Poor soil fertility is regarded as the underlying factor limiting productivity in African agriculture. Substantial knowledge has been accumulated on different approaches to manage soil fertility in smallholder farms in Africa. Nevertheless the lack of adoption of various technologies, or the absence of widespread testing and experimentation by farmers, are often disappointing. We argue that this is related to the lack of integration of available knowledge and the lack of holistic and participatory approaches that foster technical and institutional change. Systems approaches, employing relatively simple summary type models across disciplines, can help disentangle the complexity of farmer decision making at farm scale. Such decision-making is often a compromise between the potential for short-term maximization of crop and livestock production and investment for sustainable production in the long term. An example is the NUANCES (Nutrient Use in Animal and Cropping systems – Efficiency and Scales) framework that assists with the analyses of different soil improving technologies in the context of farmers' strategies. It allows for spatial and temporal variability of resource use and access to land, labour, technologies and markets. Spatial patterns of soil fertility are often very pronounced in Africa, with farmers preferentially allocating manure, mineral fertilizers and labour to fields close to the homestead, resulting in strong soil fertility gradients at farm and village level. We show that clear scope for field-specific fertilizer management recommendations exists, provided they are based on local soil knowledge and diagnosis. Tools such as NUANCES are useful as they allow for ex-ante analyses and allow targeting technologies to specific types of farmers, and for identification of more appropriate technologies. They can for example help in determining optimum allocation strategies of mineral fertilizers across farms to enhance their efficiency. To promote innovations for increasing productivity through improved nutrient use in Africa will require holistic and dynamic approaches that foster both technical and institutional change, based on solid understanding of the farmer context. This involves the participatory development of soil management technologies with coordinated efforts to experiment and extend alternative institutional arrangements that link farmers with input-dealers, rural bankers and traders and strengthens the innovative capacities of the various stakeholders involved. A major challenge is to find pathways for action that recognize the heterogeneity of institutions and the diversity and conflicting agendas of the actors involved. Systemic inquiry by multi-disciplinary ‘teams', involving facilitators from different institutions, is needed. Such teams need to place more emphasis on farmer experimentation and adaptation according to the prevailing agro-ecological and socio-economic conditions rather than on technology prescriptions.