Hierarchy was the dominant organisational structure of the twentieth century for reasons that made sense: when communication was slow and information was expensive, centralising decisions at the top and executing them down a chain of command was often the most efficient arrangement. Senior leadership had context that middle and front-line levels lacked, and the cost of pushing decision-making down the hierarchy — information collection, training, coordination — often exceeded the benefit.
Those conditions no longer hold. Communication is instant, information is cheap, and the organisations that have adapted most successfully to the current environment are those that have restructured decision authority to sit at the level closest to the information relevant to the decision. This restructuring takes different forms in different organisations, but the most coherent framework for understanding it is what organisational theorists have begun calling fractal governance.
The Fractal Principle
A fractal structure is self-similar at different scales — the same pattern repeats whether you’re looking at the organisation as a whole or at a small unit within it. In organisational terms, fractal governance means that small teams operate with the same clarity of purpose, decision authority, and accountability structures as the organisation as a whole. Each unit is not a subordinate piece of a larger machine but a functioning organisation in miniature, with its own strategy, metrics, and decision-making capacity within defined boundaries.
Spotify’s “squad” model — individual cross-functional squads with full ownership of a product area, organised into “tribes” aligned with broader product domains — is the most widely discussed implementation of this principle. Amazon’s “two-pizza team” rule, which holds that no team should be larger than can be fed by two pizzas, reflects the same instinct: small, autonomous units outperform large, coordinated ones for most tasks in fast-moving environments.
What Actually Changes in Practice
The practical shift from conventional hierarchy to fractal governance involves three changes that are easy to describe and genuinely difficult to implement.
First, decision rights have to be genuinely transferred, not just nominally. The failure mode of most “empowerment” initiatives is that formal authority is delegated while real authority — the ability to make consequential decisions without senior override — is retained at the top. Teams that can make decisions only until those decisions become uncomfortable for leadership are not autonomous teams; they’re conventionally managed teams with a different name.
Second, the boundaries within which autonomous decision-making is permitted need to be explicit and stable. Autonomy without clarity is chaos. Teams that don’t know the limits of their authority either avoid decisions (reverting to asking permission for everything) or exceed their mandate (creating coordination problems with other teams). The architectural work of defining what each unit owns — and what it doesn’t — is as important as the decision to give them ownership in the first place.
Third, accountability must be real and legible. Autonomous teams need clear metrics that reflect their actual impact, evaluated honestly and consistently. The accountability structures in fractal organisations are not softer than those in hierarchies — they’re different. Where hierarchy holds people accountable for following processes, fractal governance holds them accountable for outcomes. This is more demanding, not less, and many people find the adjustment uncomfortable.
When Fractal Governance Doesn’t Work
The fractal model is well-suited to knowledge work in fast-moving environments where the person closest to the problem typically has the most relevant information. It is poorly suited to industries where consistency and coordination are the primary value drivers: regulated financial services, large-scale manufacturing with interdependent production lines, or emergency services where hierarchical command structure under pressure is genuinely safer.
Scale matters too. The coordination costs of fractal governance — the communication overhead required for teams operating autonomously to stay aligned — grow with organisational size in ways that eventually overwhelm the benefits. The companies making fractal models work at the largest scale (Google, Amazon, Spotify) invest enormous resources in the infrastructure — technical, cultural, and structural — that makes autonomous teams coherent across the whole. For smaller organisations without those resources, a less elaborate version of the principle often serves better than the full implementation.
2 thoughts on “Fractal Organizational Governance”