The Structural Anatomy of Organisational Transformation Failure

Organisational transformation initiatives fail at a rate of approximately 70%, a figure that has remained stable across three decades and at least five distinct methodological waves: Total Quality Management, Business Process Reengineering, Lean Six Sigma, Agile, and Digital Transformation. The persistence of this rate across methodologies that explicitly position themselves as corrections to their predecessors suggests that the failure rate is not a property of any individual methodology but a structural constant of the system in which all methodologies operate. This paper proposes three contributions. First, it introduces two formal constructs for modelling engagement depletion at the individual and organisational levels: Foveation of Care (FoC), a finite, non-renewable unit of employee engagement within a given employer relationship, consumed by every demand on attentional resources from every layer of the organisational hierarchy; and the Change Ceiling (CC), the highest point of collective engagement that remaining FoCs can sustain, which ratchets downward with each abandoned or declared-but-not-completed initiative and does not recover. Second, it identifies a demand generation cycle operating through the alumni networks of major consulting firms in which methodology development, talent export to corporate leadership, demand generation by alumni-executives, consulting-led delivery, and failure diagnosis form a self-reinforcing loop whose incentive architecture is structurally misaligned with transformation success. Third, it names five invariant failure mechanisms that operate inside every transformation regardless of methodology. The paper develops these constructs through a case study of the Danish Ét Fælles Inddrivelsessystem (EFI) project and applies the framework predictively to the AI transformation wave, arguing that while the failure rate will be similar, the consequences will be categorically different due to compressed donemine timelines, invisible failure modes, and the absence of labour-market containment boundaries.