14  Agile Organizational Structures

ImportantLearning Objectives

You will be able to:

  1. Evaluate flat hierarchies: what delayering achieves, what it cannot, and the hidden structures that emerge when formal ones are removed.
  2. Design a squad-and-tribe structure, using chapters and guilds to balance mission focus against disciplinary depth.
  3. Explain the team-of-teams network model and the shared-consciousness mechanisms that make it work.
  4. Select among structural options, including hybrids and Kotter’s dual operating system, for a given organization’s work portfolio.

14.1 Introduction

Chapter 13 established what agile organizations are for; this chapter surveys the structural repertoire through which the design is realized. Four families dominate practice, arranged roughly by how far each departs from the classic pyramid. Flat hierarchies keep the pyramid but compress it, removing layers and widening spans. Squad-and-tribe structures replace functional departments with mission teams, then weave discipline back in horizontally. Networked team-of-teams models connect autonomous units through shared consciousness rather than reporting lines. Self-managing structures, glimpsed at Buurtzorg and Haier in Chapter 13, distribute even the residual managerial role into the teams themselves (Frederic Laloux, 2022).

No family is the destination; each is a tool with a price. The recurring analytical thread, inherited from Chapter 3’s Spotify warning and Chapter 11’s delegation test, is that structure names are cheap and decision rights are dear: two organizations with identical charts can sit at opposite ends of the agility spectrum depending on what their teams may actually decide. Read every structure in this chapter as a bundle of decision rights and coordination mechanisms wearing a diagram.

flowchart LR
    P["Classic pyramid"] --> F["Flat hierarchy:<br>fewer layers,<br>wider spans"]
    F --> S["Squads and tribes:<br>mission teams plus<br>chapters and guilds"]
    S --> N["Team of teams:<br>networked autonomy,<br>shared consciousness"]
    N --> SM["Self-management:<br>authority fully<br>in the teams"]
    style P fill:#ffebee,stroke:#C62828
    style F fill:#fff8e1,stroke:#F9A825
    style S fill:#e3f2fd,stroke:#1976D2
    style N fill:#e8f5e9,stroke:#388E3C
    style SM fill:#ede7f6,stroke:#7E57C2

14.2 Flat Hierarchies

14.2.1 What Delayering Buys

Flattening removes management layers and widens spans of control, and its benefits follow directly from Chapter 13’s cost analysis of hierarchy: fewer levels for information to degrade through, fewer stations at which decisions queue, lower coordination overhead, and managers who, with twelve reports instead of four, physically cannot micromanage and must lead through the alignment mechanisms of Chapter 13 instead. Flattening also changes the career geometry HR administers: with fewer rungs, progression must be redefined around growth in scope and mastery rather than accumulation of titles, which is why delayering programmes fail in HR before they fail anywhere else, when reward and status systems still price altitude (Natal Dank & Riina Hellström, 2020).

14.2.2 What Delayering Cannot Do

Flatness is a dosage question, not a destination. Wider spans thin the coaching relationship Chapter 9 made central, one manager cannot hold weekly check-ins with twenty-five people, and removing formal structure does not remove structure: informal hierarchies of tenure, charisma, and access form in the vacuum, harder to see and impossible to hold accountable. The self-management literature itself is explicit that flat organizations succeed by replacing management with explicit mechanisms, advice processes, conflict-resolution rituals, transparent information, not by deleting management and hoping (Frederic Laloux, 2022). A useful discipline: for every layer removed, name the mechanism that now performs the layer’s real functions, prioritization, development, integration, or expect the organization to grow a shadow version.

WarningCommon Misconception: Flatter Is Always More Agile

The relationship is curvilinear. Extreme flatness in a large organization starves people of development attention and starves the enterprise of integration, producing either drift or the covert re-emergence of hierarchy. The agile design goal is the fewest layers consistent with real coaching and real integration, typically far fewer than a machine bureaucracy carries, but rarely zero, and always paired with the lateral structures the rest of this chapter supplies (Darrell K. Rigby et al., 2020).

14.3 Squads, Tribes, Chapters, and Guilds

14.3.1 The Two-Axis Design

The squad-and-tribe family, popularized by Spotify and adopted at scale by ING and ANZ in earlier chapters, solves a specific dilemma: mission focus versus disciplinary depth. Organize purely by function and you get depth with hand-offs; organize purely by mission and specialists scatter, craft standards drift, and nobody develops the recruiters. The model runs both axes at once. Squads, Chapter 11’s teams, own missions. Tribes group squads working related missions, small enough, conventionally under roughly 150 people, for relationships to carry coordination. Chapters are the horizontal homes: all members of one discipline across a tribe, led by a chapter lead who develops them, sets craft standards, and typically holds the line-management relationship, while their daily work is directed by squad priorities. Guilds are voluntary communities of interest spanning the whole organization, lightweight carriers of knowledge that Chapter 17 develops (Stephen Denning, 2018).

For HR the model is doubly relevant: as the structure client organizations adopt, with the role-redesign consequences ANZ illustrated, and as a structure for HR itself, people-experience squads owning journeys such as joining or growing, an HR tribe coordinating them, chapters maintaining depth in analytics, talent acquisition, and employment law (Natal Dank & Riina Hellström, 2020).

NoteThe Squad Model’s Components and Their HR Analogues
Component Axis Function HR-function example
Squad Vertical (mission) Owns an outcome end to end Onboarding-journey squad
Tribe Vertical (mission cluster) Aligns related squads, allocates people People-experience tribe
Chapter Horizontal (discipline) Craft depth, standards, line care People-analytics chapter
Guild Horizontal (interest) Voluntary knowledge exchange Facilitation guild across the firm

14.3.2 Matrix Without the Misery

Structurally this is a matrix, mission axis and discipline axis, and classic matrix organizations earned their bad reputation through dual bosses issuing conflicting work orders. The squad model’s refinement is a clean split of the two managerial questions: the squad (through its product owner and backlog) answers what you work on; the chapter lead answers how you grow and how well you work. When the split is honoured, the matrix’s chronic conflict has no channel; when chapter leads start directing daily work or squads start setting pay, the old misery returns wearing new names. The split is, once again, a decision-rights specification, and HR is its registrar: role descriptions, appraisal inputs in the Chapter 9 manner, and reward authority must be written to match (Natal Dank & Riina Hellström, 2020).

14.4 Networked Teams and the Team of Teams

14.4.1 Shared Consciousness, Empowered Execution

The third family generalizes beyond tribes: a team of teams, in which many autonomous units coordinate as a network. Its canonical account comes from General Stanley McChrystal’s Joint Special Operations Task Force in Iraq, which found that a superbly efficient military hierarchy was losing to a networked adversary that adapted faster than orders could travel. The redesign replaced need-to-know silos with radical information sharing, a daily all-hands briefing eventually spanning thousands of participants, embedded liaisons between units, and decision authority pushed to the teams in contact with events, a combination McChrystal summarizes as shared consciousness plus empowered execution (Stanley McChrystal et al., 2015). The formula restates Chapter 13’s law: transparency wide enough that autonomy is safe.

The corporate translation is direct. Shared consciousness is built by the transparency infrastructure this book has assembled, open backlogs, visible OKRs, listening data returned to teams, and by synchronization rituals at network scale: the quarterly business review where tribes show work and dependencies, the demo day open to all. Empowered execution is Chapter 3’s delegated decision rights, now with an explicit condition attached: the more context teams share, the more authority can safely sit with them, and attempts to empower without informing produce the chaos that hierarchy’s defenders predict (Stanley McChrystal et al., 2015).

flowchart TD
    SC["Shared consciousness:<br>radical transparency,<br>synchronization rituals,<br>liaisons"] --> EE["Empowered execution:<br>decisions at the edge,<br>within intent"]
    EE --> AD["Faster adaptation<br>than any escalation<br>channel allows"]
    AD -->|"results and learning<br>shared back"| SC
    style SC fill:#e3f2fd,stroke:#1976D2
    style EE fill:#e8f5e9,stroke:#388E3C
    style AD fill:#fff8e1,stroke:#F9A825

14.4.2 Hybrids and the Dual Operating System

Most enterprises cannot and should not convert wholesale, the contingency verdict of Chapters 4 and 13. John P. Kotter (2014) offers the transitional architecture: a dual operating system, in which the managerial hierarchy continues to run reliability-critical operations while a volunteer-powered network of initiative teams, drawn from the same workforce, drives strategy and innovation at speed, the two connected through people who live in both. Kotter’s insight is that the network is not a programme office but a second circulatory system, and that hierarchies do not need to be demolished to regain the agility they had as startups, they need the network side deliberately built and legitimated. For organizations early in transformation, the dual system is often the honest first structure: agile islands with real decision rights, connected to a stable core, expanding as evidence accumulates (John P. Kotter, 2014; Darrell K. Rigby et al., 2020).

TipPractitioner Insight: Sequence Structure Last, Not First

The tempting transformation sequence, announce the new chart, then hope behaviour follows, is backwards. Structures formalize ways of working; they cannot create them. Run the Part II practices inside the existing structure first: form a few genuine cross-functional teams, give them real backlogs and real decision rights, let chapters emerge from the specialists who support them. When the informal network is visibly working, the reorganization ratifies reality instead of gambling on it, and the ANZ-style costs of big-bang restructuring shrink accordingly.

14.5 Case Studies

14.5.1 Case Study 1: W.L. Gore, The Lattice That Predates the Vocabulary

W.L. Gore & Associates, maker of Gore-Tex, has run since 1958 on what founder Bill Gore called a lattice organization: no conventional job titles or bosses, small plants deliberately capped near 150 to 200 people so everyone can know everyone, leaders defined as those whom others choose to follow, sponsors, not managers, responsible for each associate’s development, and commitments made voluntarily rather than assigned. Compensation is set by committees informed by peer rankings of contribution. Gore has been profitable for decades, is a fixture of best-workplace lists, and produced category-defining innovations, from waterproof fabrics to Elixir guitar strings, from associates’ self-initiated projects. The case matters because it shows the network design’s essential features, natural leadership, peer accountability, capped unit size, working half a century before agile gave them names.

Discussion Questions:

  1. Gore caps plant size in the region of Dunbar’s number. Which coordination mechanisms in this chapter does small size substitute for, and what does the cap cost?
  2. Peer-ranked contribution sets pay at Gore. Evaluate this against Chapter 9’s warnings about wiring peer input to reward: why might it work here?
  3. Leadership at Gore is followership-defined. What would HR’s leadership-pipeline work (succession, promotion) look like in such a system?

14.5.2 Case Study 2: Morning Star, Self-Management by Contract

Morning Star, the California company processing a large share of the United States tomato crop, operates with no managers at all. Each colleague annually writes a personal mission statement and negotiates a Colleague Letter of Understanding (CLOU) with the co-workers their work most affects, specifying commitments, metrics, and stepping-stones, in effect, a peer-to-peer contract replacing the job description and the boss. Decisions of any size follow an advice process: anyone may decide after consulting those affected and those with expertise. Conflicts escalate not up a hierarchy but through a defined peer mediation ladder. Frederic Laloux (2022) documents Morning Star as proof that self-management can run heavy, capital-intensive, seasonal manufacturing, not merely boutique knowledge work, while insiders note the model’s demands: negotiating CLOUs is intense social work, and colleagues who prefer being told what to do tend to leave.

Discussion Questions:

  1. The CLOU replaces both job description and manager. Which of the deleted layer’s functions, from this chapter’s delayering discipline, does it perform, and which does it leave unowned?
  2. Compare Morning Star’s advice process with a squad’s delegated decision rights. Which scales more easily, and why?
  3. Design induction for a company where the first task is negotiating your own accountabilities. What capabilities must HR build in every joiner?

14.6 Summary

NoteChapter Summary

Agile structure is a repertoire, not a template. Flat hierarchies buy speed and information fidelity by removing layers, but flatness is curvilinear: each deleted layer’s real functions must be reassigned to explicit mechanisms or shadow hierarchies grow (Frederic Laloux, 2022). The squad-tribe-chapter-guild model runs mission and discipline as simultaneous axes, its stability resting on a clean decision-rights split between what squads direct and how chapters develop (Natal Dank & Riina Hellström, 2020; Stephen Denning, 2018). The team-of-teams network scales autonomy through McChrystal’s formula, shared consciousness plus empowered execution, transparency wide enough to make edge decisions safe (Stanley McChrystal et al., 2015). Kotter’s dual operating system offers the transitional hybrid: a legitimated innovation network alongside a reliability hierarchy (John P. Kotter, 2014). Gore and Morning Star prove the deep end is habitable, and profitable, given decades of consistent mechanism-building. Structures, however, only formalize behaviour; Chapter 15 turns to the culture that determines whether any of these charts comes alive.

TipKey Terms

Flat hierarchy · Span of control · Delayering · Shadow hierarchy · Squad · Tribe · Chapter · Guild · Two-axis matrix · Team of teams · Shared consciousness · Empowered execution · Dual operating system · Lattice organization · Advice process

Summary

Concept Description
Flat Hierarchies
Structural repertoire Flat, squad-based, networked, and self-managing families, each a priced tool rather than a destination
Flat hierarchy A compressed pyramid with fewer layers and wider spans of control
Delayering dividend Less information degradation, fewer decision queues, and managers forced to lead by alignment
Career geometry Redefining progression as scope and mastery once title rungs are removed
Curvilinear flatness The finding that agility rises with flattening only until coaching and integration starve
Shadow hierarchy Informal structures of tenure and access that fill the vacuum left by deleted formal ones
Layer-replacement discipline Naming the explicit mechanism that inherits each removed layer's real functions
Squads and Tribes
Squad A small cross-functional team owning a mission end to end
Tribe A cluster of related squads small enough for relationships to carry coordination
Chapter The horizontal disciplinary home providing craft standards, development, and line care
Guild A voluntary, organization-wide community of interest carrying knowledge across boundaries
What-how split Squads deciding what is worked on while chapters develop how well people work
Matrix conflict channel The dual-boss conflict that returns whenever the what-how split is violated
Networks and Hybrids
Team of teams Many autonomous units coordinating as a network rather than through reporting lines
Shared consciousness Radical transparency and synchronization giving every team the context of the whole
Empowered execution Decision authority at the edge, made safe by the breadth of shared context
Synchronization rituals Network-scale rhythms such as open demo days and quarterly reviews of work and dependencies
Dual operating system Kotter's hierarchy-plus-network hybrid for driving change while running operations
Sequence structure last Building working practices first so reorganization ratifies reality rather than gambling
Case Evidence
Gore lattice Gore's title-free, sponsor-based, size-capped organization running since 1958
Morning Star CLOU Morning Star's peer-negotiated contracts replacing job descriptions and managers