In recent years, there have been calls for a better understanding of postgraduate research supervision as a teaching and learning activity (e.g., Johnston, 1999). One of the strategies suggested to improve teaching is for practioners to engage in reflection (Schön, 1983). However, as an insider to a culture, reflective practice can be difficult because it is hard to uncover the hidden assumptions of practice and appreciate the power relationships within the curriculum (Brookfield, 1995). This difficulty was addressed in reflecting on my own supervisory practice by (a) deriving themes for reflection about supervision from an interview with an experienced supervisor using a narrative inquiry approach (Holloway & Jefferson, 2000), (b) undertaking a critical analysis of my practice in relation to these themes and the associated literature, and (c) reconciling differences in the experienced supervisor's perspective and my perspective. Five key themes were derived from an interview with an experienced supervisor. These were (a) the supervisor's conceptions of research and doctoral study, (b) timely postgraduate completion, (c) the distinguishing characteristics of doctoral students in education, (d) withdrawal and attrition rates, and (e) the development of supervisory practices. A comparison of my perspective on these supervisory themes with that of an experienced supervisor revealed substantial differences in our supervisory practices. For example, the experienced supervisor appeared to consider the doctorate as "a personal journey of discovery" (Pearson & Brew, 2002). In contrast, I hold a "trading conception" in which products of research, such as publications, grants and networks, are central to the research process (Pearson & Brew, 2002). From a trading conception, a thesis assumes a prominent role as the cumulative demonstration [performance] of what has been achieved during the doctoral candidature. This conception of research is consistent with a view of the curriculum as an enculturation process (Joseph, Bravmann, Windschitl, Mikel, & Green, 2000) and a systems approach to curriculum in which various knowledge components are required for the final performance (Toohey, 1999). Additional differences between the experienced supervisor's perspective and my perspective repeatedly indicated his foci on process and my focus on product. Notwithstanding the differences in these perspectives on supervision, they are compatible from a knowledge-management approach. Zhao (2003) argues that based on a knowledge-management approach research supervision is a process in which students are transformed into knowledge workers. This process commences with Inputs in the form of a research student and the research environment, which together undergo a Knowledge Conversion Process (e.g., knowledge creation) that ultimately culminates in Outputs including a researcher (i.e., the postgraduate student), the completion of a research degree, and research products. Based on this model, the experienced supervisor and I appear to be focusing on different parts of knowledge management in supervision. Whereas, the experienced supervisor is focusing on the knowledge conversion process, I am focusing on the outputs. Although both foci are integral in Zhao's model, these differing perspectives have the potential to generate conflict if the experienced supervisor and I worked on the same supervisory team without an awareness of the other's stance. However, this potential conflict can be avoided by varying the prominence of each perspective at appropriate times during the student's candidature. The reflective approach adopted has provided insight into the origins of differences in supervisors' perspectives. Additionally, it has raised my awareness of ways to enhance the quality of research supervision as a teaching and learning activity within the knowledge economy.