The Death of Mid-Level Management Was Not a Cultural Shift — It Was a System Failure

For more than two decades, we have been told a comforting story.

Mid-level management disappeared because it was bloated. Because hierarchies were slow. Because autonomy, agility, and flat organizations were said to be better.

That story is incomplete — and dangerously misleading.

Mid-level management did not disappear because organizations discovered it was useless. It disappeared because a specific design logic made its existence appear unnecessary.

In the early 2000s, large organizations began borrowing heavily from software engineering and agile development. These models promised faster adaptation, higher resilience, and better outcomes by empowering teams to make local decisions. Authority was pushed downward. Hierarchies were flattened. Decision-making was meant to happen where the work happened.

Within that logic, keeping a mid-level manager began to look incoherent.

If teams were empowered to plan their work, prioritize tasks, and self-correct, then what was the manager’s role — especially if they were no longer the primary business decision-maker?

The conclusion felt obvious: remove the layer.

What was quietly forgotten was that decision-making authority was only one of several functions mid-level management performed — and the easiest one to see from the outside.

The Hidden Function We Removed

Historically, mid-level managers were not merely supervisors. At their best, they performed three essential system functions:

1. Translation of context

They converted abstract strategy into concrete, local meaning. They answered the question employees still ask today: “What does success look like in my work, this week?”

2. Performance coaching

They observed work closely enough to distinguish effort from impact, learning from avoidance, growth from stagnation.

3. Informal trust repair

They absorbed tension before it escalated. They mediated misunderstandings before HR policies were triggered. They restored working relationships quietly, humanly, and early.

These functions were rarely documented, measured, or standardized — but they were real. They were the connective tissue of organizational performance.

When organizations flattened, these functions did not become unnecessary. They became orphaned.

Flattening Was a Cost Optimization, Not a Design Evolution

The first major wave of flattening began in the early 2000s, driven by globalization, ERP systems, and shareholder pressure for efficiency. The second wave accelerated sharply after COVID, when remote work made visibility harder and managerial roles easier to question.

Research consistently shows that delayering was justified economically, not systemically:

• Middle management was framed as “overhead,” not as a performance infrastructure.

• Digital dashboards replaced proximity.

• KPIs replaced judgment.

• Reporting replaced conversation.

What was removed was not bureaucracy. What was removed was organizational sense-making capacity.

Organizations optimized for headcount reduction while assuming that coordination, accountability, and trust would somehow self-organize.

They did not.

Teams Became Execution Units Without Authority or Clarity

Modern teams are told they are “empowered. ”In practice, they are often accountable without authority and measured without context.

Key decisions moved upward or sideways into opaque processes:

• Strategy is defined elsewhere.

• Metrics are chosen elsewhere.

• Trade-offs are decided elsewhere.

Yet performance conversations are expected to happen locally, between individuals who lack:

• Clear outcome definitions

• Legitimate decision rights

• Psychological safety to challenge metrics they did not design

This already creates a structural contradiction: teams are responsible for results, but not for defining what results mean.

It gets worse.

Teams are asked to define their own OKRs — but without control over budgets, staffing, or scope. They are asked to report KPIs — but without visibility into the priorities or values those KPIs are meant to serve. They are told they are “responsible for the budget” — but only in the narrow sense of tracking and explaining variance, not in the authority to reallocate, renegotiate, or redesign the work.

Responsibility is pushed downward. Authority remains centralized.

This is not empowerment. It is responsibility theater.

In the past, mid-level managers absorbed this contradiction by translating strategy, buffering constraints, and negotiating trade-offs. Once that layer disappeared, the contradiction was pushed directly onto teams and individuals.

At the same time, performance accountability was quietly split in two — and broken in the process.

Formal “people managers” are now often responsible for too many teams and too far from day-to-day work to observe real contribution. They carry nominal responsibility for performance reviews, engagement, and development — but are not responsible for business outcomes. Business leaders, meanwhile, are responsible for results — but not for the people experiencing the consequences of how those results are pursued.

The result is a new kind of managerial fiction: people managers with impressive titles who feel like strangers to the teams they formally “own.”

You can escalate issues to them, but they are too far from the work to change anything meaningful. They do not control strategy. They do not control metrics. They do not control resources.

When financial controllers decide it is time to cut costs, these managers are not accountable for the decision — only for executing it.

Performance management under this model loses its moral and operational center:

• No one close to the work can legitimately judge performance.

• No one with authority is close enough to understand it.

• Accountability flows downward; decisions flow upward.

Not because people became fragile —but because the system fragmented responsibility for results, people, and decisions into roles that no longer align.

Why Performance Conversations Feel Broken

When employees resist performance discussions today, it is often framed as a cultural or generational problem. People are said to be fragile. Unmotivated. Afraid of feedback.

The data — and lived organizational experience — suggests otherwise.

Performance management did not become difficult because people changed. It became fragile because the object of performance became unclear.

Most modern performance systems rely heavily on KPIs. But KPIs describe the success of processes, not of individual contribution.

A KPI typically reflects the combined output of systems involving:

• Multiple teams

• External partners and vendors

• Historical decisions

• Budget constraints

• Tooling quality

• Market timing

Sometimes hundreds of people influence a single KPI.

Yet performance conversations are held with individuals.

There is no reliable translation layer between the two.

Employees are asked to explain or defend outcomes they only partially influence. Managers are asked to evaluate performance they cannot causally attribute. Both sides know, intuitively, that the link is weak — but the conversation proceeds anyway.

This creates a quiet erosion of trust.

Goals feel abstract and constantly shifting, because they are derived from aggregate indicators rather than concrete work. Metrics feel disconnected from lived reality, because they summarize systems, not actions. Evaluation happens far from observation, because the people who see the work are not the ones defining success. Trust repair is formalized too late, because damage occurs long before KPIs turn red.

Sociological and organizational research is clear on this point: trust emerges from repeated, interpretable interactions, not from policies, dashboards, or post-hoc reviews.

Interpretability means being able to answer a simple question:

“I did X — and that clearly contributed to Y.”

When systems can no longer provide that line of sight, people adapt.

They stop volunteering nuance. They avoid ownership of ambiguous outcomes. They minimize exposure to metrics they do not control.

Avoidance is not disengagement. It is rational self-defense in an unreadable system.

Performance conversations feel broken not because feedback is unwelcome —but because the system can no longer explain how individual effort turns into collective value.

Why Organizations Doubled Down on KPIs Anyway

Given these limitations, the obvious question is: why did organizations lean even harder into KPIs?

Because KPIs scale where interpretation does not.

As organizations grew larger, more distributed, and more regulated, leaders faced a real constraint: they could no longer personally observe work, nor rely on informal judgment without exposing themselves to legal, financial, and reputational risk.

KPIs offered something extremely attractive:

• Apparent objectivity

• Auditability

• Comparability across teams and time

• Defensibility in front of boards, auditors, and regulators

In other words, KPIs did not replace interpretation because they were better —they replaced it because they were easier to standardize, justify, and defend.

What was lost in that trade-off was causality.

Metrics became proxies for value, then gradually substitutes for understanding. Signals meant to inform judgment became targets used to replace judgment. And once interpretation disappeared, organizations had no internal mechanism left to explain why results occurred — only whether numbers moved.

This is how performance management quietly transformed from a learning system into a reporting system.

And once performance is reduced to reporting, conversations stop being about improvement. They become about compliance, exposure, and narrative control.

That is the moment performance conversations truly break.

This Was Not a Leadership Failure — It Was a Design Failure

Blaming individual managers misses the point.

What failed was not leadership intent, competence, or commitment —it was organizational design.

We removed a layer that performed critical system functions:

• Translated strategy and values into concrete, day-to-day action

• Anchored accountability in shared, local understanding

• Repaired trust informally, before issues crossed reporting or compliance thresholds

And we replaced it with abstractions:

• Tools that measure without explaining

• Processes that report without repairing

• Systems that optimize outputs while ignoring human sense-making

Leadership did not stop caring. Managers did not suddenly become incapable.

The system simply stopped giving them a place where interpretation, judgment, and repair could happen legitimately.

Organizations did not become flatter. They became structurally deaf — unable to hear what work actually sounds like on the ground.

What This Post Is Not Arguing

This is not a call to restore old hierarchies. Mid-level management, as it existed, had real and well-documented flaws.

But removing a system function without replacing it with an equivalent mechanism is not transformation. It is amputation.

If organizations want performance, accountability, and trust to coexist again, they must deliberately rebuild the functions that were lost — translation, interpretability, and early trust repair — without nostalgically rehiring titles.

This is a design problem, not a cultural one.

It requires new operational models, not motivational language or value statements.

And that is where the real work begins.

References & Further Reading

1. Mintzberg, H. (2009). Managing. Berrett-Koehler.https://www.bkconnection.com/books/title/managing

2. Galbraith, J. R. (2014). Designing Organizations: Strategy, Structure, and Process at the Business Unit and Enterprise Levels. Jossey-Bass.https://www.wiley.com/en-us/Designing+Organizations%2C+3rd+Edition-p-9781118409955

3. Womack, J. P., & Jones, D. T. (2003). Lean Thinking. Simon & Schuster.https://www.lean.org/WhoWeAre/NewsArticleDocuments/Lean_Thinking.pdf

4. Edmondson, A. (2018). The Fearless Organization. Wiley.https://www.wiley.com/en-us/The+Fearless+Organization-p-9781119477243

5. Gulati, R., Puranam, P., & Tushman, M. (2012). “Meta-organization design: Rethinking design in interorganizational and community contexts.” Strategic Management Journal. Meta‐organization design: Rethinking design in interorganizational and community contexts 

6. OECD (2017). Trust and Public Policy: How Better Governance Can Help Rebuild Public Trust. Trust and Public Policy 

7. Bernstein, E., Bunch, J., Canner, N., & Lee, M. (2016). “Beyond the Holacracy Hype.” Harvard Business Review. Beyond the Holacracy Hype