Softw Technol
Performance alignment work: how software developers experience the continuous adaptation of team performance in lean and agile environments
Inf. Softw. Technol.
Investigating architectural technical debt accumulation and refactoring over time: a multiple-case study
Inf. Softw. Technol.
Cited by (24)
- Infinite technical debt2022, Journal of Systems and SoftwareCitation Excerpt :
- Preventing technical debt with the TAP framework for Technical Debt Aware Management2022, Information and Software TechnologyCitation Excerpt :
- PENTACET data - 23 Million Contextual Code Comments and 500,000 SATD comments2023, arXiv
- The Type to Take Out a Loan? A Study of Developer Personality and Technical Debt2023, arXiv
- Process Debt: Definition, Risks and Management2023, SSRN
- On technical debt in mathematical programming: An exploratory study2022, Mathematical Programming Computation
This pair enables stakeholders to be aware of unresolved TD and decide what to do about it. These are among the most time-consuming and least effective steps (Liu et al., 2020), as TD occurrences are associated with a lack of progress and waste of time, having a negative influence on developers’ morale (Besker et al., 2020). Moreover, developers tend to report the same TD, requiring a manual component in the identification and monitoring processes (Zazworka et al., 2013).
The TAP framework tackles the deadline problem by integrating TD management into project management, i.e., integrating the repayment of a project’s TD as part of the project itself. Regarding TD management, Besker et al. found that it “promotes a culture within organizations in which developers are supported and appreciated for managing their TD” and that it “leads to a virtuous cycle where the right culture and TD prevention mechanisms reinforce each other” [48]. This paper presents a comparative case study of a TD management framework developed by practitioners.
View all citing articles on Scopus
Recommended articles (6)
- Research articleInformation and Software Technology, Volume 70, 2016, pp. 130-139
- Research articleJournal of Systems and Software, Volume 167, 2020, Article 110593
- Research articleInformation and Software Technology, Volume 148, 2022, Article 106926
- Research articleJournal of Systems and Software, Volume 167, 2020, Article 110618