An Empirical Investigation on Software Practices in Growth Phase Startups.

Abstract

Context: Software startups are software-intensive early-stage companies with high growth rates. We notice little evidence in the literature concerning engineering practices when startups transition to the growth phase. Aim: Our goal is to evaluate how software startups embrace software engineering practices. Methodology: We conduct a survey guided by semi-structured interviews as an initial step, to be followed by field questionnaires as part of a future exploratory study. We use open coding to identify patterns leading to themes we use to state our hypotheses. To identify our samples, we use purposive sampling. Results: Specifically, we analyze seven startup cases during the first qualitative phase. We obtain five anti-patterns (no-documentation, no-agile, no-code intellectual property protection, cowboy programming, no-automated testing) and corresponding patterns (readable code, ad-hoc project management, private code repositories, paired and individual programming, ad-hoc testing) in adopting software engineering practices. We state 10 corresponding hypotheses we intend to corroborate by surveying a more significant number of software startups. Contribution: This study, throughout its recommendations, provides an initial road map for software startups in the growth phase, allowing future researchers and practitioners to make educated recommendations.