A case study research on Scrum Framework

IN THE FIELD OF TECHNOLOGYDEGREE PROJECT TECHNOLOGY AND LEARNINGAND THE MAIN FIELD OF STUDYCOMPUTER SCIENCE AND ENGINEERING,SECOND CYCLE, 30 CREDITS

, STOCKHOLM SWEDEN 2019

A case study research on Scrum Framework

ELENI ZIKOPI

KTH ROYAL INSTITUTE OF TECHNOLOGYSCHOOL OF ELECTRICAL ENGINEERING AND COMPUTER SCIENCE

ELENI ZIKOPI MASTER THESIS

A case study research on Scrum FrameworkEleni Zikopi

School of Computer Science and EngineeringKTH, Royal Institute of Technology

[email protected]

I. ABSTRACT

Agile methods have been widely implemented in the indus-try while teaching these methods is a substantial part of Com-puter Science and Engineering program curricula. Evidence-driven data of agile practices, methods, and tools have beenextracted based on empirical studies with students, however,there is an important need for more anecdotal evidence toconfirm these findings. In order to fill this gap, this paperexplores the perceptions and the applicability of the ScrumFramework in a student research team in an industrial context.Empirical data have been gathered through interviews with thecase study participants combined with a survey. The analysisreveals that student experiences are mainly positive and thatthey can easily grasp the benefits of Scrum Framework. Thechallenges of implementing Scrum in such a setting mainlyconcern the balance between coach and self-organization,formulating accurate user stories when researching, finding themost suitable estimation method, as well as planning whenconducting research. The empirical findings may potentiallybe extrapolated in student scrum teams or even in industrialsettings. Finally, this research paper should encourage otherresearchers to investigate the adoption of Scrum in a studentsetting.

II. SAMMANFATTNING

Agila metoder har implementerats i stor utstrackning inombranschen, medan undervisning i dessa metoder ar en vasentligdel av utbildningarna for datavetenskap och teknik. Bevis-driven data om agila arbetssatt, metoder och verktyg har tagitsfram fran empiriska studier med studenter, men det finns ettviktigt behov av mer anekdotiska bevis for att bekrafta dessaresultat. For att fylla detta gap utforskar den har rapportenuppfattningarna och tillampligheten av ramverket Scrum iett studentforskningsteam i ett industriellt sammanhang. Em-piriska data har samlats genom intervjuer med deltagarna ifallstudien i kombination med ett frageformular. Analysenvisar att studentupplevelserna ar framst positiva och att manlatt kan forsta fordelarna med Scrum. Utmaningarna med attimplementera Scrum i en sadan miljo galler framst balansenmellan coaching och sjalvorganiserande, formulera exaktaanvandarhistorier nar man forskar, finner den mest lampligauppskattningsmetoden samt planerar forskningsrelaterat ar-bete. De empiriska fynden kan eventuellt extrapoleras i Scrum-team med studenter eller till och med i industriella miljoer.Slutligen bor detta forskningsarbete uppmuntra andra forskareatt undersoka inforandet av Scrum i ett studentkontext.

III. KEYWORDS

agile methodology; Scrum; empirical study; team perfor-mance;

IV. INTRODUCTION

In order to deliver value rapidly and to operate in anever-changing environment, certain methods, and solutionsthat improve efficiency and transparency are needed. Agilepractices have proven to be an efficient method to improvemanagement of changing customer demands and market needs[1]. Scrum is described as a framework within which peoplecan address complex and adaptive problems, while produc-tively and creatively delivering products of the highest possiblevalue [2]. Scrum is also one of the key frameworks of agilesoftware development. One of the main functions is settingsequential goals to be fulfilled in a prearranged period oftime [3]. Scrum framework was proposed by Jeff Sutherlandtogether with Ken Schwaber in 1993 to overcome estimationfaults, when requirements are not fully known, generated bythe Waterfall method [4].

Scrum framework is constructed to adapt for customer’sexpectations and therefore supports the development of high-quality products [5]. Some of the key benefits over conven-tional development frameworks are the short iteration cycles,small team sizes, strong communication channels as well asearly delivery which allows teams to receive near real-timefeedback [6]. Scrum teams simulate evolutionary, adaptiveand self-correcting systems in such a way that enables rapidadaptation to change in the business and technical domain[3], [7]. Additional features of Scrum are team autonomy,end-to-end responsibility, transparency and cross-functionalityincluding all the expertise required for every stage of thedeveloping process. On the contrary to traditional managementpractices which embrace control and predictability, Scrum ismotivated by uncertainty and creativity [3].

Although agile practices have been broadly and successfullyapplied in industry, due to their developmental nature, theystill ignite great interest for both academia and industry. Agilemethodology is a common working standard today on whichresearch has been extensively conducted [6], [8], [9]. Univer-sities are also starting to teach agile methodology to preparestudents how to work after academia [10]. While currentstudies of applying Scrum exist, the following case has neverbeen tried nor studied. This study specifically aims to identifythe main requirements, challenges and key success factors withregards to applying Scrum on a team of students working


together on a common problem statement. The purpose is toexamine whether this framework is suitable when implementedby people who have no previous experience in working agileor together before.

V. BACKGROUND

A. Related research: Implementation of Scrum in industrialand academic context

Agile methodology has not only been implemented insoftware engineering, but also in various types of productdevelopment, both in academic and industrial context [11]. Inthese cases several success factors have been identified withthe following being the most common ones [12]–[15]:• continuous feedback• visibility in the process• decrease in development cost• rapid adaptation to change• involvement of multiple roles and simultaneously execu-

tion of the work• training and learningAdditional to that, teams working with agile frameworks

have experienced increase in overall communication, knowl-edge sharing, enhancements in team culture as well as betteroverview of the stage of the project [16]. Communication isthe area where the most improvements are reported due to thedaily ceremonies [16]. Nevertheless, issues related to areas oforganization and documentation are mentioned [16].

Implementing agile practices is also usually experiencedpositively by employees. Laanti et al conducted an extensivesurvey-based research on the opinions of Nokia’s employeeson agile transformation, containing more than 1000 respon-dents in seven different countries, indicating that 60% arepositive in working in iterative structure [17].

Another study on continuous release at Ericsson describesrelated challenges and positive effects of implementing large-scale Scrum [8]. Among challenges were overcommitment byexternal pressure, managing maintenance work and balancingbetween efficiency, while the positive effects reported wereincrease in flexibility, decrease development lead time andhigher employee motivation [8].

Applicability of agile methodology in a student contextfor practical uses has been thoroughly assessed positively byMahnic in a case study using Scrum [10]. Nevertheless, across-program investigation of students’ perceptions of agilemethods showed that agile method is more difficult to beapplied in the academic environment than in the industrial onewhich usually stems from schedule issues [18]. However, thesame study indicates that students are very enthusiastic aboutcore agile practices and that there are no significant differencesin the perceptions of students of various levels of educationalprograms [18]. Another empirical study revealed positive per-ceptions about Scrum while the empirical evaluation based onsurveys confirmed the anecdotal evidence about the strengthsof Scrum reported in literature [13].

One of the main requirements to achieve high competencein a Scrum team is the compliance and adoption of the agilevalues: transcendence, autonomy and cross-functionality [3].

A team focused study conducted by Moe et al showed thatteams had difficulties managing themselves since all the teammembers were familiar with high individual autonomy, butnot team autonomy [19]. Additionally, lack of trust preventedthem from shared leadership and feedback, values essential inall agile practices [20].

Agile practices have primarily positive impact, even thoughchallenges when using them were revealed. Further researchto bridge the gap between theory and implementation isneeded, since most of agile method effectiveness available hasanecdotal foundation [20]. According to Dingsoyr et al, onparticular agile methods such as Scrum, the current state oftheory and research lies in its nascent phase [20]. Therefore,further empirically validation of agile applicability is vital[10].

B. Scrum OverviewAccording to the Scrum Guide, Scrum is founded on em-

pirical process control theory and constitutes a repetitive andincremental approach aiming at controlling risk and increasingpredictability [2]. The framework is grounded in three pillars:transparency, inspection and adaptation, meaning that the pro-cess is visible and understandable to those involved, artifactsare inspected towards the Sprint goal, while the process isflexibly and rapidly adjusted [2].

1) Scrum Terms:• Product Backlog: is a dynamic list of every known item

required for the final product. The items decided to bedeveloped in the current sprint called Sprint Backlog [2].

• Sprint: a consistent time frame of short period of time,usually 2-3 weeks, during which the team works towardsachieving a releasable ”done”, increment of the finalproduct [2].

• Estimate: provide an approximate calculation of softwaredevelopment size, effort, complexity and duration for aparticular agile project [21].

Story points: are units of measure expressing therelative size of a task in terms of complexity and effort[21].

Ideal Time: expresses the actual amount of timeinvested to complete a task, subtracted by the additionaleveryday activities [21].

2) Scrum Practices:Scrum evolves around four time-boxed events, providing theopportunity for inspection and adaption [2]:• Sprint Planning: during planning the team decides and

estimates upon the Sprint Backlog and goal. It is theactivity that initiates the sprint.

• Daily Scrum: is a 15-minute daily meeting in which theteam discuss the status of work done since the previousdaily Scrum and forecasts the work to be done until thenext one.

• Sprint Review: is held at the end of every iteration,during which the increments and product backlog areinspected and adjusted for the next sprint.

• Sprint Retrospective: is the event during which the teamexamines critically itself and creates a plan of actionstowards its enhancement.


3) Scrum Roles:A Scrum Team consist of the following roles [2]:• Product Owner: is the interface between the team and

the stakeholders of the product, accountable for main-taining the value of the product, as well as managing theproduct Backlog.

• Scrum Master: is a servant leader charged for managingthe practices, ensuring that everyone follows Scrum rulesas well as facilitating the events.

• Scrum Team: is consisted of professionals who worktowards delivering an increment of the product at the endof each iteration.

The Scrum team is characterized by self-organization andcross-functionality. These characteristics create the ground forthe team to organize itself, as well as consist of all the requiredcompetencies to execute the necessary work. This specificdesign empowers flexibility, ingenuity and productivity.

C. Purpose and Research Question

The purpose of this research is to explore the Scrumframework and in what way it promotes the team performanceand engagement of a student team compared with other nonagile projects or ways of working. The intent is to clarifywhether alternative practices could be implemented in thetraditional Scrum method when it is applied in a team thathas not previously been exposed nor tried Scrum. Finally, thegoal of this exploratory study is to point out these practicesfor the implementers of Scrum, in order to enhance team’sperformance in terms of productivity, cohesion, learning andintegration. Thus the main research question of this study isthe following:How does Scrum framework impacts the performance of astudent team compared with non agile ways of working?Additionally, the following sub-questions are investigated:

1) What are the applicable practices promoting team perfor-mance and motivation in the context of a student Scrumteam?

2) What alternatives can be implemented in order to renderScrum more suitable for this specific context?

D. Defining Team Performance

Team performance is a multidimensional notion. A visiblegoal can fuel a team but it can not guarantee success. Teamperformance can be evaluated from an end user’s customerperspective, but also from the individual’s interaction aspect aswell. Whether the team members enjoy working on a commongoal or not can define the productivity in the long-term [22].Therefore, aligning internal dynamics with external relationscould empower team performance. Teams are ruled over bythe principle of ”equifinality”, meaning that a team can reachits goals by a variety of means and under different conditions[22].

In order to evaluate the team’s effectiveness and how this isinfluenced by implementing Scrum, it is required to formulatethe performance criteria. According to Team PerformanceAnalysis suggested by Thomson (2013) [22], performancecould be measured on the following criteria:

1) Productivity depends on the existence of a clear goalas well as on the team’s ability to adapt to change.However, this criterion is also determined on whether theoutput resulted from team’s work meets the demands ofthe stakeholders and on their performance standards (e.g.quality, quantity, innovation).

2) Cohesion reflects the nature of team members relation-ship. An effectively functioning team is more possibleto be maintained and strengthened for sustainable futurework.

3) Learning indicates the opportunities a team provides to itsmembers to grow and develop. That does not imply thatthe team should serve the individual needs, rather thatsuccessful teams intrigue and challenge their members.

4) Integration disclose team’s ability to be aligned withthe interest and goals of the organization that they areinvolved.

E. Linking Group Dynamic theories with Scrum

Group development theories claim that groups developin a customary pattern. Group patterns serve as indicatorsover time for developmental changes. Various models ofgroup development suggest different developmental stagesthat a group undergoes, however some commonalities can bedistinguished. Researchers have identified five stages of groupdevelopment. According to Wheelan these are the following:dependency and inclusion, counterdependency and fight, trustand secure work and, eventually, productivity and termination[23].According to empirical studies, groups having difficultiesin overcoming the first two phases of group development[24]. In Scrum teams, the team pass through the first andsecond developmental stage of ”dependency and inclusion”as well as ”counter-dependency and fight” in a rapid fashion.Scrum teams from their nature have the following significantcharacteristics: transparency, cross-functionality and self-organization. These characteristics can be found in thelater stages of the group development model. Therefore, afunctional scrum team has a bigger potential to reach stages”trust and secure” and ”work and productivity” faster, if theteam follows the agile principles and practices. Therefore,Scrum accelerates collaboration and helps to seize the team’sfull potential. Scrum practices, on the other hand increasecommunication and collaboration between the team membersas well as embrace openness and trust, required to resolveconflicts. Thus, Scrum teams have the potential to reach thestage of ”performing” faster. Further, the cross-functionalityand diversity of the scrum team provide all the requiredcompetences to execute the teams goal, but also to be morecreative and collaborating.According to Wheelan’s developmental stages, teams can notstay in the fourth phase of “Work and Productivity” eternally.The performance growth of the team is eventually slowingand declining towards a new steady state. Research confirmsthat this phase is depending on careful management and onthe smooth transition to the next steady phase. If the transitionis delayed, performance growth will be terminated [25]. That


happens when members forget to cultivate what has beenachieved and formulate it as a learning. Thus, it is importantto change tasks, roles and raise the expectations. Givingfeedback and managing the rules for how to collaborateare vital practices to continue develop. Scrum empowerscontinuous improvement through the practices of reviewand retrospective which provide a formal circumstance forinspection.Empirical researches on group development also showeddirect connection between the developmental stage and theleadership style [26]. For a team to be effective the properleadership style is needed to meet the requirements of thegroup. The performing stage of the group development goesalong with less directive leadership [23]. Scrum teams aredirected by the Scrum Master which acts as a facilitator or aservant leader to the teams needs.

F. How does the stage of the group influences team produc-tivity?

According to Wheelan there is a link between perceivedgroup maturity and performance, as well as a link of per-formance with the structure and organization of a group[24]. Wheelan (2005) found that in order for a group toreach the full potential which happens during the ”Work andProductivity” stage it takes approximately 6 months. Anotherimportant factor revealed from empirical studies indicate thatsmall groups are able to reach mature levels of group devel-opment faster than larger groups [27]. Factors such as culture,diversity, complexity of tasks, etc., play a significant role.Group development studies also show that the nature of therelationship between the team members plays a determiningpart in teams performance [27]. Tuckman’s ”developmentalmethod”, Carnall’s ”coping cycle” and ”Johari window” reflectthat increasing the communication, self-awareness and un-derstanding between individuals has positive effects in teamsperformance.

G. Scope of the StudyThe aim of this study is limited to the application of Scrum

Framework in this specific team. It focuses on what sortof effects this framework has on the team performance andwhat challenges emerge. The goal was not to find out ifScrum can have an impact in teams effectiveness, but whatkind of effects in this specific context. However, this researchwill only highlight different options in terms of combinedpractices trying to conclude on which of them are applicableon that specific case. Further, topics such comparing theteam’s performance with other teams within the organization,developing metrics and measurement methods for assessingScrum as well as assessing the software product itself are notincluded in the scope. This study will only cover this specificteam and not anyone else in the organization.

VI. METHODOLOGY

A. Case StudyFor this study a qualitative approach was chosen, through

a case study including observations, interviews and question-

naires. The case study approach was selected for analysingthe practices, ceremonies, dynamics and outcomes of applyingScrum in a real-life context. According to Yin [28], case studyis regarded as the appropriate approach for research questionsformulated in an exploratory way in order to understand thesetting and how systemic actions are linked to the results [29].The study focused on a student self-organized Scrum team,working on a common problem statement, located at the officeof Nordea Bank in Stockholm formed for the needs of thisstudy. The industry party in this case study, Nordea, is oneof the leading Nordic financial corporations, working withan agile approach. The case was to have a team of studentsworking together as a Scrum team researching in the areaof electronic identification (eID) mechanism. The team wasconsisted of a Scrum Master, a market analyst and a researchteam. The size of 5 members and the roles of the team werestable for the whole period. The case lasted from January toMay 2019 and the team ran 8 Sprints 2 weeks each. Theproduct backlog was prepared from the team members afterdiscussion with the Product Owner. Each user story containeda short description of the task, acceptance criteria and anestimate on either complexity or time.Following the Scrum framework, each sprint started with asprint planning and ended with a review and a retrospective.Every day the team held a daily Scrum meeting. When thetime estimation method was used, every individual decidedon their own the time to accomplish the assigned task, whilewhen complexity estimation was implemented the estimationwas decided by the team.The Product Owner was the supervisor of the team in thecompany, interlinking the team with the project needs. TheScrum master acted as a servant leader, facilitating the team’sprocesses by reducing the impediments found in the team’sway, promoting an agile culture, encouraging the improvementof the team and ensuring the workflow.The team was cross-functional with the students having var-ious academic backgrounds and research angles. The teamhas never worked together before, and team members rep-resent different knowledge domains since it was comprisedof the Scrum Master with background in management, themarket analyst with a combination of technical and businessbackground, and the research team with two individuals, withprogramming skills and software development background andone with more theoretical and mathematical background. Inaccordance with the principle of self-organization, the teamdecided on what they worked with, which platform and toolsto use, as well as how to proceed in the implementation.

B. Interviews

Interview as a data generation method is largely appliedin case studies [30] and it is a common technique to collectempirical data in qualitative studies [31]. In this certaincase, interviews are used both to collect impressions andperceptions of the case participants but also of external actorswho have implemented Scrum to track their experience andknowledge in order to establish a reference point accordingto which challenges, practices and improvement factors can


be identified. The interviewees were two individuals withacademic experience of Scrum and one with industrial. Semi-structured interviews were used, consisted of open-ended ques-tions around which the discussion was unfolded. The questionswere flexible in order to maintain the flow of the conversation.This type of interviews ignite interviewees’ ability to proceedin details and introduce issues relevant to the topic [30]. Theinterview guide started with an introduction of the case studymain focus and questions to get to know more about theparticipant and his or her previous experience of the Scrumframework. The main questions followed which were dividedin different focus areas such as practices, team performance,engagement, and challenges when using Scrum.The results are categorized into clusters of progress. Cate-gorization is used to assist analysing data interpretation byclustering findings under relevant categories. Interviews wereused in combination with observations. The observations madeby actively attending the setting and participating in the teamplanning, daily Scrum, retrospective and review sessions. Ac-tive participation in the research was vital in order to providewith alternatives and improvements of the Scrum framework.

C. Questionnaire

Team participants’ opinions were gathered through a ques-tionnaire as well. The questionnaire included 28 questions,some of which were open questions, some were statementsto confirm or reject and some were assessments on teamperformance criteria according to Thomson (2013) [22]. Thequestions dealt with the use of Scrum and its practices, butalso the team performance criteria which were introduced tothe participants based on the aforementioned definitions. Inthe question formulation, the baseline for comparison was notexplicitly defined. The questions were open for interpretationby the respondents assuming that they would compare withtheir prior experiences or any non-agile project.

VII. EMPIRICAL FINDINGS AND ANALYSIS

A. Observation of Scrum Team

Through being present on a daily basis during the fivemonths case study a nuanced comprehension of the team’sbehaviors, perceptions and positions were generated. Duringthe project a total of 8 sprints were examined, planningand retrospective meetings were attended and notes weretaken. During the case study, the following parameters ofprogress were observed: product backlog, Scrum practicesand self-organization principle.From an analysis of the Scrum practices it became evidentthat all of them were improved during time and practice.Scrum planning started as quite a time-consuming practice.The first Sprint lasted two and a half hours, while the last oneonly one hour and 15 minutes. This can be attributed to thefamiliarity with the process which in turn had an impact onthe duration and the confidence of the team, resulting in moreefficient meetings. More precisely, the team needed 3 sprintsin order to get familiar and optimize the processes accordingto its needs. Since this case was a trial and error study,the team tried several alternatives in terms of estimation

methods. The team tried both story points estimation and timeestimation method. However, the analysis of the estimationmethods opened some dilemmas regarding the applicabilityof Scrum in research teams.

1) Product Backlog:In the start of the project, the Product Owner preparedthe initial Product Backlog containing a list of prioritizedrequirements. Consequently, in the next sprints the teamstarted preparing and planning the sprint backlog. A recurringissue on planning sessions was the vague nature of theacceptance criteria and the definition of what is consideredas completed, deriving from the uncertainties and the natureof research itself. That was particularly evident in the first3 planning sessions, when the team was still trying to getfamiliarized with the Scrum framework. Thus, the practicewas both time-consuming and inefficient. In order to makethe planning meetings more efficient, individual meetingswith the Scrum Master and Product Owner were introduced.The goal of this meeting was to discuss the user storiesin depth, set new requirements for the project, but mainlyassisting in formulating the acceptance criteria of each taskand thus optimizing the planning meetings.

2) Scrum practices:Further, the results of the analysis of the planning sessionsdisplayed that the capacity in estimating how complex a task israpidly improves as well. The sprint burn-down charts, whichproject the rate of completion of tasks on each day of thesprint, indicated the team’s ability in estimating and plan-ning improved incrementally from iteration to iteration. Theprogress of estimating is displayed in the burn-down chartsfrom Sprint 1,4 and 6 shown in Figures 1,2,3 accordingly.The gray line represents the optimal rate of completion whilethe red reflects the actual one. The distance between thosetwo reflect the work in progress. Thus, it is quite evidentthat the rate of completion from Sprint 1 to sprint 6 wasradically improved. There are other factors that also impact therate of task completion such as task complexity and externalfactors, however the team’s ability in planning was apparentlyimproved.In the first 3 Sprints, the user stories concerned the pre-studyphase and literature review. During the sprints 4-6, the codinguser stories started, which were typical software developmentstories, with a combination of theory research when needed.In the final sprints, the stories concerned coding, the resultsof the studies and finalizing the reports.

Concerning which estimation method was more applicable,that was one of the main challenges of implementing Scrumin an inexperienced cross-functional team. Time estimationmethod had more accurate results for concrete tasks withspecific requirements and tangible results, such as writingthe report, as it was something relatable to similar tasks.Furthermore, having such various backgrounds and expertisein different knowledge areas resulted in various perspectiveson the complexity of the tasks. On the other hand, estimatinghighly uncertain tasks, such as developing a model or


Fig. 1: Burndown Chart Sprint 1


prototype with vague acceptance criteria and no or fewrequirements, story points estimation was further accurate asit relayed in the simple explanation of complexity instead ofthe workload. Thus, both methods worked for the team but invarious tasks. Time estimation method was more accurate inconcrete tasks while complexity estimation was applicable inhighly complex tasks. Nevertheless, estimates accuracy wasinfluenced by external factors, e.g. deadlines from schools,the nature of the tasks it self, change of research scope.Regarding Scrum practices applicability, every Scrum practicewas applicable to some extent, however the most valuableone was the planning as it facilitated the structure of theworkload instead of working ad hoc. The daily Scrum,further, was a pleasant experience as it brought insights inthe team members work, creating space for discussions andfeedback. It also raised the confidence and motivation to meetthe Sprint goal when the team members started updating eachother on their personal progress. Review and retrospectivemeetings were a good opportunity for the team to discuss onthe acceptance criteria and try to further refine the definitionof ”done”, as well as suggestions for potential refinements.These meetings also increased the communication flowsamong the team members and gave the chance to providefeedback. Retrospective and and planning brought valueto the overall team performance, by prioritizing the workand setting some standard ways of working which typicallyincreases the performance.

3) Self-organization principle:Considering the principles of self-organising, the team wascompletely responsible for the progress and the Product Ownerwas more like a mentor, monitoring the project. The team wasnot provided with a detailed requirements specification, butan overview of the project thus the team members organisedand managed their workload by themselves. However, thatcreated some confusion and uncertainty which was also


expressed during the planning sessions.Additionally, when the team realized their autonomy to adjustthe work to meet the Sprint goal, the practices and mainlythe daily Scrum changed from a structured to an unstructuredfashion. The team was able to self organize, identify whatkind of daily Scrum discussion was more suitable in order towork together as a team. They were also able to revise theplans towards reaching the Sprint goal.

B. Interviews and Questionnaire

The empirical data from the interviews combined witha questionnaire answered by the case participants revealedthe following results, clustered under the categories: com-munication, motivation and engagement, additional value andsatisfaction to work, scrum benefits, difficulties and challengesand team performance.

1) Communication:The Scrum practices had a positive impact in the team com-munication as the regular stand-up meetings provided withan overview of everyone’s workload, enabling for greatercommunication flow. Additionally, the daily meetings, togetherwith the sprint planning and retrospective facilitated the onboarding as a team and boosted the first two- dependency andinclusion, counterdependency and fight- developmental phaseswith faster pace. Nevertheless, the team turned out to be quiteseparated in their work, thus Scrum practices did not havegreat impact on the collaboration of the team members.

2) Motivation and Engagement:Scrum framework had either positive or neutral impact on themotivation of the team members. Some of them expressedthat they felt motivated due to the constant structure of theworkload into smaller tasks akin to a to-do list and by settinggoals for the upcoming weeks. However, some of themwere quite skeptical whether it was the Scrum practices thatmade them feel motivated or the team setting and the abilityto brainstorm together. Some of them also mentioned thatmoving a post-it as from work in progress to complete, hada positive psychological effect.“By expecting me to vote and participate in the planningsessions, it automatically made me engaged in what the otherteam members were doing. We could then share our thoughtsand ideas on each other’s work and help each other improvethe quality of the work.” -Case Participant 1


TABLE I: Questionnaire Results

Statement Positive NegativeThe product becomes a seriesof manageable chunks. 3 1

Progress is made, even whenrequirements are not stable. 4

Everything is visible to everyone. 3 1Team communication improves. 4The team shares successes alongthe way and at the end. 3 1

Stakeholders see on-time deliveryof increments. 3 1

A culture is created where everyoneexpects the project to succeed. 3 1

Scrum benefits stated by Rising L. and Janoff S. (2000)

3) Value and Satisfaction to work:Scrum benefits in the specific case were mostly aboutstructuring and organising the work, creating a do-to list,reflecting on what has been done contributing in constantimprovement of the work and bringing the team together.According to the case participants Scrum brings additionalvalue to the team work as it forces communication, facilitatessome processes due to the dependency on each other andempowers the information sharing between members. Theyalso reported additional satisfaction to work when it comesto checking out user stories and positive impact on theenjoyment, as they would describe the overall experiencewith Scrum affirmative. Further, the team participants wouldrecommend the framework in students teams working on acommon goal or problem statement, or to a research teamworking on the same research question/ hypothesis.“It let me create a to-do list, which helped me to knowwhat to do next when I’m finished with my tasks. It alsoput pressure on me to finish it before the end of the sprint.Another factor was the social part of it, it made us all learnthe new methodology, giving us another reason to stick withthe plans of each sprint.”-Case Participant 2

4) Scrum Benefits:The second part of the survey included seven statementsconcerning Scrum benefits stated by Rising and Janoff [32].The case participants were asked to either confirm or reject thestatements based on their experience. The results are displayedin Table I and they disclose an importantly positive level ofconfirmation with all statements.The case study participants strongly agreed that progressis made even when requirements are not stable, the teamcommunication channels improve, the framework assist inbreaking down complex tasks in smaller manageable ones,everything is visible to everyone, the team share successesalong the way, on time delivery of increments and developmentof a success culture.

Besides these statements Scrum benefits mentioned in theliterature study were also confirmed in the interviews withthe case study participants. Among them were the frequentteam communication to develop a shared understanding andpeer pressure, sharing knowledge, expertise and experience,transparency in the work management as well as the structurednature of the method ensuring that the team actual produces

value, are only some of them [12]–[15].5) Difficulties and Challenges:

The greatest challenges emerged in the case study wereformulating precise user stories, deciding on accurate estimatesand finding the most suitable estimating method, finding thebalance between self-organization and coaching as well asplanning for two weeks ahead while doing research.“In the research part, sometimes it has been holding me backbecause I need to research an area and then another topiccomes up which I need to investigate.”-Case Participant 3

As mentioned above the case participants were skepticalwhether the framework is suitable when one is conductingresearch. In research one might change track completely orrealise that the intended result is not feasible. Trying toinvestigate undiscovered areas may deviate one′s course aswell. Nevertheless, the framework is quite result oriented,since the intended goal of every user story is to produce andprovide increments of the final product. When researchingthough, some steps function as enablers to proceed in theactions that would actually provide value. For instance gainingknowledge or exploring new areas is an inevitable part of theprocess in order to implement something concrete afterwards.However, the team members had to work ad hoc at times, asit was required in order to execute the planned tasks.“I am not sure if it is the best method for research in general,since it seems that Scrum is very result based and it does nothave the right scientific ”mentality”.-Case Participant 4

The estimation part of the framework was the mostchallenging one. The under investigation area was quitevague and complex and resulted in great uncertainty whenestimating. It also turned to be quite challenging to formulateproper acceptance criteria and predict the duration or effortrequired to execute them. Another challenge that has broughtup when working with Scrum was the non-compliance withthe concept of “done”. There were user stories that werenot precise enough in describing the requirement details.Research is not a predictable or linear process, so predefiningthe detailed requirements in the first place was challengingor even impossible. Thus, there were user stories which werequite vague since the under examined research area wasunknown and others that the acceptance criteria were differentthan the outcomes. Nevertheless, difficulties stemming fromscheduling issues reported in literature were also evident inthis case, since the team had to face change the researchscope, different school deadlines which influenced the processof estimation and the accuracy of the estimates [18].“Research is pretty vague, so it is very hard to make properacceptance criteria for the user stories, which also makes itvery hard to estimate time and complexity.”-Case Participant4At that point it is essential to highlight the agile estimationmethods’ subjectivity. Estimations are based on individual’sintuition and hence may lead to creating a biased situation.At the same time, these methods may lead to errors in caseof an inexperienced agile team. Therefore, there is strongneed of an experienced person in the team. ImplementingScrum in a team without previous experience requires activeguidance and coaching. However, given the principle of


self-organization that Scrum framework demands, there isa need to find the proper balance between coaching andself-management. Additionally, time and experience arerequired until a team can grasp the full potential and benefitsof implementing Scrum.The interviews with the team members also raised somequestions regarding how active should be the presence of theProduct Owner. Despite that the research team developedsome solutions that were in accordance with the initial projectspecification, these individual solutions never came togetherto provide the final end-product, i.e. the proof of concept.Rather, different aspects of the under examined problemwere investigated and individual functions were produced.However, the team did not manage to integrate them in onepractical solution. This might have been solved with a moreactive participation and interaction with the Product Owneror with stricter project specification requirements. However,since the project was in the scope of different thesis, eachindividual had various requirements and deadlines fromexternal stakeholders and institutions. On the other hand,greater guidance from an external supervisor might hadsacrifice the principle of self-managing. Thus, it was leftup to the team members to be completely responsible anddecide on the progress, the direction and the methods forthe implementation of such a solution. Nevertheless, learningthrough experience and exploration was greater in such a way.

6) Team Performance:Finally, the team performance was evaluated based on four cri-teria suggested by Thomson (2013) [22]. The case participantswere given a definition of productivity, cohesion, learning andintegration, according to which were asked to answer whetherScrum framework had an impact and in which way.All the participants were affirmative that the framework hasa positive impact on the productivity. From an analysis oftheir answers resulted that Scrum improves productivity byfacilitating the process of creating a structured and clear two-weeks plan that brings value towards the final goal to a greatdegree. It also increases the productivity as tasks and goals getfurther concrete, as well as increases the chance of meeting thedemands of the stakeholders. However, to a certain degree theframework limits the ability of working ad hoc which studentsusually do work in this fashion. Nevertheless, there is someuncertainty whether Scrum increases the chances of meetingthe demands more than other frameworks do.Concerning cohesion the participants were also positive thatScrum practices increase the cohesion of the team. They statedthat Scrum practices have a substantial social aspect whichincrease the cohesion automatically as they force constantcommunication and information sharing contributing to ahealthy team culture and mutual understanding.Learning defined as the opportunities provided to the teammembers to grow and develop,was recognized as the most con-siderable aspect of Scrum. Some team participants expressedthat constant learning is the major benefit of Scrum, wherethe retrospective is a particular effective practice. The practiceenables open discussion, where everyone can contribute onthings to improve including oneself. Nevertheless,considering

that the case was a research project some participants wereskeptical that Scrum due to its result oriented nature mighthinder scientific learning since it forces to produce concreteresults.Finally, integration considered as the team’s ability to bealigned with the interest and goals of organization, was achallenging criterion for the certain case. Given that it wasa case study, the team members worked independently fromthe other teams in the organization. Thus, the frameworkdid not bring integration towards the interest and goals ofthe organization as the participants were executing the tasksneeded to fulfill the expectation of the institutions.

VIII. DISCUSSION

The purpose of this research was to explore in whichway Scrum framework impacts the performance of a studentteam. Additionally, the following sub questions were alsoinvestigated 1) what are the applicable practices promotingteam performance and motivation in the context of a studentScrum team and 2) what alternatives can be implemented inorder to render Scrum more suitable for this specific context.Results of the study confirmed both anecdotal evidence fromempirical studies and from the literature review about Scrumbenefits, however important aspects when implementingScrum in a student team were also highlighted. Analysisof the results revealed some important challenges whenimplementing Scrum in inexperienced people.Results of this empirical study showed that Scrum do impactthe performance of a student team in a positive way. Scrumpractices play an important role in the bonding of theteam, particularly in the first two developmental stages. Thedaily practices promote greater communication flow, betteroverview of the stage of the project while structuring acomplex project into smaller and concrete tasks increaseproductivity. Students reported positive perceptions on thescrum practices. Interviews with the case study participantsindicated that the students felt motivated and engaged withthe project due to the constant structure of the workload intosmaller tasks, while they also reported that scrum practicesbring additional value to work and have a positive impact onthe team culture.Scrum practices, on the other hand increase communicationand collaboration between the team members as well asembrace openness and trust, required to resolve conflictsand pass through the phase counterdependency and fight.Thus, Scrum teams have the potential to reach the stage of”performing”, according to developmental theories, faster.Further, the cross-functionality and diversity of the scrumteam provide all the required competences to execute theteams goal, but also to be more creative and collaborating.Discussing in person every day and using the whiteboardimproved the team’s amiability and openness while peopleseem more inclined to relay valuable information quickly,thus the trust among the team members increased faster.Additionally, the use of post-its rendered identifiable eachindividual’s task, while one’s contribution and performancewas also displayed. The presence of the whiteboard created


a healthy peer pressure and boosted the overall performanceand motivation of the team. ”Gamification” of the practices,such as moving post-its and voting on estimation, providedan enjoyable aspect that resulted in greater work satisfaction,motivation and thus better performance.However, some dilemmas regarding the applicability ofScrum in the specific context were revealed. Team’scommon understanding on the end-product and requirementspecifications is vital. It is necessary that the team hasfound the right level between abstract and detailed when itcomes to formulating the requirements. This fact intensifiesthe importance of the presence and active participation ofthe Product Owner in order to keep a consistent line whenformulating the requirement specifications and in their turnthe user stories.Analysis of the results of Sprint retrospective meetingsrevealed that the non-compliance with the concept ”done”,stems from uncertainty and the nature of research itself whichin turn results in decrease of motivation and cause uncertaintyin the planning practice.Cross-functionality of the team played a twofold role in theproject. On the one hand, it raised the confidence to achievethe team’s goal and the motivation of the team members sincethe team consisted of all the required competences to executea complex project. On the other hand the cross-functionalitycomplicated the estimation process. It was evident thatthe team members due to their various backgrounds andknowledge areas did not share the same perspective ontask-complexity and size of tasks.This is why trying time estimation method in the first sprintshelped the team to get familiarized with the process since it isa more concrete method and everyone could understand howto estimate time. Instead, in such a cross-functional team, noteveryone could understand how complex a task actually is.Nevertheless, there is always a degree of uncertainty sincein the first sprints the team could not accurately predict howlong would reading/ understanding and writing a paper couldtake.Solution to the above problems would be setting a cleargoal in the beginning, get an introductory course on howto implement scrum framework and its practices e.g. howto handle planning and estimation, participate in moreexperienced teams to see how they handle the practices.The results could be extrapolated in the following ways: fromthe standpoint of implementing Scrum in a newly formedteam in an industrial context, as well as from the introductionof students to the industry. Thus, the findings concerningthe case participants perceptions on Scrum are valid underthe assumption that Scrum is implemented according tothe Scrum guidelines and all the practices which affectsatisfaction and motivation are followed. Concerning thesuccess of the project, the roles of Scrum Master and ProductOwner have a substantial impact since they can decide onthe final outcome. Furthermore, the study showed the activepresence of an experienced team member to act as a coachcan also impact the success of the project in a positive way.In order to maximize the validity of the results, the case studywas implemented and integrated in an industrial environment.

The case participants worked on a real research problem andfollowed the Scrum guidelines as strict as possible. The casestudy participants were senior master students in their lastyear of studies, with some of them having working experiencein software engineering projects. Therefore, the results of thestudy could find applicability in the introduction of Scrumin a company, or in the implementation of Scrum to withlimited or no knowledge of Scrum.

IX. CONCLUSION

Empirical studies with students participants can supplyindustry and academia with evidence-driven results onagile processes and practices. This case study describedone more implementation of the Scrum Framework on astudent research team focusing on the assessment of theteams performance and motivation as well as the impact ofScrum practices on the former. The main conclusions of thisinvestigation are related to the research questions of the study.

A. RQ 1: How does Scrum framework impacts the perfor-mance of a student team compared with non agile ways ofworking?

Results of the study displayed that Scrum has positiveimpact on the performance of the team. The main benefits ofusing Scrum were observed on the communication, motivationand engagement with the project, satisfaction and additionalvalue to work. Nevertheless, the analysis of the result revealedcertain challenges when using the framework in such a setting.Insufficient communication and active presence of the ProductOwner, absence of coaching or mentoring from an experiencedindividual as well as difficulties in formulating and prioritizingrequirements when doing research can compromise the successof the project.However, the result-oriented nature of the framework createda flow in the productivity of the team. Although value wascreated and progress was made even when there was greatuncertainty in the research project, dilemmas about the effec-tiveness of the result oriented nature in the investigation andlearning have been aroused.The case participants confirmed with high rate of agreementthe following statements on Scrum benefits:

• The product becomes a series of manageable chunks.• Progress is made even when requirements are not stable.• Everything is visible to everyone.• Team communication is improved.• The team shares success along the way and at the end.• Stakeholders see on-time delivery of increments and a

culture is created where everyone expects the project tosucceed.

It should be noted that unanimity was reported in the commu-nication and constant progress aspects.


B. RQ 2: What are the applicable practices promoting teamperformance and motivation in the context of a student Scrumteam?

Scrum practices were in total applicable in such a context.Daily stand up meetings had a positive impact on the com-munication channels and increased the motivation to reachthe Sprint Goal. Planning and breaking complex tasks intomanageable chunks created flow in the team performance.Retrospective and Review provided the ideal opportunity toinspect and adapt in more suitable ways of working for theteam and gave the opportunity to reflect on the progressof the work. However, estimating the effort or time neededto complete the tasks was the most challenging part. Giventhe lack of experience and the vague nature of research,dilemmas aroused whether estimating the workload in researchis applicable as well as which estimation method provideswith better results. Based on the results, tasks with concreteresults such as coding, time estimation method is applicable.For complex tasks with few or no requirements, complexityestimation is more suitable.

C. RQ 3: What alternatives can be implemented in order torender Scrum more suitable for this specific context?

Taking everything into consideration, the case study con-firmed that students are able to reach the proposed benefitsfrom empirical studies and theory within a time-frame of atleast three sprints. However, in order for a team to reach highperformance stage six months are suggested. Experience isalso an important factor to reach high performance. Therefore,an experienced team member and the active participation ofthe Product Owner are vital to drive the project forward andgrasp the benefits of Scrum in such a setting. Nevertheless,balance between coaching and self-organization is importantto be found, so as not to sacrifice Scrum principle of self-management.

X. ACKNOWLEDGEMENTS

The support and confidence I received during the masterthesis study is highly appreciated. I would like to thankmy supervisors for the continuous support, feedback andinspiration they provided throughout the project.SupervisorsHannes Salin, supervisor NordeaFelix Albrecht, supervisor NordeaChris Rosenquist, supervisor KTH

I would also like to thank team Hodor who shared theirlearnings and thoughts and let me explore the role of ScrumMaster. Their engagement, motivation and energy inspired me.

Last but not least, a special thanks to the people that theyare always by my side, my family and friends, supporting meand believing in me in such an insatiable way.

XI. REFERENCES

REFERENCES

[1] A. Cockburn and J. Highsmith, “Agile software development: The peoplefactor,” Computer, vol. 34, pp. 131–133, 2001.

[2] K. Schwaber and J. Sutherland, “The Scrum guide,” 2017.

[3] J. Sutherland, The art of doing twice the work in half the time. NewYork: Crown Business, 2014.

[4] S. M. Wanjala and G. M. Muchiri, “A review of agile software effortestimation methods,” 2016.

[5] K. Beck, M. Beedle, A. van Bennekum, A. Cockburn, W. Cunningham,M. Fowler, J. Grenning, J. Highsmith, A. Hunt, R. Jeffries, J. Kern,B. Marick, R. C. Martin, S. Mellor, K. Schwaber, J. Sutherland, andD. Thomas, “Manifesto for agile software development,” 2001.

[6] C. Berger and U. Eklund, “Expectations and challenges from scalingagile in mechatronics-driven companies: A comparative case study,”in Proceedings of the Sixteenth International Conference, XP 2015,Helsinki, Finland, May 25–29, 2015. Switzerland: Springer Interna-tional Publishing, 2015, pp. 15–26.

[7] T. Dingsøyr, S. Nerur, V. Balijepally, and N. B. Moe, “A decade ofagile methodologies: Towards explaining agile software development,”Journal of Systems and Software, vol. 85, pp. 1213–1221, 2012.

[8] V. Heikkila, M. Paasivaara, C. Lassenius, and C. Engblom, “Continuousrelease planning in a large-scale scrum development organization atericsson,” vol. 149, 06 2013.

[9] T. Dingsøyr, T. Dyba, and P. Abrahamsson, “A preliminary roadmapfor empirical research on agile software development,” in Agile 2008Conference, Aug 2008, pp. 83–94.

[10] V. Mahnic, “Teaching scrum through team-project work: Students’perceptions and teacher’s observations,” International Journal of En-gineering Education, vol. 26, pp. 96–110, 01 2010.

[11] R. Levy, M. Short, and P. Measey, Agile Foundations: Principles,Practices and Frameworks. UK: BCS, 2015.

[12] G. Goncalves, R. Shigemura, P. D. B, R. Santana, E. Silva, A. Dakwat,F. Jose Miguel, L. Dias, P. Tasinaffo, and A. Cunha, “An agile developedinterdisciplinary approach for safety-critical embedded system,” 042017.

[13] V. Mahnic, “A case study on agile estimating and planning using scrum,”Elektronika ir Elektrotechnika, vol. 5, pp. 123–128, 05 2011.

[14] T. Myklebust, T. Stalhane, and N. Lyngby, “Application of an agiledevelopment process for en50128/railway conformant software,” 092015.

[15] S. C. Misra, V. Kumar, and U. Kumar, “Identifying some importantsuccess factors in adopting agile software development practices,”Journal of Systems and Software, vol. 82, no. 11, pp. 1869 – 1890,2009, sI: TAIC PART 2007 and MUTATION 2007. [Online]. Available:http://www.sciencedirect.com/science/article/pii/S016412120900123X

[16] K. Hiltunen, “The agile method scrum in development of safety crit-ical applications: A case study about challenges and opportunities fordevelopers and verifiers,” p. 60, 12 2018.

[17] M. Laanti, O. Salo, and P. Abrahamsson, “Agile methods rapidlyreplacing traditional methods at nokia: A survey of opinions on agiletransformation,” Inf. Softw. Technol., vol. 53, no. 3, pp. 276–290, Mar.2011. [Online]. Available: http://dx.doi.org/10.1016/j.infsof.2010.11.010

[18] G. Melnik and F. Maurer, “A cross-program investigation of students’perceptions of agile methods,” in Proceedings. 27th International Con-ference on Software Engineering, 2005. ICSE 2005., May 2005.

[19] N. B. Moe, T. Dingsøyr, and T. Dyba, “A teamwork modelfor understanding an agile team: A case study of a scrumproject,” Information and Software Technology, vol. 52, no. 5,pp. 480 – 491, 2010, tAIC-PART 2008. [Online]. Available:http://www.sciencedirect.com/science/article/pii/S0950584909002043

[20] T. Dyba and T. Dingsøyr, “Empirical studies of agile software devel-opment: A systematic review,” Information and software technology,vol. 50, no. 9-10, pp. 833–859, 2008.

[21] M. Cohn, Agile Estimating and Planning, ser. Robert C. Martinseries. Prentice Hall Professional Technical Reference, 2006. [Online].Available: https://books.google.dk/books?id=j0eFmAEACAAJ

[22] L. Thompson, Making the Team: A Guide for Managers, 5th ed. PearsonEducation, 2014.

[23] S. A. Wheelan, B. Davidson, and F. Tilin, “Group developmentacross time: Reality or illusion?” Small Group Research, vol. 34,no. 2, pp. 223–245, 2003. [Online]. Available: https://doi.org/10.1177/1046496403251608

[24] S. A. Wheelan, “Group process: A developmental perspective (2nd ed.),”Small Group Research, 2005.

[25] A. White, “From comfort zone to performance management,” 01 2009.[26] S. A. Wheelan, “Creating effective teams: A guide for members and

leaders.” 2005.[27] ——, “Group size, group development, and group productivity,” Small

Group Research, vol. 40, no. 2, pp. 247–262, 2009. [Online]. Available:https://doi.org/10.1177/1046496408328703


[28] R. K. Yin, Case study research design and methods, 4th ed.Thousand Oaks, Calif Sage Publications, 2009. [Online]. Available:http://openlibrary.org/books/OL16841559M

[29] L. McLeod, S. MacDonell, and B. Doolin, “Qualitative research on soft-ware development: A longitudinal case study methodology,” EmpiricalSoftware Engineering, vol. 16, pp. 430–459, 08 2011.

[30] B. J. Oates, Researching Information Systems and Computing. SagePublications Ltd., 2006.

[31] C. Seaman, “Seaman, c.: Qualitative methods in empirical studies ofsoftware engineering. ieee transactions on software engineering 25(4),557-572,” Software Engineering, IEEE Transactions on, vol. 25, pp. 557– 572, 08 1999.

[32] L. Rising and N. S. Janoff, “The scrum software development processfor small teams,” IEEE software, vol. 17, no. 4, pp. 26–32, 2000.

www.kth.se

TRITA -EECS-EX-2019:462

Documents

A case study research on Scrum Framework