Why Powderless DTF Printer is T-shirt Printing Game Changer.pptx
Software Startup Engineering: A Systematic Mapping Study
1. Software Startup Engineering: A Systematic Mapping
Study
Anh Nguyen-Duc, Vebjørn Berg, Jørgen Birkeland,
Ilias Pappas, Letizia Jaccheri
EMSE Summer
school 2018
1
2. About myself
• 2011-2014, Ph.D from Norwegian University of
Science and Technology
– PhD title “Supporting coordination of software
development across organizational boundaries”
• 2015 – currently, entrepreneurs, startup
enthusiast, startup researcher
– Software Startup Labs
• 2017 – currently, Associate Professor at University
of South Eastern Norway
– Human Computer Interaction, Project driven
courses
4. Case - Muml
• Vision: To be the fastest provider of validated breaking news - by
capitalizing on live, super-local user-generated content + validation
technology
• Funding:
– 100k Eur from Innovation Norway,
– 50k Usd from Google News Initiatives
– 50k Nok self-funding
• Outsourced development team in Asia
• A fully functional product
4
6. Case - Muml
• Terminated operation after 2 years 7 months 18 days
• Reasons for failures
– Explainable factors
• Human factors
• Financial factors
• Process factors
• Technology factors
– Unexplainable factors
6
Certain failures can be avoid by learning
either from past experiences or systematic
and external knowledge
7. Startups vs. SMEs
Startup companies are unique:
Little or no operating history
Limited resources
Multiple influences
Dynamic technologies and markets
7
8. Motivations
• Startup researchers have called for a further attention
to engineering approaches
• Systematic mapping study helps to identify the current
status in the area and pave the way for more empirical
studies examining startups
8
N. Paternoster, C. Giardino, M. Unterkalmsteiner, T. Gorschek, P. Abra- hamsson, Software
development in startup companies: A systematic map- ping study, Information and Software Technology
56 (10) (2014)
E. Klotins, M. Unterkalmsteiner, T. Gorschek, Software Engineering Knowledge Areas in Startup
Companies: A Mapping Study, Vol. 210 of Lecture Notes in Business Information Processing, 2015, pp.
9. Research Questions
• RQ1: How has software startup research changed over time in
terms of focused knowledge areas?
• RQ2: What is the relative strength of the empirical evidence
reported?
• RQ3: In what context has software startup research been
conducted?
9
12. Results
RQ1: How has software startup
research changed over time in
terms of focused knowledge areas?
Focus areas are software process,
management, construction, design, and
requirements, with the shift of focus toward
process and management areas.
RQ2: What is the relative strength
of the empirical evidence reported?
The rigor of primary papers was higher
between 2013-2017 than that of 1994-2013.
RQ3: In what context has software
startup research been conducted?
Thematic concepts representing the software
startup context include innovation, lack of
resources, uncertainty, time-pressure, small
team, highly reactive, and rapidly evolving.
12
15. RQ1- SE Process
• The need for adapting software development process to startup context:
– Contextual factors: project’s scope, magnitude, complexity, and changing
requirements
• The lack of guidance how startups can establish their methods
• More studies to contribute to the adoption of agile practices in startups
15
16. RQ1- SE Professional Practices
• Startup processes need to cover both business and engineering aspects
• Startup developers need to acquire not only technical competence but also
business mindset
• Developers in software startups typically prioritize speed related agile practices
rather than quality related ones
• A possible research area is to investigate how universities can facilitate learning and
to support the specific needs of practitioners that are to work in software startups
16
17. RQ1- SE Management
• Relates software project management to business model experimentation and
customer development
• Startups struggle with how pivoting should be performed at diferent lifecycle
stages?
• Startup-aware outsourcing is a feasible option for early-stage startups
• A need for more research to identify how startups explicitly manage risks
17
18. RQ1- Software Requirements
• Requirements mainly were elicited through the founders’ assumptions and
interpretations of the market
• Minimum Viable Products (MVPs) are effective tools for requirements elicitation
• The lack of studies investigating how software startups perform requirements
engineering processes
18
19. RQ1- Software Design
• Requirements mainly were elicited through the founders’ assumptions and
interpretations of the market
• The lack of studies investigating how software startups perform requirements
engineering processes
19
20. RQ1- Software Quality
• Testing is critical to startups’ success, but often overlooked
• The most common testing techniques:
– unit tests (37%)
– pilot clients (25%)
– functional tests (25%)
– specialist testers (13%)
• Future research on how startups can learn from established companies’ systematic
testing processes
20
21. RQ1- Software Construction
• There does not exist a clear understanding of how entrepreneurs can use the
different tools efficiently to meet their specific needs
• A software tool portal can be helpful to support software construction
21
25. RQ3- Thematic Concepts, 1994-2017
Thematic Concepts
Frequency 13’-
17’ (#27)
Frequency 94’-
13’ (#47)
Innovation/Innovative 15 19
Uncertainty 14 15
Small team 11 12
Lack of resources 9 21
Little working/operating history 9 3
Time-pressure 7 17
Rapidly evolving 5 16
New company 5 8
Highly reactive 3 19
Highly risky 3 8
Third party dependency 2 12
One product 2 9
Not self-sustained 1 3
Low-experienced team 0 9
Flat organisation 0 5
25
26. Final remarks
• Software startups find it hard to apply theory in practice
• Future work on certain research themes, i.e startup evolution models,
human aspects, and consolidation of contextual factors
• Multivocal Literature Review can be the next review on software startup
26
Editor's Notes
60% of startups do not survive in the first five years
75% of venture capital funded startups fail
startup as a temporary organization that seeks a scalable, repeatable, and profitable business model, and therefore aims to grow
Lean manufacturing in toyota
Lean prototyping
Researchers have provided lessons learned and advice studies, paying less attention to specific tools and frameworks.
Two reasons for this are increased importance of startups, and increased focus on researchers providing high-quality research.
Startup literature provides an inconsistent use of thematic concepts describing startups.
From 1994-2013, the highest number of primary papers within a single year was 7 (2008).
In comparison, 2016 and 2017 constituted 9 and 11 papers respectively.
Between 1994-2013 “software design” and “software requirements” are the most represented knowledge areas,
However, software engineering process” and “software management” have received significant attention from the community between 2013-2017.
Lean manufacturing in toyota
Lean prototyping
Lean manufacturing in toyota
Lean prototyping
Lean manufacturing in toyota
Lean prototyping
Lean manufacturing in toyota
Lean prototyping
Lean manufacturing in toyota
Lean prototyping
Lean manufacturing in toyota
Lean prototyping
Lean manufacturing in toyota
Lean prototyping
Figure 7 shows the contribution types of primary papers between 1994-2017, separating the periods before and after 2013. The most frequently provided contribution types between 1994-2013 were advice and model, while lessons learned was most represented between 2013-2017.
The least frequently used ones combined from both studies were framework, guidelines, and tools.
Startup Start-up Very small entity Very small company Very small enterprise
It is usual that researchers specify the product orientation of the startups (e.g., B2B/B2C).
The number of startups under investigation is in the range from 1-20 startups. The most frequently used number of startups was found to be 3-5. The number of employees is usually in the range of 2-25, depending on the lifecycle stage of the company. The age of the investigated companies is usually in the range 1 month to 3 years.
Startups use different software development methods.
No more than two papers mentioned whether the investigated companies had received any funding