Deep divePeople & Organizational Performance

Spans and layers analysis: measuring structure and redesigning it safely

Spans of control and management layers are the two figures most often used to describe an organization's structure, and the two most often misused. This deep dive explains how to compute them from HR data, how to read the patterns, why the right span depends on what a manager actually does, how AI tools are changing managerial work, and how to redesign without breaking accountability or losing expertise.

Reviewed 7 min read

On this page
  1. Span of control, layers and the terms around them
  2. Why structure shapes both speed and cost
  3. Computing spans and layers from HR system data
  4. Manager archetypes and the span each can sustain
  5. Reading the results: common patterns and what they usually signal
  6. What AI tools change in the work of managing
  7. A redesign sequence that keeps accountability intact
  8. People risks when removing management layers
  9. Hypothetical example: flattening a regional operations function
  10. Questions and answers
  11. Sources

Span of control, layers and the terms around them

This page is about reporting structure. Where AI work itself should sit, centrally or in business units, is a separate choice covered in AI operating models.

Span of control
The number of people reporting directly to a manager on the solid line. Count positions, including vacancies being filled, rather than names.
Layers
The number of reporting steps between the chief executive and the frontline. Measure it per chain, then summarize by function, because averages hide very deep pockets.
Single-report manager
A manager with exactly one direct report. Sometimes justified for a deputy or a specialist pair; often a sign that a management title was used as a reward.
Player-coach
A manager who also carries substantial individual delivery work, which limits the time available for each report.
Dotted-line relationship
A secondary reporting relationship, common in matrix organizations, that should be recorded but left out of the primary span.

Why structure shapes both speed and cost

Every layer adds a handoff. Decisions travel up for approval and back down for execution, information is summarized at each step, and the frontline waits. Deep structures therefore tend to be slow, and they often leave the people at the top feeling well informed while seeing a filtered picture.

Narrow spans add cost directly, through management salaries, and indirectly, through close supervision that leaves capable people little room to decide. Very wide spans fail the other way: managers who cannot coach, review work or notice problems early. The aim is not the widest possible span but a structure in which each manager has enough people to justify the role and few enough to lead well.

Computing spans and layers from HR system data

  1. Extract positions, not just people

    Pull every position with its manager position, including vacancies, interim cover and contractors who manage employees. People-only extracts undercount spans in teams that are mid-hiring.

  2. Resolve the reporting tree

    Treat the solid line as primary, store dotted lines separately, and fix loops and orphaned positions before computing anything. Every position should trace up to the chief executive.

  3. Compute per manager and per chain

    Calculate each manager's direct span and each position's layer, then the depth of the longest chain in each function.

  4. Tag manager archetypes

    Classify each manager by the work they do, using the archetypes below, since the archetype determines which span is reasonable.

  5. Validate with leaders

    Walk each function's leader through the results. HR data is often out of date, and a short review catches moves, informal teams and mislabeled roles before anyone draws conclusions.

Manager archetypes and the span each can sustain

ArchetypeNature of the workWhat limits the spanSpan tends to be
Player-coachDelivers individual work alongside managingTime left after their own deliveryNarrower
Specialist leadLeads experts on complex, varied problemsThe depth of review each piece of work needsNarrow to moderate
Operational supervisorRuns a team doing similar, well-defined workScheduling, exceptions and quality checksWider, especially with good tooling
Manager of managersCoordinates other managers and sets directionCoordination across units and decisions escalated to themModerate
Project or portfolio leadLeads temporary teams drawn from several functionsThe number of parallel workstreamsVaries with the portfolio

These are directions, not benchmarks. Group managers by archetype before comparing spans, so a specialist lead is never judged against a call-center supervisor.

Reading the results: common patterns and what they usually signal

  • If

    One function has many single-report managers.

    Then

    Check whether management titles stand in for a missing specialist career path, and create that path before merging teams.

    Removing the title without an alternative loses the people it was meant to keep.

  • If

    A small function has a deep chain.

    Then

    Collapse intermediate layers that only pass information through, and reassign their decisions.

    Layers that neither decide nor coach add delay without adding control.

  • If

    Senior specialists have narrow spans.

    Then

    Leave them if the review burden is real; widen them only if the leads are mainly coordinating.

    Expert review is expensive to dilute.

  • If

    Spans are very wide and escalations are frequent.

    Then

    Add team leads or tooling for scheduling and reporting before adding a management layer.

    Escalations often come from coordination load rather than missing hierarchy.

What AI tools change in the work of managing

Much of an operational manager's week goes on coordination: compiling status reports, chasing updates, building schedules, routing requests and preparing summaries for the next level up. These tasks suit AI assistance, and as they shrink, some managers can lead more people without losing quality. The effect is uneven. Operational supervisors gain most; player-coaches gain less, because their limit is their own delivery work.

Coaching, judgment about people, conflict resolution and accountability for outcomes stay human. Be careful with tools that evaluate the people themselves: AI used to monitor and evaluate workers' performance or behavior is a high-risk use under the EU AI Act, and inferring employees' emotions is prohibited outside medical or safety purposes1. Analyze managerial roles with the same task-level AI impact assessment used for any other role.

A redesign sequence that keeps accountability intact

01Design principles02Decision rights03Structure options04People and roles05Transition plan06Post-change review
  1. Design principles

    Agree what the structure must optimize for and which constraints are fixed.

  2. Decision rights

    Decide who decides what before drawing boxes, so every decision has one owner.

  3. Structure options

    Model a few alternatives against the principles, not against a target span.

  4. People and roles

    Match people to new roles openly, using criteria agreed in advance.

  5. Transition plan

    Sequence moves so handovers finish before old roles close.

  6. Post-change review

    Check after a few months whether decisions are faster and managers are coping.

Conceptual sequence for redesigning spans and layers. It shows dependencies between stages, not durations or results.

People risks when removing management layers

Skipping consultation

Early signalRole changes are announced before employee representatives hear about them.

MitigationCheck obligations early. In the EU, the collective redundancies directive (Directive 98/59/EC) requires consultation once planned dismissals reach set thresholds2. In Germany, outside the smallest companies, fundamental changes to an establishment's organization that may significantly disadvantage staff require informing and consulting the works council3.

Expertise leaving with the layer

Early signalExperienced managers resign because the only alternative offered is a lower title.

MitigationOffer expert or principal roles with real scope, and capture knowledge that sits only with departing managers.

Change by spreadsheet

Early signalThe new structure was drawn from span data without anyone describing the work.

MitigationValidate every option with the leaders who will run it, and test it against real decisions and workloads.

Overloaded remaining managers

Early signalSpans widen while reporting, approvals and administration stay the same.

MitigationRemove or automate coordination work before or alongside the widening, not after complaints arrive.

Hypothetical example: flattening a regional operations function

Questions and answers

How often should spans and layers be reviewed?

Run a light refresh from HR data every year, and a full review when strategy, a major system or the business mix changes. Spans drift quietly as teams grow, managers leave and interim arrangements become permanent, so a regular extract catches problems before they harden. Keep the same archetype tags between reviews so you compare like with like.

Should we set a target span of control?

Set ranges by archetype, if you set anything. A single company-wide target pushes specialist teams into unnatural merges and lets oversized operational teams look acceptable. Ranges agreed per archetype, with a rule that exceptions are explained rather than banned, give leaders guidance without forcing every function into the same shape.

How should specialists without direct reports be treated in the analysis?

Leave them out of span calculations but keep them in the layer count as individual contributors, so you can see where senior expertise sits. If specialists are given reports mainly to justify their grade, the fix is a parallel career track with comparable recognition, not a different way of counting.

Does removing management layers always reduce cost?

Not reliably. Savings from fewer manager roles can be absorbed by higher pay for wider roles, severance, lost expertise and slower work while new teams settle. Treat cost as one outcome among several, and judge the redesign mainly on decision speed, clarity of accountability and whether managers can still support their people.

Sources

  1. Regulation (EU) 2024/1689 (Artificial Intelligence Act), Article 5 and Annex III — EUR-Lex · checked 10 October 2026
  2. Council Directive 98/59/EC on the approximation of the laws of the Member States relating to collective redundancies — EUR-Lex · checked 10 October 2026
  3. Betriebsverfassungsgesetz (Works Constitution Act), § 111: operational changes — Federal Ministry of Justice (Germany) · checked 10 October 2026

More in People & Organizational Performance

Back to People & Organizational Performance

Next step

Get a second view on your spans and layers data

Share an anonymized reporting-line extract and the question behind it. We will point out data issues to fix first and how to group managers before comparing spans.

Discuss an organization review