What the research found
Growth stages and the irony of solved problems. Larry Greiner, in research first published in Harvard Business Review and republished as an HBR Classic in 1998, described organisational growth as a sequence of phases — creativity, direction, delegation, coordination, collaboration — each ending in a crisis produced by the practices that resolved the previous one: a leadership crisis, then an autonomy crisis, then a control crisis, then a crisis of red tape. His central argument was that a leadership approach is never simply right or wrong in the abstract; it is right for a given phase and becomes the seed of the next problem once the organisation outgrows it. Managers, in his words, “experience the irony of seeing a major solution in one period become a major problem in a later period.”
Coordination, not alignment, as the binding constraint. Donald Sull, Rebecca Homkes and Charles Sull, drawing on a multi-year study of more than 250 companies and several thousand managers for MIT Sloan and London Business School, found that only 9% of managers said they could reliably count on colleagues in other functions all of the time. Their central finding was that most organisations’ vertical alignment — the standard target of leadership attention — is already sound. The actual constraint is horizontal: cross-unit coordination, which requires a different kind of leadership attention than alignment does, and one that becomes more, not less, important as an organisation grows and its units multiply.
Where decisions actually stall. Paul Rogers and Marcia Blenko, partners at Bain & Company, surveyed executives at more than 350 global companies and found only about 15% considered their organisation structurally set up to outperform competitors. Their research located decision-making difficulty at a small number of recurring structural seams — between global and local, centre and business unit, function and function, and organisation and outside partner — rather than attributing it broadly to individual capability. This points toward a leadership requirement that is architectural rather than behavioural: designing where decisions get made, not simply making more of them personally.
Incentives outrunning stated intentions. Steven Kerr’s research, developed at City University of New York and later applied at General Electric, examined reward systems across manufacturing, insurance, and settings well beyond business, finding a consistent pattern: people orient toward what is actually rewarded rather than what is stated as the goal. As organisations grow and reward systems formalise — metrics, targets, review cycles — the gap between what leadership says it wants and what the system actually reinforces becomes both more consequential and harder for leadership to see directly, because leadership experiences its own stated intentions, while employees experience the system.
Psychological safety as the condition for information to travel. Amy Edmondson’s Harvard Business School field study of 51 manufacturing teams found that psychological safety, not team confidence, predicted whether people surfaced problems and raised concerns. In a small organisation, information often reaches leadership through proximity, regardless of whether the underlying environment is genuinely safe. As distance between leadership and the front line increases with growth, proximity stops compensating, and psychological safety becomes the primary determinant of whether leadership hears about problems at all — a shift from an accidental advantage to an explicit requirement.
Barriers that compound rather than sit in isolation. Michael Beer and Russell Eisenstat, writing for Sloan Management Review from Harvard Business School, identified a set of mutually reinforcing “silent killers” of strategy implementation, including unclear priorities and poor vertical communication — barriers that are, by design, hard for leadership to see because no single instance looks like a failure. Their point was structural: these barriers reinforce each other, so that addressing one without the others tends not to hold, and a barrier “has to be turned into a capability” before others can be resolved either.
Where the evidence agrees
Set alongside each other, these six bodies of work agree on more than their differing terminology suggests. All six locate the central challenge of growth not in individual leadership capability declining, but in the growing distance between where decisions are made and where the information needed to make them well actually lives — whether that distance is described as coordination failure, structural bottleneck, incentive drift, or psychological unsafety. None of the six attributes the core problem to leaders becoming less capable as organisations scale. All six describe mechanisms that were adequate substitutes at small scale — proximity, informal relationships, personal modelling — quietly ceasing to function as the organisation grows past the size at which they can reach everyone who needs them, without necessarily being replaced by anything.
A second point of agreement is less obvious but arguably more consequential: in every one of the six accounts, the gap in question is largely invisible to leadership from the inside, at least initially. Sull’s coordination failures don’t announce themselves as coordination failures — they surface as missed deadlines or duplicated work. Kerr’s incentive drift is invisible precisely because leadership experiences its own stated intent, not the reward system as employees experience it. Beer and Eisenstat make this explicit, describing their barriers as commonly invisible to the leadership responsible for them. This shared feature — that the mechanism is not something leadership can simply notice by paying closer attention — is arguably the single most practically important point of agreement across all six bodies of work, because it means the absence of visible problems is not, on its own, evidence that these mechanisms are functioning well.
Where the evidence diverges, and what remains unresolved
The six bodies of work diverge, however, on where leadership attention should be primarily directed, and the disagreement is genuine rather than merely terminological. Greiner’s framing points toward sequencing — the claim that different leadership approaches are correct at different stages, which implies leadership requirements change in a roughly predictable order. Rogers and Blenko’s research points toward structure — the claim that the underlying issue is architectural and largely independent of stage, since their decision bottlenecks appear across companies of very different sizes and growth trajectories. These two views are not strictly incompatible, but they suggest different diagnostic starting points: one asks “what stage are we in,” the other asks “where in our structure does this specific decision actually stall.”
A second unresolved tension sits between Kerr’s incentive-based explanation and Edmondson’s safety-based explanation for why information and intended priorities fail to reach leadership. Both are well evidenced independently, and both plausibly operate at once in most real organisations, but neither body of research establishes which, if either, is the primary driver in a given case — or whether they interact, with an unsafe environment making people more likely to orient toward measured incentives precisely because raising unmeasured concerns feels risky. That interaction is a genuinely open question the existing research does not resolve.
A third, smaller disagreement worth naming: Beer and Eisenstat’s compounding-barriers account implies that partial progress on any single mechanism is close to worthless until several are addressed together, since the barriers reinforce each other. Sull’s coordination finding, by contrast, is more agnostic on sequencing — it identifies where the problem sits without making a strong claim about whether it must be addressed alongside other issues to matter. An organisation trying to decide whether to tackle one of these six mechanisms in isolation, or treat them as a connected set, will not find a single answer in the literature; it will find two studies that were never designed to be reconciled with each other making genuinely different implicit assumptions.
The same pattern appears across many kinds of organisation; see Other Sectors for how it is read elsewhere.
Questions worth considering
- Which of these six mechanisms — coordination, structural bottlenecks, incentive drift, safety, sequencing, or compounding barriers — best describes what is currently changing in this organisation, and is it possible more than one is operating simultaneously?
- Has any mechanism that worked well at a smaller scale — proximity, a founder’s personal attention, informal relationships — been consciously replaced as the organisation has grown, or has it simply stopped functioning unnoticed?
- If leadership’s stated priorities and the organisation’s actual incentive structure were compared directly, would they still point the same direction?
This perspective is part of 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.