Measuring impact across grant-funded portfolios
This guide addresses the grant-giver problem: the difficulty funders face when evidence generated across multiple grants cannot be interpreted consistently at portfolio level. Where organisations define outcomes differently, use different measures, and apply different evaluation approaches, data accumulates as separate accounts of delivery rather than as a coherent picture of change. As a result, funders can confirm activity but cannot reliably determine whether investment is producing intended outcomes across the portfolio as a whole, creating operational, strategic, and accountability risk.
Drawing on a review of funder practice, the guide examines the main ways this problem has been approached in the literature, including governance reforms, reporting standardisation, flexible evaluation design, learning systems, data integration, framework design, capacity building, and outcome-focused reporting. While each approach improves specific aspects of evaluation, the review shows that none resolve the underlying issue of producing evidence that can be interpreted coherently across heterogeneous portfolios of funded activity. The guide then illustrates how these challenges emerge in practice through three scenarios: a centrally commissioned programme with external evaluation control; a partnership-based delivery model with shared outcomes but fragmented measurement; and a small-grants model where outcome definition and evaluation design are largely delegated to funded organisations. In each case, evidence is generated but cannot be aggregated into a clear portfolio-level account of performance.
In response, the guide sets out a structured approach to designing evaluation systems that align outcomes, indicators, tools, and methods before data collection begins, supported by the resource requirements needed to implement this in practice. This enables project-level data to be combined into a coherent portfolio-level evidence base that supports decision-making across complex, multi-partner programmes.
