Abstract
In an environment where technological development largely determines economic growth and social change, predicting these developments is critical for strategic decisions in organizations. Previous work widely uses S-curves to model technological development in three phases: introduction, growth, and maturation. However, S-curves with a single growth phase (single-sigmoid) fail to model developments with multiple growth phases. Especially fast-moving technologies, such as computer-science technologies, can exhibit multiple growth spurts when new growth-inducing features are introduced. To fill this research gap, we build on previous seminal work exploring computer-science subcategories on arXiv. We hypothesize that the growth of subcategories on arXiv might be modeled more accurately by S-curves with multiple growth phases (multi-sigmoid). We show that nine of the 20 technology subcategories analyzed exhibit multi-sigmoid growth patterns. For instance, we find a significant second growth phase for the subcategories “Information Retrieval”, “Cryptography”, and “Systems and Controls”. Altogether, our results expand the TechMining literature by showing that S-curves with multiple growth phases (multi-sigmoid) describe.