Improving Software Development in Early-Stage Startups

Abstract

Startup companies are important drivers of economic growth globally. Over the last two decades, software startups have become a part of mainstream culture, and have, in the process, become associated with innovativeness and various success stories. Many of the current and up-and-coming tech giants, the so-called unicorns with a valuation of over one billion USD, are examples of these startup success stories, some more well-known than others. However, past this illustrious image, the vast majority of startups fail, and in up to 98 % of new business ideas in general fail. Software startups operate in a unique context often characterized by disadvantage that stems from various factors that vary by startup. This unique nature of the software startup context presents issues when it comes to applying existing knowledge of Software Engineering (SE) (or Information Systems Development (ISD)) into the startup context. Various research findings, existing SE/ISD methods, and lessons learned from practice come from more established software organizations such as multinational corporations. For example, Agile methods are more equipped to tell an organization ‘how’ to develop software in a situation where the needs of the customer are well understood. On the other hand, startups often operate in a situation where it is also unclear ‘what’ should be developed and there is no clear customer in sight yet. This dissertation focuses on better understanding the software startup context in SE, with a focus on how software startups develop software. To this end, the dissertation ultimately proposes a method for early-stage software startups. The dissertation comprises five academic articles, out of which three are conference publications and two are journal publications. The articles utilize qualitative methods to approach the different issues in each article. The results of the dissertation further our understanding of how software startups work, and the method presented in the fifth and final article of the dissertation will ideally help early-stage startups work more systematically.

Keywords: startup, software startup, software engineering, software development method, software development practice, decision-making, the essence theory of software engineering