Improving the Cost Structure of Sensemaking

In many everyday contexts people interact with information systems in order to make sense of a domain of interest. However, what this means and how it can best be supported are poorly understood. In particular, there has been little research on how to develop system representations that simplify naturally occurring sense making processes by matching people’s conceptualizations of the domain. In this paper we draw on Klein et al’s data-frame theory and Russell at al’s notion of cost-structures in sensemaking to propose an approach to understanding sensemaking that supports reasoning about system requirements. The two key elements of the approach are the identification of the process and the transformational steps within that process that could benefit from support to reduce costs, and the identification of primary concepts which are cued by information in the context of a given sensemaking task and domain, and around which users integrate information to form a structured understanding. Our general principle is that by understanding a sensemaking transformation in terms of its source data and the integrating structures it creates, one is better able to anticipate the evolving information needs that it tends to invoke. We test this approach with a case study of fraud investigation performed by a team of lawyers and forensic accountants and consider how to support the elaboration of prototypical user-frames once they have been invoked.