In this paper we seek to empirically study the use of location-awareness of others in the context of mobile collaboration. We report on a field experiment carried out using a pervasive game we developed called CatchBob!. Using both quantitative and qualitative data, we show the underwhelming effects of automating location-awareness. Our results indeed shows that automating this process does not necessarily improve the task performance and that it can be detrimental to socio-cognitive processes involved in collaboration such as communication or the modeling of partners' intents. The paper concludes with some potential impacts for location-based application practitioners.