The humanitarian sector produces more conflict analysis than at any point in its history. It also produces more programmes that fail to achieve their intended social outcomes at community level. This paper argues that the relationship between these two facts is not coincidental. The dominant producers of context intelligence generate analysis calibrated to the scale and decision-making needs of coordination architectures in Geneva, Brussels, and New York. Practitioners who must implement programmes in specific sub-districts receive analysis that is accurate at the macro level and nearly useless at the operational level. This mismatch reflects a structural incentive: the organisations that produce analysis respond to the audiences that fund them, not to the practitioners who need it. Drawing on field observations across five operational theatres — Ethiopia, the Democratic Republic of Congo, Syria, Jordan, and Lebanon — this paper identifies five specific operational failures that follow from this structure, and proposes a design framework for a sub-national operational intelligence function that the sector currently lacks.