‘Educational reform movements would not have been able to get started without quantitative comparisons’, Niklas Luhmann once remarked, underlining how quantification itself fuels a need for decision-making. Rather than bringing rationality or transparency to the exercise of public power, numbers throw into relief new differences that motivate (political) action. Past research along the lines of his systems theory has mostly explored how quantitative differences yield new worlds: either as a doubling of reality that sets out the probable against the real or as the emergence of world society, itself resulting from the thus created statistical reality. With the present contribution, we seek to relate this world-constructing capacity of quantified data to the debate on transnational governance and accountability policies in education. Rather than seeing it as mere technocratic aid (as official discourse might present it) or as an instrument of the powers-that-be (as many social scientists insist), quantitative datafication is discussed against the background of steering limits. That is: as a repercussion from the paradoxical necessity to govern a world that appears ungovernable.