Rice is a strategic crop for poverty reduction and food security in Sub-Saharan Africa (SSA). Improving rice productivity in SSA requires increased use of inorganic fertilizers. Low adoption rates and low application intensity are explained by low and heterogeneous profitability of fertilizer application. Soil properties influence crop yield response to fertilizer and hence profitability. Soil properties are known to vary even within a small geographical area, making it difficult for farmers to learn from other farmers’ experiences. Therefore, providing site-specific information on soil properties is expected to encourage farmers to use fertilizer by indicating the appropriate rate to apply and reducing uncertainty about the returns to fertilizer investment. In the relevant literature, many studies have shown the positive effects of providing site-specific information on farmers’ fertilizer use, but most studies deal with multiple soil properties to generate the information, which results in high financial costs of implementation, reducing the economic viability. In this study, a randomized controlled trial was conducted in the central highlands of Madagascar. The objective was to examine whether binary information on the expected effectiveness of nitrogen fertilizer on lowland rice plots, based on a single soil property, helps farmers to increase and optimize fertilizer use. At this seminar, I will report the results and findings obtained so far.