Continuance in expertise-sharing networks: a social perspective

As engineering firms, R&D groups, and technical organizations recognize the centrality of their engineers' expertise to their performance, they are widely investing in knowledge management (KM) initiatives. Contemporary KM initiatives increasingly include expertise-sharing networks that help answer questions about who knows what. These systems allow organizations to locate and leverage the specialized engineering and technical expertise that is held in the minds of dispersed individuals. However, stories of such expertise-sharing networks that languish from under-use and abandonment abound and the issue of continuance has received very little attention in prior research. We explore this understudied issue. We develop a model of expertise-sharing network system continuance through a four-year observational study of 418 users of two such systems and then empirically test it using multiperiod data collected from 122 users of four such systems. The concept of irretrievable investments was used to guide theoretical development in the initial observational phase of the study. The study makes several unique theoretical contributions. First, it develops a model that illustrates how irretrievable post adoption investments (sunk costs) by individual users of expertise-network systems increase continuance. We empirically show that the model explains approximately half of the total variance in continuance intention. This model advances continuance beyond the traditional expectation-satisfaction model of initial adoption to more advanced post adoption stages of use and theoretically incorporates the network-specificity aspect of post adoption investments in explaining continuance. Specifically, we show that individual users': 1) reputation among peer users of a system increases continuance; 2) system-mediated relationships with other users of the system increase continuance; and 3) investments in personalization of a system initially diminish continuance. Another notable contribution is the development and validation of several new measures for expertise-sharing network constructs.