Organisations are rarely incoherent because of what they contain; more often, they are incoherent because of how what they contain is arranged. Organisational design is typically treated as a problem of parts: the right roles, the right reporting lines, the right allocation of decision rights. Coherence, in this framing, is something that follows from correct specification, and its absence is evidence that some element has been wrongly defined or wrongly placed. Yet this way of thinking sits uneasily with a familiar experience. Organisations that possess every correct element still fail to cohere, while others, assembled in ways no framework would endorse, hold together with an ease that resists explanation. What differs between them is not visible in their parts, which leaves us with a question: if coherence cannot be secured by the correctness of elements, are the elements the wrong place to begin?
Organisational design is most often approached as a problem of specification. Charts are drawn, roles are defined, spans of control are calculated, and decision rights are allocated through tools built for the purpose: operating model reviews, accountability matrices, competency frameworks, periodic restructures. The assumption running through these methods is that an organisation is the sum of its defined components, and that if each component is correctly specified and logically positioned, order will follow. When it does not, the response is to respecify: redraw the chart, rewrite the roles, restructure once more. Taken together, these approaches locate design in the accuracy of parts and the logic of their allocation, and they treat the document that records this allocation as a faithful account of the organisation itself.
The limitation of these methods is not that they are wrong, but that they are elemental. They assume that coherence resides in components: that a role, correctly written, will behave as written; that a structure, once specified, will be experienced as specified. What they struggle to account for is an observation available to anyone who has moved between organisations: two of them can hold identical charts and radically different coherence. Same titles, same functions, same lines of reporting; entirely different experiences of order. If the elements are the same and the outcomes are not, then whatever produces coherence does not live in the elements.
A more useful way of approaching this is through composition. In art, composition names the arrangement of elements such that meaning becomes legible. It governs balance and imbalance, hierarchy, focal points, tension, the movement of the eye across a surface. A painting can contain all the right elements, rendered with skill, and still fail if the composition is wrong, because composition determines how a work is perceived before it is interpreted, and perception always precedes explanation. Crucially, composition operates on the level of relation, not parts. Meaning emerges not from the elements themselves but from spacing, proximity, contrast, and absence; painters give this last one a name, negative space, the deliberate emptiness that lends everything else its weight. Two identical elements can communicate entirely different things depending on where they sit. Composition, in other words, is a theory of relational order. Organisations are not paintings, but the parallel is instructive. Titles, roles, and functions are elements; structure lies in their relations. What matters is the proximity of decision-making to the work it governs, the spacing between authority and execution, the contrast between responsibility and power, and, no less, the negative space: what is deliberately left undefined, and whether that absence has been composed or merely neglected.
It is important to note that composition offers no argument, it does not persuade or justify. Viewers sense that a composition works, or does not, before they can articulate why. Coherence, in this sense, is perceptual before it is analytical; structure communicates itself prior to conscious reasoning. You cannot convince someone that a composition works. It either reads or it does not. The same holds inside organisations. People feel where decisions actually originate, where responsibility concentrates, whether authority and accountability correspond, long before, and often despite, the official account. Structure is sensed, not declared; it announces itself through coherence or announces its absence. This is why a design that must be continually explained has, in compositional terms, already failed: explanation is doing the work that arrangement was meant to do. The reframing shifts organisational design out of the language of specification and into the quieter register of arrangement.
Furthermore, composition treats constraint not as a limit on order but as its condition. Frame, medium, scale, and material all impose boundaries, and composition emerges because of them, not despite them. Remove constraint and hierarchy collapses, focus dissolves, meaning diffuses. Hierarchy, seen this way, functions like visual focus: the eye needs to know where to land. In organisations, clarity depends on the same economy of attention, on knowing where decisions originate and where responsibility concentrates. Flatness is often pursued as a virtue, a refusal of structure in the name of openness, but flatness without compositional hierarchy produces in organisations what it produces on canvas: noise, everything present and nothing legible. Seen through this lens, a well-drawn chart loses its status as proof of design. The question, then, is no longer whether an organisation contains the right elements, but why the chart is so often mistaken for the composition.