Data-driven Innovation : An exploration of outcomes and processes within federated networks

The emergence and pervasiveness of digital technologies are changing many aspects of our lives, including what and how we innovate. Industries and societies are competing to embrace this wave of digitalization by developing the right infrastructures and ecosystems for innovation. Similarly, innovation managers and entrepreneurs are using digital technologies to develop novel products, services, processes, business models, etc. One of the major consequences of digitalization is the massive amounts of machine-readable data generated through digital interactions. But this is not only a consequence, it is also a driver for other innovations to emerge. Employing analytical techniques on data to extract useful patterns and insights enables different aspects of innovation. During the last decade, scholars within digital innovation have started to explore this relationship between analytics and innovation, a phenomenon referred to as data-driven innovation (DDI). Most theories to date view analytics as variable that affects innovation in performative terms and treats it as a black-box. However, if the innovation managers and entrepreneurs are to manage and navigate DDI, and for the investors, funders and policymakers to take informed decisions, they need a better understanding of how DDI outcomes (i.e. market offerings such as products and services) are shaped and how they emerge from a process perspective. This dissertation explores this research gap by addressing two research questions: “What characterizes data-driven innovation outcomes?” and “How do data-driven innovations emerge in federated networks?” A federated network is a type of – increasingly common – contemporary innovation structure that is also enabled by digital technology. The dissertation is based on a compilation of five articles addressing these questions. The overall research approach follows a multiple case study design and the empirical investigation takes place in two case sites corresponding to two EU-funded projects. As a result, a classification taxonomy is developed for data-driven digital services. This taxonomy contributes to the conceptualization of DDI outcomes grounded on static and dynamic characteristics. In addition, a DDI process framework is proposed that highlights the importance of exploration, the temporal relationship between data acquisition and innovation development, and the various factors that influence the process along with examples of their contextual manifestations. Finally, social and cognitive interactions within federated networks of DDI are explored to reveal that the innovation teams rely on data-driven representations to facilitate various stakeholders’ engagement and contribution throughout the process. These representations eventually stabilize into boundary objects that retain the factual integrity of the data and analytical models but are also flexible for contextual interpretation and use. These findings contribute to the current discourse within digital innovation by introducing the lens of data analytics to conceptualize a specific type of digital artifacts, and well as providing a rich descriptive account of an extended digital innovation process. They also contribute to the discourse on data-driven innovation by providing an empirical account of DDI from a process viewpoint