Every flat organization has a hierarchy.
Some put it on the org chart. Others let the new hire discover it during a disagreement.
The discovery is usually quick. One person's opinion ends the debate. Another person can commit resources without asking. A third has no formal title but can make a proposal disappear through silence. The company may have removed management layers. It has not removed differences in authority, expertise, access, status, trust, tenure, or influence.
This is not proof that flat organizations are dishonest or doomed. Fewer layers can shorten communication paths, give people more autonomy, and keep decisions close to the work. Self-managing organizations can distribute authority in thoughtful, formal ways that differ sharply from a conventional managerial hierarchy.
The comforting myth is that removing managers removes hierarchy itself.
Management work still exists. Priorities must be set. Conflicts must be resolved. Performance and growth need attention. Resources must be allocated. Someone has to decide when reasonable people disagree. If the formal structure does not say how these things happen, informal structure will.
Flat can be a useful design choice. "Structureless" is usually an inaccurate description of where power went.
What the evidence says
- Jo Freeman's influential 1970s essay argued that groups do not escape structure by refusing to formalize it. Informal elites can become harder to challenge because their power is not acknowledged.
- Freeman wrote about women's liberation groups, not modern companies. Her argument is a primary historical analysis, not a universal empirical law.
- A study of 6,234 video game startups found a trade-off: additional hierarchy was associated with stronger commercial success and a potentially small cost to creative novelty in that setting.
- A scholarly review of self-managing organizations found that radical decentralization does not simply mean an absence of structure. It involves formal, systematic redistribution of authority.
- The choice is not hierarchy or freedom. It is what kind of hierarchy, where authority sits, and whether people can see and challenge it.
Flat describes layers, not power
An organizational hierarchy can mean several related things.
There is hierarchy of formal authority: who can direct work, allocate resources, evaluate performance, and make final decisions. There is hierarchy of expertise: whose judgment carries weight in a particular domain. There is hierarchy of status and influence: who is listened to, included, trusted, or feared. There is hierarchy of access: who can reach the people who control consequential choices.
Removing managerial layers changes the first form. It may affect the others. It does not erase them.
A founder can say everyone is equal while retaining ownership, compensation authority, and final say. A self-managing team can distribute decisions across roles while still deferring to expertise. A formally hierarchical company can create wide autonomy within clear boundaries.
These arrangements are different, but none is hierarchy-free.
The distinction matters because "flat" often carries moral meaning. Flat sounds modern, human, and creative. Hierarchy sounds slow, controlling, and beige. Once structure becomes an identity, people may defend the label instead of examining the work.
The better questions are practical:
- Which decisions should sit close to the work?
- Which decisions need integration across the organization?
- What authority does each role hold?
- How can people challenge a decision or remove a bad leader?
- Who performs the management work?
- Which informal power has become consequential enough to make visible?
A layer count cannot answer any of these by itself.
Jo Freeman's warning was about invisible power
In 1970, feminist activist and political scientist Jo Freeman gave a speech based on her experience in women's liberation groups. Versions of the argument later circulated under the title The Tyranny of Structurelessness.
Freeman's central claim was not that every group should imitate a bureaucracy. It was that informal structure forms whether a group acknowledges it or not. Friendship networks, access to information, time, skills, and social approval can create elites. When those relationships are unwritten and denied, the people exercising power may be difficult to identify, hold accountable, or replace.
She proposed principles for more democratic structuring: delegate specific authority to specific people for specific tasks, require accountability to those who selected them, distribute authority widely, rotate tasks, allocate work through rational criteria, share information, and provide equal access to needed resources and skills.
The context matters. Freeman was writing about a social movement in a particular historical period, not reporting a controlled study of technology startups. Her essay should not be used as proof that every informal team becomes tyrannical.
Its durable insight is narrower and stronger. Refusing to discuss structure does not stop relationships from shaping power. It stops the group from examining them openly.
Modern organizations repeat the pattern when they announce "no bosses" but leave employees to infer whose approval is safe to skip.
Self-management is more structure, carefully placed
Serious self-management does not ask everyone to decide everything.
It assigns authority through roles, teams, processes, advice requirements, elected positions, peer agreements, or other explicit mechanisms. The authority may be distributed and changeable rather than concentrated in permanent managers. It is still authority.
Michael Lee and Amy Edmondson's review of self-managing organizations defines them as efforts to decentralize authority radically and systematically across the organization. Their review draws together research on post-bureaucratic organizations, humanistic management, and organizational democracy while emphasizing that scholarly understanding of radical decentralization remains limited.
The words "systematically" and "limited" are both important.
Self-management is not achieved by deleting titles and telling people to act like owners. It needs a system that specifies decision rights, coordination, conflict resolution, performance, compensation, staffing, strategy, and change. Otherwise the familiar management work remains, performed by whoever has the confidence, relationships, or spare time to claim it.
Nor is there enough evidence to promise that one model works across every organization and task. Case studies of distinctive companies are useful for generating ideas. They are not universal operating instructions.
The less a company relies on the ordinary manager role, the clearer its alternative mechanisms need to be.
Hierarchy carries a trade-off, not a verdict
Layers can slow information, distance leaders from work, and concentrate decisions too high. They can also divide complex work, support coordination, develop people, and give decisions a clear home.
Saerom Lee examined this trade-off in a study of 6,234 video game startups founded between 1971 and 2015. The analysis used credits and job titles to estimate hierarchical levels and examined the first game developed by each company. It distinguished creative success from commercial success.
The findings challenged the idea that flatter was uniformly better. Additional hierarchical levels were associated with stronger commercial success and a potentially marginal reduction in creative novelty. The proposed explanation involved division of labor, coordination, and the increasing complexity of turning creative work into a commercial product.
The limits are substantial. These were video game startups, hierarchy was inferred from job information, and observational data cannot settle how every company should organize. Creative ratings and commercial outcomes also capture only part of organizational health.
The value of the study is not a prescription to add managers. It is evidence that structure can help one outcome while modestly hurting another. Organizational design has trade-offs because organizations pursue several things at once.
A creative studio, emergency service, global manufacturer, and 6-person research team do not need the same coordination system. "Flat" cannot answer what kind of work they do or which failure matters most.
Informal hierarchy forms around useful things
Not all informal power is sinister.
People defer to someone because they have sound judgment. They ask a long-serving colleague because that person remembers the system. A trusted connector helps 2 teams cooperate. Expertise and trust should influence work.
The problem appears when influence becomes consequential, unacknowledged, and unaccountable.
If 1 person can veto proposals without formal authority, the group needs to understand why. If a founder's "suggestion" functions as an instruction, calling it advice does not create autonomy. If access to an inner circle determines whose ideas receive attention, the company has a hierarchy of access whether or not the org chart admits it.
This connects to the org chart nobody drew. Informal advice and trust networks help organizations function. They can also exclude newcomers, overload central people, and route around formal accountability.
Healthy flat organizations do not attempt to suppress every informal relationship. They observe where real influence sits and decide which parts should remain relational, which need broader access, and which have become formal enough to name.
Visibility is not bureaucracy. It is the precondition for a useful argument about power.
Management work does not disappear with managers
Remove a management layer and several things can happen.
The team may gain autonomy and handle the work collectively. A lead role may emerge. The founder may absorb the decisions. Senior specialists may perform coaching and coordination without recognition or time. Difficult people issues may wait because nobody wants to claim authority. Performance feedback may become everybody's responsibility and therefore nobody's regular job.
Each outcome is possible because the role and the work are not the same.
Management work includes:
- setting and reconciling priorities;
- allocating resources;
- building capability;
- giving feedback and addressing performance;
- resolving conflict;
- hiring and staffing;
- connecting work across boundaries;
- carrying information and decisions;
- protecting people from impossible demand;
- deciding when consensus ends.
A self-managing system can distribute these responsibilities across people and roles. It must still allocate capacity and accountability for them.
Watch for unpaid management. If the most dependable employee becomes the default coordinator, coach, mediator, and quality check while retaining a full individual workload, the organization has created a manager-shaped burden without the title, authority, or compensation.
Flatness should not depend on invisible labor.
Consensus can hide a veto
Flat teams often prefer consensus because it appears to distribute power. Consensus can improve decisions and commitment when people have relevant information, enough time, and a real ability to disagree.
It can also conceal authority.
A decision waits until the highest-status person is comfortable. A persistent participant outlasts everyone else. People who dislike conflict stop objecting. A person with urgent customer information cannot move because the whole group has not met. The company calls this collaboration.
Not every decision deserves the same method. Define categories:
- Individual decisions for work within a clear role boundary.
- Advice decisions where an owner must consult affected or expert people, then decide.
- Consent decisions where the group proceeds unless someone identifies a defined material objection.
- Consensus decisions for the smaller set where durable collective agreement is essential.
- Escalated decisions when local choices create organization-wide consequences.
Then state who chooses the method and how long it has.
The point is not to import a fashionable governance vocabulary. It is to stop "we decide together" from meaning "nobody knows when the decision is finished."
A useful flat organization can answer hard questions
The health of a flat structure is easier to judge through its answers than its layer count.
Ask:
- Who can decide this, and how would a new employee know?
- What happens when 2 roles make incompatible decisions?
- Who gives feedback, handles poor performance, and supports growth?
- How are unpopular or low-status voices protected?
- How can informal leaders be challenged or replaced?
- Who owns cross-team coordination?
- Where can someone appeal a decision?
- Which management work is being performed without time or recognition?
Clear answers do not make the organization bureaucratic. They make autonomy usable.
An org chart may still help. It can show teams, roles, formal accountabilities, and the places where authority is deliberately distributed. The chart does not need to resemble a pyramid. It does need to tell the truth.
Be especially careful with titles. Inflated or improvised titles can create an informal hierarchy while the company continues to describe itself as flat. If the first 10 employees are all "Head of" something, the hierarchy has not vanished. It is waiting for the next 10 people to ask what the titles mean.
Make power visible enough to improve
Hierarchy is not automatically good. Flatness is not automatically humane.
Good organizational design puts authority where the work needs it, limits that authority, makes it legible, and creates ways to challenge it. Some companies will do that with managers and relatively few layers. Others will use distributed roles and self-managing teams. Many will combine the 2.
The label matters less than whether people can see how decisions happen, understand what they own, access the information they need, and hold consequential power accountable.
Remove layers when they add delay without judgment or support. Distribute authority when people close to the work can use it well. Keep enough structure to coordinate complexity and protect people from hidden rule.
A flat organization still has a hierarchy. The honest ones tell you where it is.