The development of recovery processes using aqueous two-phase systems (ATPS) has been limited by the extensive experimental work required to establish the optimal system parameters to ensure selective partitioning of the product of interest. Although using full factorial experiments has been demonstrated to be an effective strategy for the characterization of the partitioning behaviour of biological products in ATPS, this approach is characterized by its costly and time consuming nature. As an alternative, the use of a robotic-aided strategy has been proposed. However, the need for high cost equipment may limit the generic implementation of this strategy. Based on practical experience using ATPS, practical strategies for the predictive design of primary recovery processes using polymer-salt systems were derived. To evaluate the generic application of the proposed strategies, four experimental models (B-phycoerythrin, C-phycocyanin, double layered rotavirus-like particles and lutein) were selected. The application of these strategies resulted in the development of simplified recovery processes for the selected experimental models. The practical review presented is considered a relevant contribution that will provide general rules to facilitate the establishment of ATPS processes, particularly for new researchers in the field. (c) 2007 Society of Chemical Industry