The great flattening is usually pitched as a universal upgrade, as if any organization could shed its middle layer and run lighter. The evidence is narrower and more useful than that. Flat works, and works spectacularly, under conditions that many companies do not meet. The mistake is not admiring the model. It is assuming the conditions come free.
Start with the clearest success. Buurtzorg supports more than 900 self-managing nursing teams and roughly 14,700 staff with no middle managers at all (Martela and Nandram, 2025). What makes that possible is not culture alone. The authors are explicit that the model works best where interdependence between teams is low, where frontline staff are highly motivated and competent, and where the service can be delivered in modular units (Martela and Nandram, 2025). Home nursing satisfies all three. A team can care for its patients largely without waiting on another team, the nurses are skilled professionals, and the work divides cleanly.
Where flat thrives
Those three conditions are not decoration on the case. They are the load-bearing structure. Low interdependence is what lets teams coordinate among themselves instead of routing through a manager, because there is little cross-team coordination to route. Motivated, competent frontline staff are what let decision rights sit at the edge without producing chaos. Modular work is what allows a team to be a complete unit rather than one station on a fragile assembly line. Remove any one and the arithmetic that lets 21 coaches support 950 teams stops working (Martela and Nandram, 2025).
Read positively, this is a checklist. An organization whose work is genuinely modular, whose people are capable of self-direction, and whose teams rarely block one another is a strong candidate to flatten and will likely find the middle layer was carrying less than it seemed. That describes a real slice of professional services, field operations, and specialist care. For those firms the great flattening is an opportunity the evidence supports.
The conditions are also partly buildable, which is the hopeful reading. Interdependence is not purely a fact of nature; it is shaped by how work is divided. A company can often lower it by redrawing team boundaries so that each unit owns a complete slice of the work rather than a fragment of it. Capability can be grown through hiring and development. Modularity can be engineered into a service that was previously delivered as one tangled process. None of this is quick, and some businesses are structurally interdependent in ways no redesign can undo. But treating the conditions as levers rather than fixed traits is what separates a company that flattens on purpose from one that simply hopes its work already qualifies.
Where it breaks
Now invert the checklist. Where work is tightly coupled, where one team constantly waits on another, the coordination a middle manager brokered does not disappear when the manager does. It becomes everyone's part-time problem, which is to say no one's, and the result is the complexity McKinsey found two-thirds of leaders already reporting in their organizations (McKinsey, 2023). Flatten a highly interdependent operation and you have not distributed management. You have deleted it and left the interdependence in place.
The cautionary history sits in the earlier wave of radical self-management. Writing in Harvard Business Review, Ethan Bernstein and his co-authors urged a balanced view of the movement against the extremes of both hype and dismissal, and noted that Zappos was then the largest company to have implemented holacracy in full (Bernstein and colleagues, 2016). Zappos later stepped back from that system. The point is not that flat structures fail; Buurtzorg shows they can scale for years. The point is that the same model can thrive in one setting and strain in another, and the difference is the conditions, not the ambition. Bernstein's call for a sober assessment reads, a decade on, as a warning against importing a structure without first checking whether your work fits it.
Read your own conditions first
The practical discipline is to diagnose before you flatten. Deloitte recorded US middle-management postings falling 42 percent from their 2022 peak, much of it driven by cost pressure rather than a considered read of whether the work could survive without the layer (Deloitte, 2025). A company that flattens because flattening is in fashion, without asking whether its work is modular and its interdependence low, is running Buurtzorg's play on a field where Buurtzorg's rules do not apply.
Ricardo Semler's own version of self-management succeeded in part because Semco pushed autonomy where the work could bear it and did not pretend one design fit every corner of the business. That is the honest lesson of the flattening evidence. The question is not whether flat is good, in the abstract it clearly can be. The question is whether your work is modular, your people ready, and your teams independent enough to hold the coordination themselves. If the answer is yes, the middle manager may genuinely be optional. If it is no, removing the layer will not make the conditions appear. It will only remove the person who was compensating for their absence.
Sources
- Martela and Nandram (2025). Buurtzorg: scaling up an organization with hundreds of self-managing teams but no middle managers. Buurtzorg runs 900+ teams with no middle managers under named boundary conditions: low interdependence, motivated frontline staff, and modular services. https://link.springer.com/article/10.1007/s41469-024-00184-y
- Bernstein and colleagues (2016). Beyond the Holacracy Hype. Argues for a balanced assessment of self-management and notes Zappos was the largest full holacracy adopter as of 2016 (a system it later stepped back from), showing the same model can strain in a different setting. https://hbr.org/2016/07/beyond-the-holacracy-hype
- McKinsey (2023). The State of Organizations 2023. Two-thirds of leaders call their organizations overly complex, the outcome of flattening tightly interdependent work. https://www.mckinsey.com/capabilities/people-and-organizational-performance/our-insights/the-state-of-organizations-2023
- Deloitte (2025). 2025 Global Human Capital Trends. US middle-management postings fell 42%, much of it cost-driven rather than a considered read of whether the work could survive flat. https://www.deloitte.com/us/en/insights/topics/talent/human-capital-trends/2025.html