The process of serendipity in knowledge work

While serendipity is generally considered a spark for innovation and new knowledge, the triggers for serendipity appear infinite and consequently information systems' support for serendipity has been difficult to realize. Research to date has tended to focus only on supplying users with unexpected triggers for serendipity (e.g., embedded links in results). We adapt a model of the serendipitous process that examines serendipity more holistically. Using previously collected data, we focus on understanding the precipitating conditions that must be present to facilitate serendipity. Results suggest that serendipity occurs during social networking and active learning, and more specifically in the act of exploratory search. Results also suggest that serendipity is not always instant -- the usefulness of triggers may not be immediately apparent and a period of incubation is sometimes necessary before recognition of the serendipitous nature of a latent trigger is attained. Implications for the design of information systems are explored and support for the incubation period is discussed.