The Middle Management Trap: Why Elimination and Preservation Both Fail

The Middle Management Trap: Why Both Sides Are Wrong

The pendulum keeps swinging. In the 1990s, companies delayered aggressively, chasing the promise of speed. Today, after watching Twitter's chaos and remembering Boeing's 737 MAX disaster, we're reconsidering. But this isn't a binary choice between preserving bureaucracy and creating organizational chaos. The real question is whether you have the discipline to fix what middle management actually does—or whether you're just choosing which failure mode you prefer.

The stakes are immediate. With economic pressure mounting and AI promising to automate coordination work, executives are once again eyeing their org charts and calculating the cost of those director and senior manager layers. Before you make that cut—or defend the status quo—you need to understand why this debate keeps producing the same expensive mistakes.

What the Debate Revealed

The initial positions appeared to align around preservation, but the reasoning exposed fundamental tensions. The org design perspective advocated keeping the layer while cutting 30% of the population and completely redesigning what remains. The people and leadership perspectives defended middle management as translators, shock absorbers, and knowledge keepers. The operator warned about the predictable chaos that follows elimination.

But in the second turn, the debate sharpened around a critical insight: redefining middle management without changing how you measure and empower them is just delayering with better PR.

The org designer pushed back on using management layers as career trophies, arguing for decoupling compensation from headcount. The people perspective countered that redefinition requires new accountability frameworks—measuring managers on team psychological safety and knowledge transfer rather than span of control. The leadership view hardened around a crucial point: redefining roles without granting real decision rights creates powerless bureaucrats, which is worse than elimination.

"If we're preserving middle management, they need real decision rights: budget allocation up to defined thresholds, unilateral hiring authority for backfills, and the power to kill projects consuming their teams. Otherwise, you're just creating expensive messenger pigeons."

The operator delivered the most pragmatic challenge: before redesigning anything, what are you willing to stop? What decisions are you truly pushing down? What meetings are you killing? Without that, transformation is just theater.

This is where most organizations fail. They want the efficiency of fewer managers but maintain the same reporting cadence, approval processes, and coordination demands. The result is predictable: overloaded managers who become bottlenecks rather than enablers, followed by either quiet rebuilding of the layers or mass exodus of talent.

The Framework: Four Tests Before You Decide

Stop asking whether to eliminate middle management. Instead, run these four tests to determine what your organization actually needs:

Test One: The Decision Altitude Test
Map the last 50 significant decisions in your organization. At what level were they made versus where they should have been made? If executives are deciding things three levels below their optimal altitude, you don't have too many layers—you have managers without authority. If decisions are being made too low without adequate context, you need better translation mechanisms.

Test Two: The Span of Control Reality Check
Calculate your executives' effective spans. If VPs have 15+ direct reports, they're managing no one effectively. The research is clear: beyond 8-10 directs, management becomes administrative rather than developmental. But here's the nuance—span of control depends on work complexity and team maturity. A VP overseeing ten autonomous product teams needs less management density than one coordinating interdependent engineering squads.

Test Three: The Knowledge Transfer Audit
Identify your critical institutional knowledge. Who holds it? How is it being transferred? When Zappos went holacratic, they didn't just lose hierarchy—they lost the sense-making work that middle managers performed. If your middle managers are primarily forwarding emails and attending meetings, that knowledge isn't being captured or transferred. But if they're the ones who know why the system works the way it does, elimination is organizational amnesia.

Test Four: The Workload Reduction Test
This is the operator's challenge, and it's the one most organizations fail. List everything you're asking middle managers to do. Now cut 40% of it. If you can't identify what to stop, you're not ready to redesign anything. Transformation without workload reduction is just adding responsibilities to already-overloaded roles.

The Nuance: When Context Changes Everything

Company stage matters enormously. A 50-person startup adding its first management layer faces different dynamics than a 5,000-person enterprise considering delayering. Early-stage companies often add management too slowly, creating burnout among founding team members. Mature companies typically add layers as defensive moves—creating roles to retain people or reduce executive spans without considering whether the work requires it.

Industry complexity is another variable. Boeing's engineering organization requires different management density than a software company. When technical decisions carry safety implications and require deep institutional knowledge, the cost of losing middle management expertise is catastrophic. When work is more modular and reversible, you can operate with flatter structures.

The quality of your existing middle management is perhaps the most important factor. If you have exceptional managers who are force multipliers, preservation and empowerment make sense. If you have mediocre managers who became administrators, the org designer's approach—cut 30% and radically redesign what remains—is correct. But be honest: most organizations have a mix, and the political reality is that cutting the bottom 30% is harder than cutting entire layers.

Remote and distributed work changes the equation. GitLab's fully remote structure with 2,000+ people demonstrates that flat organizations can work when you're intentional about communication architecture and decision rights. But they've also been deliberate about measuring manager effectiveness on team engagement and knowledge transfer rather than traditional metrics.

Where to Start: Five Concrete Actions

First, conduct the decision rights audit. Before changing your structure, map where decisions are actually being made versus where they should be made. This reveals whether you have a layer problem or an authority problem. Most organizations discover they have the latter.

Second, redefine and measure what matters. Stop measuring middle managers on span of control and administrative efficiency. Start measuring them on team psychological safety scores, retention of high performers, knowledge transfer effectiveness, and the number of decisions they enable others to make. What you measure is what you get.

Third, grant real decision rights or eliminate the role. Give middle managers budget authority up to defined thresholds, unilateral hiring authority for backfills, and the power to kill projects. If you're not willing to grant these authorities, you don't actually want managers—you want coordinators, which is a different (and less expensive) role.

Fourth, create parallel advancement paths. Decouple career progression from hierarchical ascension. A senior individual contributor who coordinates five teams without direct reports should have compensation and status equivalent to a director. Otherwise, you're forcing people into management roles they don't want and shouldn't have.

Fifth, radically reduce the workload before redesigning anything. Identify what you'll stop: which meetings you'll kill, which approval processes you'll eliminate, which reports you'll discontinue. Transformation without workload reduction is just adding responsibilities to already-overloaded roles, which accelerates the failure cycle.

The Real Choice

The debate about eliminating middle management is a proxy for a harder question: do you have the organizational discipline to be clear about what work actually needs to be done, who should do it, and what authority they need? Most companies don't. It's easier to reorganize than to make those fundamental choices about decision rights, accountability, and workload.

The organizations that succeed aren't the ones that eliminate layers or preserve them—they're the ones that design roles around actual value creation and then have the courage to stop everything else. That's uncomfortable work. It requires saying no to executives who want both efficiency and control, to managers who want authority without accountability, and to systems that exist because they've always existed.

But here's what makes this moment different: AI is about to automate much of the coordination and information synthesis work that middle managers currently perform. The question isn't whether to preserve or eliminate these layers—it's whether you can redesign them around the irreplaceable human work of judgment, development, and sense-making before technology forces your hand. The companies that figure this out now will have a decisive advantage. The ones that keep debating will find the decision made for them.