The increasing volume of waste, and the low remanufacturing rates for End-of-Life (EoL) products contrast with the goal of a sustainable circular economy. The essential role in enabling the reintegration of subassemblies, components, and materials into the production cycle is the disassembly. Efficient and effective disassembly is therefore a key for achieving a high level of circularity, but also a significant obstacle due to its complexity. This paper presents a study of the manual disassembly processes conducted in a learning factory, focusing on the disassembly depth (complete vs. partial disassembly) and organisational form (disassembly line vs. single workstation). The objective is to analyse disassembly processes, to identify learning effects, and to derive recommendations for designing efficient disassembly systems. Furthermore, the study highlights how a flexible and adaptable learning factory infrastructure can support research on disassembly system design. To achieve these goals, over 180 test runs were performed on a product within the learning factory under various disassembly conditions. The study evaluates the impact of organisational form and disassembly depth on processing times. The findings are used to develop recommendations for designing disassembly processes and systems.