A new approach for better assessing consumer accept ability of improved traditional cassava products has been developed. It involves several successive steps. Qualitative surveys were conducted all along the food chain to have a good knowledge of the variability of processes, varieties, and products, and identify quality criteria of a good cassava product with the various stakeholders. Active and effective participation of women processors was essential in the adoption of new processes or varieties. Their observations and comments on cassava behaviour throughout the processing were used to identify the ability of new varieties or processes to make a good product. Physicochemical analyses of products complemented processors' selection by identifying quality traits. Focus groups and sensory evaluation were used to generate sensory descriptors and classify cassava products in separate groups according to their sensory properties. The acceptability and preferences of improved products by a large number of consumers were assessed using a "all-in-one" method coupling hedonic test, JAR "Just All Right" test, CATA "Check All That Apply" question and "willingness to pay”. Three case studies on bobolo or “bâton de manioc” and gari in Cameroon, and on gari in Benin will be given as examples and detailed. The methodological approach was different for each case, depending on the product, the local conditions, and the objectives of the study. The originality of the approach was the participatory involvement of a group of women processors in the selection, at each step of the processing, of the improved varieties with appropriate characteristics for traditional uses. This might be very useful for providing information to breeders early in varietal improvement programmes. (Texte intégral)