Photo by Blake Wisz on Unsplash

McKinsey made headlines in 2012 when they declared middle management obsolete. With digital tools enabling direct communication between executives and frontline workers, the thinking went, why waste money on layers of people in between? That message resonated. Over the next decade, Fortune 500 companies slashed middle management roles by nearly 15%. Startups bragged about their flat organizational structures. Consultants got rich selling delayering strategies.

Then something unexpected happened. Companies started quietly rehiring middle managers at accelerating rates.

The Hidden Costs of Flattening Hierarchies

When Zappos eliminated their entire middle management tier in 2013, CEO Tony Hsieh celebrated the move as liberation from "the management hierarchy." For about eighteen months, it worked beautifully. Then the cracks appeared. Critical information got lost between executive vision and frontline execution. New employees had nobody to mentor them. Project coordination became chaotic. Attrition spiked. Within three years, Zappos quietly began rehiring managers.

Zappos isn't alone. Microsoft underwent aggressive delayering in the early 2000s under Steve Ballmer. Product quality suffered. When Satya Nadella took over in 2014, one of his first moves was restoring managerial layers and creating clearer accountability structures. The company's stock price, which had flatlined for a decade, more than tripled in the following years.

Research from the Harvard Business Review analyzing 30 years of corporate data found something counterintuitive: companies with 3-4 layers of management between executives and frontline workers actually outperformed those with fewer layers on nearly every metric—profitability, innovation speed, employee retention, and customer satisfaction.

What Middle Managers Actually Do (Besides Attend Meetings)

Here's what gets lost when you eliminate middle managers: context. Individual contributors see their specific task. Executives see strategic vision. Middle managers are the translation layer. They understand why the executive's directive matters and how it applies to the team in front of them. They catch misaligned priorities before they waste months of work. They know which individual contributors have the expertise to solve specific problems.

Take a manufacturing company that decided to cut middle management layers. Their frontline supervisors suddenly reported directly to the VP of Operations. The VP, managing 47 people instead of 7 managers, couldn't possibly know that Maria in the packaging line had solved a similar efficiency problem three years ago, or that the new initiative conflicted with another priority that hadn't been communicated yet. Projects overlapped. Expertise was duplicated. Efficiency dropped 8%.

Middle managers also serve as shock absorbers. They absorb unrealistic demands from above and translate them into achievable goals. They absorb complaints from below and filter out the noise to identify real problems. They protect teams from constant organizational chaos. Remove them, and you expose workers to raw executive whim.

The Rise of the "Dual Ladder" Problem

Some companies tried a middle path: create "individual contributor" promotion tracks so talented specialists could advance without managing people. This solved part of the problem—it kept expertise in the organization. But it created something worse: resentment.

Why? Because you can't actually have equal status without power. If a Senior Principal Engineer and a Senior Manager have the same title and salary but different decision-making authority, everyone knows who actually has seniority. It's theater. And talented people see through theater quickly.

Google discovered this the hard way. They created elaborate IC (individual contributor) levels to compete with management tracks. Initially, it worked. But by 2016, internal surveys showed Senior Principal Engineers, despite impressive titles, felt like second-class citizens when their opinions were systematically overruled by managers. The solution wasn't to create better titles—it was to hire more managers to actually respect the expertise these engineers brought.

The Real Issue Nobody Mentions: Speed vs. Quality

The original delayering argument had one legitimate advantage: faster decision-making. With fewer layers, approvals move quicker. Early Amazon—relentlessly fast, minimal hierarchy—moved at breathtaking speed and could outmaneuver slower competitors.

But there's a threshold. Once an organization reaches a certain size or complexity, fewer layers don't speed things up anymore—they bottleneck them. A single VP getting pinged with requests from 60 direct reports doesn't make decisions faster. They make decisions worse and more slowly.

What companies actually needed wasn't fewer managers—it was better managers making faster decisions. The problem was never the structure. It was the quality of leadership and the decision-making authority given to managers.

This connects to a broader organizational challenge. As research on remote team management shows, when people don't have clear reporting structures and support, engagement plummets. The same principle applies to organizational hierarchies—clarity and support matter more than flatness.

What Companies Are Actually Doing Now

Smart organizations aren't going back to 1990s corporate structures. They're being intentional about span of control. Microsoft's "span of control" became a key metric under Nadella—they aimed for 8-10 direct reports per manager rather than 3-5 or 60+. This gives managers enough leverage without overloading them.

They're also being ruthless about what middle management actually does. The worst middle managers simply relay messages up and down. The best ones add value at each level: mentoring individuals, translating strategy into tactics, protecting teams from chaos, and solving cross-functional problems that nobody else sees.

General Motors, after struggling with bloated middle management in the 1990s and overcorrecting by cutting too many, now uses a more nuanced approach. They maintain managerial layers but demand that each one adds specific value. If a manager's primary function is "attending more meetings and reporting up," they're eliminated. If a manager develops talent, coordinates across silos, and removes obstacles for their team, they're protected and invested in.

The Honest Takeaway

The worst kind of middle manager is someone who exists purely to enforce hierarchy. But the best kind is invaluable—they're mentors, translators, problem-solvers, and filters for chaos. The optimal organization doesn't minimize management. It optimizes for it. That means hiring the right people into management roles, giving them real authority to make decisions, and measuring them on what they actually deliver rather than how many direct reports they have.

The era of "flatter is always better" is quietly over. Companies that realized this early are stealing talent and market share from those still operating under the old religion of delayering.