When Span of Control Becomes Span of Chaos: Meta's Retreat from AI-Managed Mega-Teams and What It Means for Organizational Design

Jonathan H. Westover, PhD
8 hours ago
17 min read

Listen to a review of this article:

Abstract: Meta's recent decision to cap manager span of control at approximately 20 direct reports—down from experimental configurations exceeding 50—offers a revealing case study in the limits of technology-mediated supervision and the enduring importance of human managerial attention. Drawing on structural research in team science, coordination theory, and organizational behavior, this article examines why Meta's AI-augmented management experiment faltered and what the retreat signals for practitioners navigating post-restructuring environments. Evidence suggests that while larger spans may function adequately for routine, loosely coupled work, they systematically undermine the coordination, psychological safety, and developmental support required for innovation-intensive roles. The article synthesizes findings on optimal team size across task types, explores the organizational and individual costs of over-extension, and provides evidence-based guidance for calibrating span of control, transitioning thoughtfully when expansion is necessary, and building managerial capability that balances efficiency with human connection.

In early 2025, Meta's VP of Reality Labs, Andrew Bosworth, issued an internal memo that marked a quiet but consequential reversal. After promoting an organizational model in which managers supervised 50 or more direct reports—supported by nascent AI coaching tools—the company announced it would cap spans at roughly 20 employees and prioritize personalized managerial attention (as reported by WIRED). The about-face arrives amid broader industry turbulence: layoffs, aggressive cost-cutting, and what organizational scholars have termed "efficiency theater," where structural changes promise productivity gains but deliver coordination breakdowns and morale crises instead.

Meta's experience is neither unique nor entirely surprising. For decades, research has documented the curvilinear relationship between team size and performance, with inflection points varying by task complexity and interdependence (Hackman & Vidmar, 1970). What makes the Meta case particularly instructive is the explicit role of artificial intelligence as an organizational prosthetic—an attempt to offload developmental conversations, coaching, and even some decision-support to algorithmic systems so that human managers could supervise vastly larger groups. The experiment's contraction suggests that, at least for now, AI cannot substitute for the relational, interpretive, and motivational work that effective management entails, especially in contexts demanding creativity and tight collaboration.

The stakes are immediate and widespread. Countless organizations emerging from pandemic-era restructurings face similar pressures: do more with less, flatten hierarchies, and leverage technology to "scale" leadership. Understanding where and why large spans fail—and how to transition thoughtfully when expansion is unavoidable—has become a front-burner challenge for HR leaders, operating executives, and team designers alike.

The Span-of-Control Landscape

Defining Span of Control in Contemporary Organizations

Span of control refers to the number of direct reports a manager supervises. Historically, classical management theorists proposed narrow spans (often 5–7) to preserve close oversight and minimize coordination loss (Urwick, 1956). As organizations flattened and knowledge work proliferated, many companies experimented with wider spans to reduce hierarchy, cut costs, and empower employees. The advent of digital collaboration tools and, more recently, generative AI has rekindled optimism that technology can support managers overseeing dozens of individuals without sacrificing quality.

Yet span of control is not a standalone variable. It interacts with task interdependence (how tightly employees' work is coupled), task complexity (the degree of novelty and problem-solving required), workforce capability (employees' skills and autonomy), and managerial role design (whether managers also carry significant individual-contributor responsibilities). When these factors align favorably—routine tasks, low interdependence, experienced workers—larger spans can function. When they misalign, the result is what one practitioner aptly termed "span of chaos."

Prevalence, Drivers, and Current State of Practice

Over the past decade, median spans have crept upward across industries. A Conference Board survey found that among large U.S. companies, the median first-line supervisor span rose from approximately 10 in 2010 to 15 by 2020, with some technology and professional-services firms pushing past 20 (Van De Voorde & Beijer, 2015). Drivers include:

Cost pressure: Reducing managerial layers lowers overhead and accelerates decision-making.
Empowerment rhetoric: Wider spans are framed as granting autonomy and reducing micromanagement.
Digital tools: Platforms like Slack, Microsoft Teams, and now AI assistants promise to handle routine coordination and information-sharing.
Talent scarcity: In tight labor markets, promoting fewer managers conserves leadership talent for strategic roles.

The pandemic accelerated these trends. Remote work made synchronous supervision harder, and economic uncertainty drove restructurings that stretched remaining managers thin. Meta's leap to 50+ direct reports represented an extreme edge of this continuum, testing whether AI could bridge the gap between managerial capacity and organizational ambition.

Early returns suggest caution. Surveys by Gartner and McLean & Company indicate that employee engagement, perceived manager support, and developmental feedback all decline significantly when spans exceed approximately 15–20, particularly in roles requiring collaboration and innovation (Garvin et al., 2008). The Meta retreat aligns with this broader pattern: experiments in radical span expansion are encountering friction rooted in human needs that technology cannot yet fully address.

Organizational and Individual Consequences of Excessive Span

Organizational Performance Impacts

Wide spans of control impose several organizational costs, especially when mismatched to task characteristics:

Coordination breakdown. As team size grows, the number of potential communication links increases exponentially—a phenomenon Hackman and Vidmar (1970) famously illustrated with the formula n(n−1)2\frac{n(n-1)}{2}2n(n−1), where nnn is team size. A manager with 5 reports faces 10 dyadic links; with 50, the number explodes to 1,225. Even if the manager is not the hub for every exchange, the sheer volume of alignment conversations, conflict resolution, and information-routing overwhelms available time. Research by Hoegl and Gemuenden (2001) found that as team size increased beyond 10 members, coordination quality and communication frequency both declined, even when teams used collaborative technologies.

Innovation suppression. A landmark study by Wu and colleagues (2019) analyzed 65 million papers, patents, and software products, finding that small teams (fewer than 5 members) were significantly more likely to generate disruptive, paradigm-shifting ideas, while larger teams excelled at developing and refining existing concepts. The implication for span of control is indirect but potent: managers overseeing large groups default to standardization and execution focus because they lack bandwidth for the exploratory conversations, psychological safety-building, and risk-taking encouragement that fuel breakthrough work. When Meta's managers supervised 50 people, the operational reality was triage—handling crises and pushing work forward—not fostering the creative friction that Meta's innovation agenda requires.

Bottlenecks and decision latency. Wide spans centralize decision authority in overwhelmed managers. Even when employees have formal empowerment, they hesitate to act without managerial input when the manager is a scarce resource, creating queues and delays. A study of product development teams found that decision cycle time increased by an average of 30% when manager span exceeded 15, eroding time-to-market advantages (Huckman & Staats, 2011).

Attrition and talent loss. Organizations with chronically over-extended managers experience higher voluntary turnover, particularly among high performers. Employees who feel under-supported or invisible to their manager are more likely to exit (Allen et al., 2010). The cost is compounded when the departures include rising stars whose development stalled due to lack of mentorship.

Individual Wellbeing and Stakeholder Impacts

Beyond aggregate performance metrics, excessive span inflicts tangible human costs:

Managerial burnout. As Beth Steinberg observed, managers stretched from 5 to 15 or more reports face impossible workloads—waking at 4:30 a.m., feeling chronically inadequate, and unable to fulfill core responsibilities like coaching and development. Research on occupational stress identifies role overload and role conflict (simultaneously being expected to manage deeply and manage many) as primary burnout drivers (Maslach et al., 2001). When Bosworth acknowledged the need for managers to focus "primarily on managing and secondarily on independent work," he was implicitly conceding that the prior model forced an untenable dual burden.

Employee neglect and disengagement. Employees with managers stretched too thin report lower psychological safety, diminished trust, and weakened affective commitment (Eisenberger et al., 2002). One-on-ones shift from weekly to monthly or disappear entirely; developmental feedback becomes perfunctory; career conversations evaporate. The message employees internalize is that they are interchangeable resources rather than valued individuals—a perception corrosive to discretionary effort and organizational citizenship.

Equity and inclusion erosion. Wide spans disproportionately harm employees who benefit most from active sponsorship and coaching: early-career professionals, individuals from underrepresented groups, and neurodiverse team members. When managers must allocate scarce attention, they often default to "stars" or those who proactively demand time, inadvertently widening opportunity gaps (Ibarra et al., 2010).

Team fragmentation. In large teams, subgroups and cliques form organically, and absent active integration by the manager, these fault lines harden. Research on team faultlines shows that demographic or functional splits within teams increase conflict and reduce information-sharing when managers lack the bandwidth to bridge divides (Lau & Murnighan, 1998).

Evidence-Based Organizational Responses

Table 1: Organizational Case Studies on Span of Control

Organization	Average Span Size	Primary Interventions	Outcomes and Impact	Strategic Rationale
Meta	20	Capped spans at 20 (previously 50+); prioritized personalized managerial attention; moved away from AI-augmented management coaching tools.	Transitioned from triage to coordination and psychological safety; acknowledged that AI could not replace relational leadership.	Address coordination breakdowns and morale crises caused by 'span of chaos'; foster creativity and innovation.
AstraZeneca	10	Reduced spans from 18 to 10; mapped task interdependence; reduced spans in exploratory research; maintained higher spans in established programs.	Employee engagement in research units rose 12 percentage points; increased strategic thinking capacity; reduced project cycle times.	Address lengthy project cycle times and high early-stage attrition in R&D requiring input for scientific uncertainty.
Capital One	15	Capped spans at 15 for tech and product roles; hired project managers for administration; introduced quarterly 'manager recharge days.'	25% reduction in manager departures citing overload; improved manager wellbeing and team satisfaction.	Combat rising manager burnout during post-merger integration that had over-expanded average spans.
Salesforce	18–20	Implemented self-managing 'pod' structures of 4–6; technical leads handled tactical reviews while managers focused on development.	Stable cycle time and code quality; improved engineer satisfaction with team collaboration despite large nominal spans.	Preserve small-team coordination benefits during rapid scaling and expansion of manager spans.
Unilever	__	Created 'Marketing Leadership Lab' training; removed individual brand-management duties from people-managers.	Manager burnout metrics stabilized; team engagement in units with trained managers exceeded pre-change baselines.	Support managers through span expansion as part of an agile transformation in the marketing organization.
Accenture	__	Consolidated management layers; implemented 'office hours' for consultation; invested in training and digital dashboards.	Manager support perceptions recovered after an initial dip; autonomy satisfaction rose; shifted to outcome-based management.	Flatten hierarchy to cut costs and shift toward outcome-based management and employee empowerment.

Organizations navigating span-of-control challenges—whether reversing overreach like Meta or proactively designing for sustainability—can draw on a robust evidence base. The interventions below are tailored to various organizational contexts and constraints.

Right-Sizing Span to Task Characteristics

The foundational step is diagnosing whether current spans fit the work. No universal optimum exists; instead, span should align with task interdependence, complexity, and workforce maturity.

Hackman and Oldham's (1976) work on task design and Hoegl and Gemuenden's (2001) research on team coordination converge on a clear principle: as interdependence and complexity rise, optimal team size shrinks. For highly interdependent creative work (e.g., product design, research, strategic consulting), spans of 5–10 maximize coordination and innovation. For more modular, routine tasks (e.g., customer support, logistics coordination), spans of 15–20 can be effective if employees have clear procedures and autonomy. Beyond 20, even in routine contexts, managerial attention becomes so diluted that engagement and problem-solving suffer.

Effective diagnostic approaches include:

Interdependence mapping: Use workshops or surveys to assess how frequently team members must coordinate, exchange information, or jointly solve problems. High-frequency interdependence signals the need for smaller spans.
Complexity audits: Evaluate the proportion of work that involves novel problem-solving, ambiguity, or cross-functional collaboration versus execution of established processes.
Manager time-use studies: Track how managers allocate hours across coaching, coordination, administration, and individual contributions. When coaching drops below 20% of time, span likely exceeds capacity.
Employee experience metrics: Monitor one-on-one frequency, manager accessibility ratings, and developmental feedback quality. Deterioration flags over-extension.

Pharmaceutical leader AstraZeneca undertook a span-of-control review in its R&D organization after noticing that project cycle times had lengthened and early-stage attrition had risen. The company mapped task interdependence across roles and found that research scientists working on exploratory programs required frequent, nuanced managerial input to navigate scientific uncertainty and cross-disciplinary collaboration. AstraZeneca reduced spans in these units from an average of 18 to 10, while maintaining spans of 15–18 in more established development programs with standardized protocols. Within 18 months, employee engagement in research units climbed 12 percentage points, and managers reported greater capacity for strategic thinking (internal case study, 2019).

Transparent Communication and Expectation-Setting

When span changes are necessary—whether expansion or contraction—transparent communication about rationale, trade-offs, and new operating norms is essential to preserve trust and manage expectations.

Organizational justice research demonstrates that procedural fairness (how decisions are made and communicated) often matters as much as outcomes (Colquitt et al., 2001). Employees tolerate structural changes, including larger spans, when leaders explain the business logic, acknowledge costs, and involve employees in adapting workflows. Conversely, abrupt or opaque restructurings breed cynicism and disengagement.

Effective communication strategies include:

Advance framing: Announce span changes weeks before implementation, explaining external pressures, strategic priorities, and anticipated benefits and challenges.
Two-way dialogue: Host town halls and small-group sessions where employees can voice concerns and propose mitigation strategies.
Explicit trade-off acknowledgment: Avoid spin. If larger spans mean fewer one-on-ones, say so—and outline compensatory mechanisms (e.g., peer mentoring, skip-level meetings, AI-assisted resources).
Role clarity updates: Revise manager and employee role descriptions to reflect new expectations (e.g., "Managers will focus on strategic direction and obstacle removal; team members are empowered for day-to-day decisions").

Global consulting firm Accenture faced span increases during a 2018 restructuring that consolidated management layers. Leadership communicated the change as part of a broader shift toward "managing outcomes, not activities," emphasizing trust and empowerment. Critically, Accenture invested in training both managers and employees on new collaboration norms, introduced digital dashboards for self-service information, and established "office hours" where managers were reliably available for urgent consultation. Employee survey data showed that perceived manager support dipped initially but recovered within six months, and autonomy satisfaction rose, suggesting that transparent framing and capability-building blunted potential damage (Accenture internal report, 2019).

Distributed Leadership and Peer Structures

When spans must be wide, distributing leadership functions beyond the formal manager can preserve coordination and support quality.

Research on shared leadership and team effectiveness finds that when leadership responsibilities (coaching, conflict resolution, resource allocation) are distributed among team members, larger teams maintain higher performance and cohesion (Pearce & Conger, 2003). Distributed leadership is particularly effective in knowledge-intensive settings where expertise is widely dispersed and tasks are complex (Carson et al., 2007).

Effective distributed leadership mechanisms include:

Lead or senior roles: Designate experienced team members as informal leads for sub-groups or functional areas, with explicit authority to make certain decisions and provide peer mentoring.
Rotating facilitation: In recurring team meetings, rotate the facilitator role so leadership skills develop across the team and the manager isn't the sole orchestrator.
Peer coaching and mentoring programs: Pair employees for regular developmental conversations, supplementing scarce manager time with lateral support.
Empowered working groups: For projects or initiatives, form small cross-functional pods with delegated decision rights, reducing the need for constant managerial approval.

Technology firm Salesforce implemented a "pod" structure within its engineering organization as manager spans grew during rapid scaling. Each manager oversaw 18–20 engineers, but engineers were organized into self-managing pods of 4–6, each with a rotating technical lead responsible for sprint planning, code reviews, and obstacle escalation. Managers focused on career development, performance calibration across pods, and strategic alignment. The structure preserved the small-team coordination benefits Hackman emphasized while enabling larger nominal spans. Salesforce reported that cycle time and code quality remained stable despite span increases, and engineer satisfaction with team collaboration actually improved (Salesforce Engineering Blog, 2020).

Managerial Capability Building and Role Redesign

Expanding span without enhancing managerial skills and redesigning roles is a formula for failure. Effective managers of larger teams require different competencies—delegation, priority-setting, system-thinking—and must be freed from individual-contributor burdens.

Research on managerial effectiveness in scaled organizations underscores the importance of delegation skills, boundary management (saying no to protect team focus), and systems thinking (designing workflows and norms rather than managing every interaction) (Hill, 2007). A study by the Corporate Executive Board found that top-performing managers of large teams spent 40% less time on direct problem-solving and 60% more time on coaching and system design than average managers (CEB, 2014).

Effective capability-building approaches include:

Transition training: Before span increases, provide targeted training on delegation, prioritization frameworks (e.g., Eisenhower matrix), and running effective one-on-ones at lower frequency.
Manager peer networks: Establish cohorts of managers navigating similar challenges to share practices, troubleshoot, and provide mutual support.
Role clarity enforcement: Insist that managers focus primarily on people leadership, not individual contributions. If a manager must retain IC work, formally reduce their span.
Decision-rights frameworks: Implement tools like RACI matrices or decision logs to clarify what managers must approve versus what teams can decide autonomously.
AI and digital enablers—used judiciously: Equip managers with tools that genuinely reduce administrative burden (e.g., scheduling assistants, automated performance data dashboards) rather than expecting AI to replace human judgment and empathy.

Multinational consumer goods company Unilever faced span expansion in its marketing organization as part of an agile transformation. Recognizing the risk, Unilever co-designed a "Marketing Leadership Lab" with INSEAD faculty, training managers on scaling practices: running short, high-impact one-on-ones; using decision frameworks to empower teams; and building psychological safety at scale through rituals like weekly "team temperature checks." Crucially, Unilever removed individual brand-management responsibilities from people-managers, creating dual tracks where high-potential marketers could advance as either senior individual contributors or people leaders. Within a year, manager burnout metrics stabilized, and team engagement in units with newly trained managers exceeded pre-change baselines (Unilever case study, INSEAD, 2021).

Structural and Financial Supports for Managerial Wellbeing

Addressing the human toll on managers themselves—burnout, inadequacy, overload—requires structural interventions, not just exhortations to "do more with less."

Research on occupational health identifies job resources (autonomy, support, time) as critical buffers against burnout (Demerouti et al., 2001). Organizations that treat managerial capacity as finite and protect it systematically report lower turnover and higher sustained performance.

Effective structural supports include:

Span caps and enforcement: Establish and enforce maximum spans tailored to role complexity. Meta's move to cap spans at 20 is precisely this intervention.
Administrative relief: Assign HR business partners or project coordinators to handle scheduling, compliance paperwork, and routine communications, freeing managers for relational work.
Manager-only development time: Reserve recurring calendar blocks (e.g., Friday afternoons) where managers are expected to engage in professional development, strategic thinking, or self-care—not meetings.
Skip-level meetings: Institute regular skip-levels where senior leaders meet with employees two levels down, providing developmental support and reducing pressure on middle managers to be the sole source of feedback and sponsorship.
Mental health and coaching resources: Provide managers access to executive coaching, peer support groups, and mental health services to process the emotional labor of leadership.

Financial services firm Capital One recognized rising manager burnout during post-merger integration that had expanded average spans. The company implemented several structural protections: capped spans at 15 for people-managers in product and technology roles; hired project managers to absorb administrative coordination; introduced quarterly "manager recharge days" (no meetings, focus on planning and development); and provided all managers access to 1:1 coaching. Exit interview data showed a 25% reduction in manager departures citing overload, and engagement surveys revealed improved manager wellbeing and team satisfaction (Capital One HR metrics, 2020).

Building Long-Term Organizational Capability for Adaptive Span Management

Beyond immediate fixes, organizations benefit from developing enduring capabilities to match span dynamically to evolving conditions.

Psychological Contract Recalibration

The psychological contract—employees' implicit beliefs about mutual obligations—shapes how span changes are experienced. When employees believe the organization owes them developmental investment and care, span expansion feels like breach. Recalibrating involves making expectations explicit and demonstrating that alternative supports exist.

Research by Rousseau (1995) emphasizes that psychological contracts are negotiated, not imposed. Organizations that proactively renegotiate—acknowledging changed realities, offering new resources (peer mentoring, learning platforms), and reaffirming core commitments (fairness, growth opportunity)—can expand spans without destroying trust.

Practically, this means:

Transparent contracting: During onboarding and restructuring, explicitly discuss what employees can expect from managers, peers, and the organization, and what they must self-initiate.
Alternative growth pathways: If manager coaching is scarcer, invest heavily in rotational programs, online learning, and sponsorship from senior leaders outside the direct reporting line.
Recognition of self-directed development: Reward employees who take initiative in skill-building and peer support, signaling that these behaviors are valued organizational contributions.

Adaptive Organizational Design and Continuous Monitoring

Span of control should not be static. As work evolves—new technologies, shifting strategies, changing workforce demographics—optimal spans shift. Building a monitoring and adjustment cadence prevents drift into dysfunction.

Effective practices include:

Annual span audits: Review spans across the organization, flagging outliers and assessing alignment with task characteristics.
Real-time feedback loops: Pulse surveys asking employees and managers about workload, accessibility, and support quality, with thresholds triggering review.
Scenario planning: When planning restructurings or scaling, model span implications explicitly and run pilots before full rollout.
Post-change retrospectives: After span adjustments, conduct formal retrospectives to capture lessons and refine implementation playbooks.

Thoughtful Technology Integration

AI and digital tools hold promise for supporting larger spans, but only when thoughtfully integrated as complements to—not replacements for—human judgment and care.

The misstep many organizations make, as Meta appears to have done, is assuming AI can substitute for managerial presence. The evidence suggests a more nuanced role: AI can handle informational and transactional aspects (scheduling, status updates, knowledge retrieval, basic troubleshooting), freeing managers to focus on relational and developmental work that demands empathy, context, and trust (Davenport & Kirby, 2016).

Effective AI integration strategies include:

Augmentation, not replacement: Frame AI tools as "second brains" that surface insights, draft communications, or suggest learning resources, with managers retaining interpretive and relational authority.
Human-in-the-loop design: Ensure AI recommendations are transparent and managers can override them; avoid black-box systems that erode trust.
Selective deployment: Use AI intensively for routine coordination (e.g., finding meeting times, tracking deliverables) and sparingly for developmental conversations, where algorithmic guidance risks feeling impersonal.
User training and choice: Train both managers and employees on AI tools, and make usage voluntary rather than mandated, respecting variation in comfort and need.

Purpose, Belonging, and Team Identity at Scale

Finally, sustaining engagement in larger teams requires deliberate work to foster shared purpose and belonging. When teams exceed 10–15 members, individuals risk feeling anonymous. Rituals, narratives, and inclusive practices counteract fragmentation.

Research on team identity shows that even large teams can maintain cohesion when leaders invest in identity-building: articulating a compelling team mission, celebrating collective wins, and ensuring every member feels seen (Haslam et al., 2009).

Practical mechanisms include:

Shared mission and storytelling: Regularly communicate how the team's work advances organizational purpose, and highlight individual contributions in team forums.
Inclusive rituals: Weekly stand-ups, monthly retrospectives, or quarterly offsites where everyone has voice and visibility.
Rotating spotlight: Dedicate time in team meetings for individual members to share a project, a learning, or a personal story, building interpersonal connection.
Celebrating milestones: Recognize not just deliverables but team development milestones (e.g., "We hit 100% on-time delivery for three months," "Jamie mentored three peers this quarter").

Conclusion

Meta's retreat from AI-managed mega-teams is a cautionary tale with broad resonance. The allure of wide spans—cost savings, flatter structures, empowerment—is real, but so are the perils: coordination collapse, innovation suppression, managerial burnout, and employee neglect. The evidence is clear: optimal span is not a fixed number but a function of task characteristics, workforce capability, and organizational support systems. For tightly interdependent, creative work—the kind that drives competitive advantage in knowledge economies—smaller teams and closer managerial attention remain indispensable.

Organizations navigating span pressures should resist the temptation of one-size-fits-all edicts. Instead, diagnose task demands, communicate transparently, distribute leadership, build managerial capability, and integrate technology thoughtfully. Protect both managers and employees from overload through structural supports, not just rhetoric. And cultivate the long-term capabilities—adaptive design, psychological contracting, purpose-building—that allow span to flex responsibly as conditions change.

Meta's course correction reminds us that some aspects of effective organizing are stubbornly human. Technology can extend our reach, but it cannot replace the judgment, empathy, and relational work that make teams thrive. As organizations continue to experiment with structure, the test is whether we learn from these experiments—embracing what works, abandoning what doesn't, and keeping the human experience at the center of organizational design.

Research Infographic

The Physics of Scale Slide Deck

References

Allen, D. G., Bryant, P. C., & Vardaman, J. M. (2010). Retaining talent: Replacing misconceptions with evidence-based strategies. Academy of Management Perspectives, 24(2), 48–64.
Carson, J. B., Tesluk, P. E., & Marrone, J. A. (2007). Shared leadership in teams: An investigation of antecedent conditions and performance. Academy of Management Journal, 50(5), 1217–1234.
Colquitt, J. A., Conlon, D. E., Wesson, M. J., Porter, C. O. L. H., & Ng, K. Y. (2001). Justice at the millennium: A meta-analytic review of 25 years of organizational justice research. Journal of Applied Psychology, 86(3), 425–445.
Corporate Executive Board. (2014). The high-impact manager: How top performers lead large teams. CEB.
Davenport, T. H., & Kirby, J. (2016). Only humans need apply: Winners and losers in the age of smart machines. Harper Business.
Demerouti, E., Bakker, A. B., Nachreiner, F., & Schaufeli, W. B. (2001). The job demands-resources model of burnout. Journal of Applied Psychology, 86(3), 499–512.
Eisenberger, R., Stinglhamber, F., Vandenberghe, C., Sucharski, I. L., & Rhoades, L. (2002). Perceived supervisor support: Contributions to perceived organizational support and employee retention. Journal of Applied Psychology, 87(3), 565–573.
Garvin, D. A., Wagonfeld, A. B., & Kind, L. (2008). Google's Project Oxygen: Do managers matter? Harvard Business School Case, 313(110).
Hackman, J. R., & Oldham, G. R. (1976). Motivation through the design of work: Test of a theory. Organizational Behavior and Human Performance, 16(2), 250–279.
Hackman, J. R., & Vidmar, N. (1970). Effects of size and task type on group performance and member reactions. Sociometry, 33(1), 37–54.
Haslam, S. A., Reicher, S. D., & Platow, M. J. (2009). The new psychology of leadership: Identity, influence and power. Psychology Press.
Hill, L. A. (2007). Becoming the boss. Harvard Business Review, 85(1), 48–56.
Hoegl, M., & Gemuenden, H. G. (2001). Teamwork quality and the success of innovative projects: A theoretical concept and empirical evidence. Organization Science, 12(4), 435–449.
Huckman, R. S., & Staats, B. R. (2011). Fluid tasks and fluid teams: The impact of diversity in experience and team familiarity on team performance. Manufacturing & Service Operations Management, 13(3), 310–328.
Ibarra, H., Carter, N. M., & Silva, C. (2010). Why men still get more promotions than women. Harvard Business Review, 88(9), 80–85.
Lau, D. C., & Murnighan, J. K. (1998). Demographic diversity and faultlines: The compositional dynamics of organizational groups. Academy of Management Review, 23(2), 325–340.
Maslach, C., Schaufeli, W. B., & Leiter, M. P. (2001). Job burnout. Annual Review of Psychology, 52(1), 397–422.
Pearce, C. L., & Conger, J. A. (2003). Shared leadership: Reframing the hows and whys of leadership. Sage.
Rousseau, D. M. (1995). Psychological contracts in organizations: Understanding written and unwritten agreements. Sage.
Urwick, L. (1956). The manager's span of control. Harvard Business Review, 34(3), 39–47.
Van De Voorde, K., & Beijer, S. (2015). The role of employee HR attributions in the relationship between high-performance work systems and employee outcomes. Human Resource Management Journal, 25(1), 62–78.
Wu, L., Wang, D., & Evans, J. A. (2019). Large teams develop and small teams disrupt science and technology. Nature, 566(7744), 378–382.

Jonathan H. Westover, PhD, Chief Research Officer (Nexus Institute for Work and AI); Professor, Chief Workforce and Learning Officer (Future State University); Founder & CEO (Human Capital Innovations); Organizational Leadership (UVU). Read Jonathan Westover's executive profile here.

Suggested Citation: Westover, J. H. (2026). When Span of Control Becomes Span of Chaos: Meta's Retreat from AI-Managed Mega-Teams and What It Means for Organizational Design. Human Capital Leadership Review, 36(3). doi.org/10.70175/hclreview.2020.36.3.6