The case for structure
Paul Rogers and Marcia Blenko’s research at Bain & Company, based on a survey of executives at more than 350 global companies, found that only about 15% of organisations were structurally set up to outperform their competitors — and located the shortfall at specific, identifiable seams: between global and local levels, between centre and business unit, between functions, and between the organisation and outside partners. This is evidence for a structural view: decision-making failure clusters at nameable points in an organisation’s design, which implies that redesigning those points — clarifying who decides what, and where — should measurably improve outcomes, independent of who occupies any given seat.
Larry Greiner’s account of organisational growth phases reinforces this from a different direction. Practices that depend on one person’s judgement or attention — informal coordination, a founder’s personal oversight — are, in his account, appropriate to an early phase specifically because the organisation is small enough for them to work. The implied trajectory is that structure is meant to progressively absorb what personal judgement used to carry, phase by phase, as an organisation matures. On this reading, the fact that some problems currently require leadership judgement is not evidence that they always will — it may simply mean the organisation has not yet built the structure that would let judgement retire from that particular role.
The case for leadership
Amy Edmondson’s Harvard Business School field study of 51 manufacturing teams complicates the structural view considerably. Her finding was that psychological safety, not team efficacy, predicted whether people surfaced problems and engaged in learning behaviour — and psychological safety is not a property of a process or an org chart. It is produced, continuously, by how people in a position of authority actually respond when something difficult is raised. A structure can mandate that concerns be escalated through a defined channel; it cannot make raising a concern feel safe. That has to be established and re-established through behaviour, repeatedly, by whoever holds authority in a given relationship — which is a leadership function that does not transfer to a process no matter how well the process is designed.
Michael Beer and Russell Eisenstat’s research on strategy implementation identifies a closely related limit. Their “silent killers” — including an unclear or conflicting senior team style and poor vertical communication — are, by their own description, mutually reinforcing and frequently invisible to the very leadership responsible for them. A structural fix aimed at one barrier in isolation tends not to hold, in their account, precisely because the barriers reinforce each other at the level of how senior people actually behave together — something a process change does not directly touch. As they put it, a barrier has to be turned into a capability before others can be as well, which describes a sequence of leadership behaviour changing, not a system being installed.
Donald Sull and colleagues’ finding — that the central execution problem is cross-unit coordination, not vertical alignment, across more than 250 companies studied — can be read to support either position, which is itself informative. Better structure could plausibly improve coordination. But coordination between people who do not have a strong working relationship is also, irreducibly, a function of trust built through repeated interaction — something structure can create occasions for, but cannot manufacture directly.
Why the disagreement persists
Part of what keeps this question genuinely open is that the two positions are not actually testing the same claim. The structural case is strongest where the underlying issue is about where decision rights sit — Rogers and Blenko’s bottlenecks are, by definition, about structure, and improving structure plausibly improves them. The leadership case is strongest where the underlying issue is about whether something feels safe to say — and safety is demonstrably relational rather than positional, per Edmondson’s findings. An organisation experiencing a decision-rights problem and one experiencing a psychological-safety problem can look similar from a distance — both show up as things not surfacing when they should — while requiring entirely different kinds of intervention to change.
A further complication is that structure and leadership behaviour are not cleanly separable in practice. A well-designed decision-rights structure can still fail if the people occupying the structure do not model the behaviour that makes it trustworthy to use; conversely, strong leadership judgement in an environment with genuinely absent structure spends much of its capacity compensating for gaps that better design would have closed for free. Most real organisations sit somewhere between the two extremes the debate implies, which is part of why neither position fully displaces the other in the research.
There is also a timing dimension worth naming. Greiner’s phase account suggests that even where a problem is currently resistant to structural solutions, it may not remain so indefinitely — an organisation too small or too immature to build a working decision-rights structure around a given problem today may be able to do so later, once other structural foundations are in place. This means the structure-versus-leadership question may not have a single stable answer even within one organisation: the same recurring problem could plausibly sit in leadership’s territory now and migrate toward structure’s territory as the organisation matures, which the static framing of “which position is right” does not capture well.
Looking for your own sector? The Other Sectors page points to the closest match.
Questions worth considering
- For a given recurring problem, is the underlying issue closer to “nobody has the authority to decide this” or “nobody feels able to raise this” — and would the same fix address both?
- If a structural change were made — a new decision right, a new forum — would it hold without also requiring a change in how people in authority actually behave when it is used?
- Which of the organisation’s current problems would genuinely be unaffected by better structure, no matter how well designed?
More on this topic: Leadership.
Your team is facing similar challenges?
Subscribe to The Riddle Ledger
Perspectives are sent as they are published, with no promotional email, and you can unsubscribe at any time.