Key messages Multi-lateral development banks (MDBs) have committed to financing climate change mitigation in agriculture and have adopted a harmonized methodology for attributing and reporting climate finance; however, design (including practice selection) and measurement of project impacts remains ad hoc. Decision-support tools such as “practice lists” help users identify and select practices for climate-smart agriculture (CSA) or specific outcomes such as soil organic carbon (SOC) sequestration. However, interviews suggest they are not widely used at MDBs. Practice lists fail to provide absolute quantification of the net mitigation or sequestration impact, which is the parameter of primary interest to MDBs. Hence, many MDBs use ex-ante estimation methods such as EX- ACT to determine the net impact of an investment. MDBs require better guidance on practice selection, recognizing that barriers such as scientific uncertainty, context-dependence, and adoption-dependence (i.e. persistence) limit the value of lists or ex-ante models. Some measurement, such as adopted in the Australian Emissions Reduction Fund, would significantly reduce uncertainties, at a potential cost tradeoff. The potential co-benefits of building SOC have led to interest in developing concessional lending programs and insurance products that factor in SOC levels, provided a means can be found to cost-effectively and reliably estimate and monitor SOC levels.