There is a peculiar form of madness — unrecognised as such because it wears the respectable garments of rationality — that consists in the relentless optimisation of means without any corresponding scrutiny of ends. It is the madness of the organisation that measures everything and understands nothing, that can tell you with exquisite precision how many seconds each employee spends on each task but cannot articulate, even approximately, what the tasks are ultimately for. This pathology, which has penetrated so deeply into the fabric of institutional life that it has become virtually invisible, deserves a name and a diagnosis. Let us call it, for want of a better term, the efficiency delusion.
The intellectual genealogy of the efficiency delusion can be traced to the work of Frederick Winslow Taylor, whose Principles of Scientific Management, published in 1911, inaugurated the systematic study of labour productivity. Taylor's central insight — that industrial processes could be decomposed into their constituent motions and each motion optimised for speed and economy — was, within its original domain, both valid and transformative. The resistance it provoked from workers and unions was grounded not in any quarrel with the principle of efficiency itself but in the recognition that Taylor's method, by reducing human labour to a sequence of measurable physical operations, had excised from the definition of work everything that made it meaningful: skill, judgement, autonomy, and the satisfaction of craftsmanship.
What Taylor could not have foreseen was the extraordinary migration of his principles from the factory floor to domains where they are not merely inapplicable but actively corrosive. The contemporary enthusiasm for quantified performance metrics in education, healthcare, policing, and scientific research represents precisely this kind of categorical transplantation — the imposition of an industrial logic upon activities whose essential character resists quantification. When a hospital measures its success by patient throughput, or a university by the number of publications its faculty produces, or a police force by its arrest statistics, what is being optimised is not the activity itself but a numerical proxy for the activity, and the relationship between the proxy and the thing it purports to represent is far more tenuous than the advocates of measurement are generally willing to acknowledge.
The consequences of this displacement are by now extensively documented. The sociologist Donald Campbell formulated the principle with lapidary concision: 'The more any quantitative social indicator is used for social decision-making, the more subject it will be to corruption pressures and the more apt it will be to distort and corrupt the social processes it is intended to monitor.' Teachers teach to the test; doctors discharge patients prematurely to improve turnover figures; researchers fragment their findings into the smallest publishable units to inflate their citation counts. In each case, the metric has ceased to be a measure of performance and has become a substitute for it — a target that is pursued for its own sake, irrespective of whether its attainment serves the purpose the activity was originally designed to fulfil.
The deeper problem, however, is not that metrics are imperfect proxies — this could, in principle, be remedied by the development of better metrics — but that the entire enterprise of quantified evaluation rests upon an epistemological assumption that is, for many human activities, simply false: the assumption that the most important aspects of performance are the ones that can be measured. In teaching, the qualities that matter most — the capacity to inspire curiosity, to model intellectual honesty, to create the conditions in which genuine understanding becomes possible — are precisely those that resist reduction to numerical form. To insist upon measuring them nonetheless is not rigour but a parody of rigour, a confusion of precision with accuracy that mistakes the map for the territory.
The antidote to the efficiency delusion is not the abandonment of measurement but its demotion — the recognition that quantitative data, however useful as one input among many, cannot serve as the sole or even the primary basis for evaluating activities whose value is inherently qualitative. This requires, above all, a willingness to trust human judgement in domains where human judgement is the only instrument adequate to the complexity of the task — a willingness that our current institutional culture, with its deep distrust of the unquantifiable, seems increasingly reluctant to extend.