The Scorecard of a 30-Year 'Bossless Company' Experiment: Eliminate Visible Hierarchy, and Invisible Hierarchy Grows

The argument for eliminating organizational charts is no longer a radical experiment limited to a few Silicon Valley companies. Around the mid-2020s, global corporations rapidly cut middle management layers amid a trend known as the 'Great Flattening.' According to an analysis of employment records by data analytics firm Live Data Technologies, the number of management-level positions at U.S. publicly traded companies decreased by approximately 6.1% between May 2022 and May 2025. Furthermore, a proprietary report released in June 2025 by Gusto, a small- and medium-sized business payroll platform, based on a sample of 8,500 payroll records, showed that the number of direct reports per manager nearly doubled from about 3 in 2019 to about 6 as of the third quarter of 2024. Gartner also projected that by 2026, one in five organizations would use AI to flatten their organizational structures and cut more than half of their middle management positions.

Moves by individual companies have been even more specific. In an official corporate memo in September 2024, Amazon CEO Andy Jassy announced that the ratio of individual contributors to managers would be increased by at least 15% by the first quarter of 2025. According to an all-hands meeting transcript obtained by Business Insider, Jassy stated directly that Amazon achieved this goal by the end of March through team consolidation and the transition of managers into individual contributor roles rather than through massive layoffs. Similarly, CNBC reported statements made directly by Brian Welle, VP of People Analytics at Google, during an all-hands meeting: Google has cut the number of managers leading small teams with fewer than three direct reports by 35% over the past year (many of whom transitioned to individual contributors rather than being laid off). Meanwhile, Citigroup, spearheaded by CEO Jane Fraser, undertook an organizational restructuring that reduced its management layers from 13 down to 8. Meta CEO Mark Zuckerberg's remark, "We don't want a management structure where managers manage managers," encapsulates the corporate sentiment of this period.

Entering 2026, this trend accelerated further. In May, Meta cut approximately 8,000 employees—about 10% of its total workforce—citing the need to secure funding for AI infrastructure investments. These cuts were concentrated among middle managers and software engineers, and the remaining organization was reorganized into small team units called 'pods.' That same month, Cisco also announced layoffs affecting approximately 4,000 employees. Payment company Block cut about 4,000 employees, or roughly 40% of its total workforce, in February. At the end of March, founder Jack Dorsey, alongside Roelof Botha of Sequoia Capital, published an essay titled 'From Hierarchy to Intelligence,' arguing that middle management layers may not be permanently necessary. It is telling that this essay directly referenced past experiments such as Spotify's squad model, Zappos' holacracy, and Valve's flat structure. Even the most radical organizational redesign attempts in the AI era are ultimately returning to these past experiments to re-examine their successes and failures.

This leads to a question positioned at the logical conclusion of such trends: Beyond reducing managers, can a company that completely removes organizational charts and reporting lines actually function properly? Nearly 30 years of corporate case studies have accumulated around this question. To cut straight to the conclusion, these cases suggest that while a 'company without an org chart' can exist, a 'company without a structure' is difficult to operate over the long term.

Companies That Erased Their Org Charts: What Did They Attempt?

The oldest case is Valve, the company that operates the gaming platform Steam. A new hire handbook released in 2012 stated, "We've had no bosses since 1996," explaining a structure where there are no official managers and no one reports to anyone else. Employees reportedly pull wheeled desks around to choose projects they want to participate in, and project proposers recruit colleagues to form teams. Performance evaluations and compensation are also known to be centered largely around peer reviews. Although recent reports, such as a work experience review published in April 2026, still describe Valve as officially maintaining a manager-free structure, this relies on descriptions by observers and current/former members rather than official corporate announcements.

A more dramatic experiment took place at the online shoe retailer Zappos. Founder Tony Hsieh began a pilot introduction of 'Holacracy' for the HR team in early 2013, and at the company-wide all-hands meeting in the fourth quarter of that November, he announced its expansion to the entire workforce of about 1,500 people. The initial goal was to complete the transition within 2014. Holacracy is a management system that eliminates managers, job titles, and traditional organizational charts, distributing authority to self-governing units called 'circles.' It was a structure with no department heads, no vice presidents, and no official bosses. Zappos became the largest company in the world to adopt this model and served for a while as a symbol of the 'bossless company.'

Buffer, a social media management tool company, also experimented with a self-management system that eliminated managers and job titles, inspired by Frédéric Laloux's book *Reinventing Organizations*. Employees participated in determining their own pay and worked through collaboration and consensus rather than directives from a boss.

The Outcome of the Experiments: Structure Returned

The subsequent trajectory of these experiments is revealing. However, the 2014 target was not met, and as the parallel operation dragged on, Hsieh sent an ultimatum memo in March 2015 to all employees instructing them to either commit to the new system or take a severance package and leave. Media reports indicated that about 14% to 18% of the workforce left the company. Afterwards, Zappos quietly changed course. According to reports by the U.S. business media outlet Quartz, Zappos scrapped the increasingly formalistic Holacracy meeting structure and reintroduced management roles. This is supported by circumstantial evidence: recent job postings by Zappos on the career platform Built In actually feature managerial titles like 'Senior Manager' and 'Marketing Manager,' and traditional executive titles such as current vice presidents are confirmed in data from corporate intelligence provider CB Insights. Jon Bunch, the executive who led the introduction of Holacracy, later described the company's structure not as a Holacracy, but as a 'market-based ecosystem' where each team is accountable for its own profit and loss. A retrospective analysis published in 2026 diagnosed that while Holacracy-style terminology like circles and 'team leads' remain at Zappos, decision-making has reverted to a familiar model combining structure, accountability, and leadership.

Buffer's reflection is even more straightforward. Reflecting on roughly a year of self-management experimentation, co-founder Leo Widrich admitted that employees felt overwhelmed by the given freedom in the absence of higher-level visions and goal-setting. He left the paradoxical lesson that "the stronger a self-managing company is, the more structure it needs," and co-founder Joel Gascoigne summarized that they learned structure and hierarchy are not the same thing, and that employees actually desired structure.

Even Valve, Which Officially Proposes a Manager-Free System, Faces Mixed Internal Reviews

Internal evaluations are mixed even at Valve, which officially champions a manager-free system. A 2023 investigative report by the investigative journalism channel *People Make Games* and numerous Glassdoor reviews left by former employees have pointed out that behind Valve's flat structure lies a 'hidden hierarchy' where influential groups hold de facto decision-making power. One former employee described the company as a 'pseudo-flat structure.' However, this criticism is based on the testimonies of current and former members and has not been established as fact through organizational data. The possibility that informal power structures can take root where official org charts disappear was a problem already raised in 1972 by sociologist Jo Freeman in her essay 'The Tyranny of Structurelessness.' The core of this criticism is that eliminating explicit rules and hierarchies does not make power itself disappear; rather, it risks shifting power to informal networks where accountability is difficult to assign.

Haier, Which Looks Like an Exception, Is Actually the Most Sophisticated Structure

On the flip side is Chinese home appliance giant Haier, frequently cited as a success story. Haier introduced the 'Rendanheyi' model in 2005, effectively dismantling its middle management layers. Former Chairman Zhang Ruimin stated in an interview with McKinsey that over 12,000 middle management layers were eliminated, and an organization of 80,000 people was reorganized into about 4,000 micro-enterprises. Each micro-enterprise possesses an independent profit-and-loss statement, contracts directly with the market or other micro-enterprises internally, and exercises its own decision-making authority, hiring power, and profit distribution rights.

What is notable is that Haier is not a 'company without a structure.' Haier replaced the structure of bosses and reporting lines with another structure: contracts and market mechanisms. The accountability of each unit became even clearer through the numbers of profit and loss, and micro-enterprises that failed to deliver performance were not allocated resources. Zhang Ruimin himself has stated that most companies find this model difficult to copy because introducing it requires transferring decision-making, personnel, and compensation powers entirely to the front lines. Rather than eliminating hierarchy, it is closer to redesigning the system so that market mechanisms perform the work that hierarchy used to do.

What the Data Tells Us: The Problem Is Not the Existence of Managers, But Their Quality

There is also data that arguments for the obsolescence of reporting structures tend to overlook. Gallup has repeatedly confirmed through various surveys since the 'State of the American Manager' report (2015) that at least 70% of the variance in team engagement is explained by managerial factors. According to the 'State of the Global Workplace: 2026' report released by Gallup, global employee engagement dropped to 20% in 2025, reaching its lowest level since 2020. This follows consecutive annual declines from 23% in 2022. The core background for this drop identified by Gallup is the sharp plunge in manager engagement. Manager engagement fell from 31% in 2022 to 27% in 2024, and further down to 22% in 2025, with the steepest drop (5 percentage points) occurring between 2024 and 2025. Global productivity losses resulting from low engagement were estimated at about $10 trillion, or roughly 9% of global GDP.

Notably, Gallup directly mentioned the side effects of organizational flattening in the report. Analyzing the case of the South Asia region, where manager engagement plummeted by 8 percentage points, Gallup pointed out that reduced manager proportions and expanded spans of control could be the underlying cause. Gallup advised that organizational flattening must not become a simple labor cost transaction, and that the engagement of managers and teams, as well as the scope of managers' workloads, must be considered together. An interesting connection is with AI. In Gallup's Q1 2026 U.S. survey, the strongest predictor of employee AI usage, excluding technical infrastructure, was whether their direct manager actively supported AI adoption. The survey finding that the AI transformation—which served as the justification for stripping away management layers—actually spreads through managers is paradoxical.

Three Things HR Must Check: Span of Control, Redefining Roles, and Information Infrastructure

This question is not foreign to Korean companies either. Across domestic enterprises and startups alike, organizational restructuring to reduce job tiers and standardize titles has continued, and the number of people and scope of work managed by a single team leader are steadily increasing. Fatigue among employees regarding 'pseudo-horizontal' organizations—where a horizontal structure is outwardly championed while actual decision-making remains concentrated among a few executives—is also a familiar experience. The questions raised for domestic HR by overseas experimental lessons ultimately boil down to the same thing: Have authorities been pushed down as much as layers were reduced, and has a design been made for who will take over the work previously done by the vanished layers?

Translating this discussion into the language of practical operational checks yields a list of three items. First is the limit of the span of control. In situations where a single manager's direct reports double in a short period, as seen in Gusto's data, companies must first define the team size and workload that the remaining managers can realistically handle before cutting layers. A column published late last month in the Forbes Human Resources Council pointed out that behind the flattening advocacy—which argues that as AI absorbs coordination tasks, managers can coach larger teams—almost no calculation has been made as to where the time and competencies required for that coaching actually come from. Based on research findings that wider spans of control can diminish managers' own engagement, Gallup similarly advises designing managers' other responsibilities concurrently when deciding on flattening. If labor cost savings achieved through layer reduction are offset by the burnout and team defection of remaining managers, it is not a financially winning trade.

Second is the redefinition of the manager's role. Employee dissatisfaction evident in the Buffer and Zappos cases was not directed at guidance- and development-oriented management, but rather at control- and approval-centric management. Stripping away coordination tasks that AI can absorb, such as aggregating reports and coordinating schedules, leaves core design questions: what work remains for managers, and how those tasks will be reflected in evaluation and compensation systems. Third is information infrastructure. Both Haier's micro-enterprise model and Buffer's salary transparency presupposed a radical level of information transparency. If an organization eliminates reporting structures without a system where everyone can see who decided what and why, information flows through informal networks, and that network itself becomes a new form of power.

We Must Change the Question: From 'Do We Need an Org Chart?' to 'Who Pays the Coordination Costs?'

So, can a company without an org chart and reporting structure function properly? The answers accumulated from nearly 30 years of corporate cases lean toward conditional. Valve's long-term survival shows that for small-scale companies with high employee self-direction and products holding an overwhelming market position, such structures can be sustained for a considerable period. However, this is closer to an observed case than a verified success formula. A causal relationship proving that the manager-free system itself is the cause of performance has not been established, and as examined earlier, testimonies that informal power holds de facto decision-making power within Valve have repeatedly surfaced. Even in such cases, management functions themselves—coordination, evaluation, and direction-setting—do not disappear. The criticism common to evaluations of the Valve and Zappos cases is that without an official structure, the risk increases that these functions will shift to informal power, repetitive meetings, or a few tacit power brokers.

Conclusion

Conversely, the trajectories of Zappos and Buffer demonstrate the price organizations pay when eliminating hierarchy becomes an end in itself. What both companies learned through experimentation was not "how to get rid of managers," but "how to distinguish between structure and hierarchy." What employees in both companies did not want was structure itself, but rather bad management—namely, approval bottlenecks and control-centric reporting. Haier's case goes a step further, showing that eliminating traditional reporting structures requires the prior establishment of a more sophisticated alternative system: clear profit-and-loss accountability, contractual relationships, and transparent information sharing.

In the 2026 business environment, this question is no longer theoretical. As seen in the cases of Meta and Block, as AI absorbs coordination tasks like compiling reports and scheduling, many companies are moving toward reducing management layers in actual organizational restructurings. However, Gallup's data and past cases consistently point in one direction: while you can eliminate an org chart, you cannot eliminate the work that the org chart did. Companies that strip away reporting structures without designing who will take over the functions of direction-setting, stakeholder coordination, and evaluation and development—and through what systems—are highly likely to end up with invisible hierarchies and higher coordination costs instead of visible ones. The real answer to whether a company without an org chart can function ultimately depends not on whether it erased hierarchy, but on where it re-inscribed the functions of hierarchy.