Liberated Company: What Works and What Fails

A brass foundry in northern France scrapped its HR department, its time clocks and its shop-floor supervisors in the early 1980s, and it thrived for twenty-five years. The FAVI story launched a wave of management transformations that is still running. Some of them held. Many stopped after two or three years, usually without an announcement.

The liberated company is one of the most discussed management ideas in Europe and one of the vaguest. This article separates what researchers actually observed from what the promise implies, then describes what organisations do when autonomy survives past year three.

Where the liberated company comes from

Isaac Getz, a professor at ESCP Business School, popularised the term in a 2009 California Management Review article. That same year he published Freedom, Inc. with journalist Brian M. Carney. In French the concept travels as entreprise libérée, and Getz also calls it the F-form, short for freedom-form organisation, as opposed to command-and-control.

The idea has older roots. It extends Douglas McGregor Theory Y (1960), which holds that people want to work and take responsibility once you stop preventing them. Tom Peters had already published Liberation Management in 1992. Researchers Patrick Gilbert, Ann-Charlotte Teglborg and Nathalie Raulet-Croset, writing in Gérer & Comprendre (issue 127, March 2017), place the model in the lineage of the human relations school of the 1930s and the participative management movement of the 1980s, a wave that peaked and then faded.

The founding case is FAVI, a brass foundry in Hallencourt run by Jean-François Zobrist from 1983 to 2009. Zobrist removed time tracking, bonuses and middle management, and handed the plant to autonomous mini-factories organised by customer. The same handful of names follows in every article since: Poult, Chronoflex, Michelin, W.L. Gore, Harley-Davidson, HCL Technologies.

The vocabulary overlaps with the teal organisation described by Frédéric Laloux, and it carries on part of what the self-management wave of the 1970s promised.

What the model actually asks for

Getz describes a posture more than a method. Companies that claim the label still share a fairly stable core.

• Remove middle layers and hand the work to small autonomous teams.
• Cut back support functions, reporting and control procedures.
• Drop the symbols of power: reserved parking, closed offices, badges, expense approvals routed up three levels.
• Turn the CEO into a "liberating leader" whose job is to shape the environment rather than issue orders.
• Trust people upfront, before they have had a chance to earn it.

The model is precise about what to remove and much vaguer about what to put in its place. That gap is where lasting transformations separate from the ones that quietly stop.

What research found on the positive side

Gilbert, Teglborg and Raulet-Croset ran an inside study of FAVI, Poult and Chronoflex between 2012 and 2015. Their findings are encouraging on several counts. Hierarchies really do flatten, shop-floor teams really do gain decision power, employees report broad satisfaction, and healthy financial results get shared with them.

The most systematic study remains Au-delà de l'entreprise libérée by Thierry Weil and Anne-Sophie Dubey, published by La Fabrique de l'industrie in 2020. The study covers ten organisations between 50 and 1,300 employees, with around ten interviews each, run across every level of the hierarchy. The report identifies six shared levers: flattening the reporting line, splitting into small units of 5 to 40 people, reducing status symbols, changing managerial posture, creating consultation spaces, and redefining support functions.

The autonomy people end up with almost always covers the "how", meaning the way work gets done. It rarely reaches the "what", the objectives, and even more rarely governance itself. Worker cooperatives are the exception, because shared ownership opens strategic decisions.

Those gains are real, and they explain the enthusiasm. They fix a good share of the problems a heavy hierarchy creates: slow decisions, filtered information, initiative that dies on the way up.

Why so many attempts run out of steam

A model that produces those effects deserves attention. The trouble shows up a few years later, once the launch energy fades and nothing solid was built in the meantime.

Informal power takes the place of formal power. Jo Freeman described this in 1972 in The Tyranny of Structurelessness. When an organisation removes its official structure, an unofficial one takes over immediately. It runs on seniority, friendship and comfort with speaking up. Nobody elected it, nobody can challenge it, and it appears in no document.

The boss comes back through the window. Gilbert and his co-authors report a striking paradox: flattening the hierarchy strengthens the image of the leader. Once the middle layers are gone, every arbitration lands on the one figure still standing. The model becomes dependent on a person. Zobrist retired in 2009 and handed FAVI to Dominique Verlant. A system that runs on the personality of its founder rarely outlives their departure.

Middle managers are left alone. Weil and Dubey describe a "destabilised" population, asked to give up its role without being offered another one, and rarely trained for the next one. They are the group that decides whether the transformation lands, and the group everyone forgets.

Workload rises while resources stay where they were. The La Fabrique de l'industrie study lists the observed effects: responsibility that becomes exhausting to carry over time, rising psychosocial risk, unexpected turnover among people who lose their bearings, and transparency that slides into peer surveillance.

Freedom favours whoever already knew how to use it. Engagement becomes uneven. People comfortable speaking up gain influence, everyone else loses it.

The best documented international case is Zappos. The retailer adopted Holacracy in January 2014, then in March 2015 Tony Hsieh offered a buyout to anyone unwilling to commit. 18% of the workforce took it. The company drifted away from the model over the following years. Medium dropped it in 2016, and GitHub walked back its manager-free structure. Holacracy is not the liberated company, yet both hit the same wall: a rulebook adopted in one block, with no gradual learning, gets rejected like a graft.

The usual examples, and what we actually know

The evidence base is narrow and it repeats from article to article. Gilbert and his co-authors turn that into a substantive critique: a model that spreads by always citing the same five or six companies raises a reproducibility problem. Economist Thomas Coutrot pointed out in 2018 how few rigorous scientific evaluations exist on the subject.

Two cases deserve a correction, because they come up constantly.

Michelin launched a programme in 2012 called MAPP, for autonomous management of performance and progress, rolled out across a large share of its plants. Production teams gained control over their work organisation, their indicators and their continuous improvement, with results measured on several sites. The group still has a reporting line, an executive committee and conventional shareholder governance. Calling it a liberated company in Getz's sense overstates it. Calling it real decision power for production teams is accurate.

W.L. Gore is the closest thing to a long-running case, with its lattice organisation in place since 1958 and no formal job titles. It also runs on a strong ownership model, a slow hiring process and a sponsor system for every new joiner, meaning a considerable amount of structure that rarely makes it into the summaries.

The accurate reading fits in one sentence. These companies gave real decision power to their operating teams, which is already a lot, and none of them removed their governance.

Liberate and structure, rather than liberate and improvise

Removing a structure without offering another one leaves a vacuum, and teams fill it with whatever is at hand: habits, friendships, prestige. The useful question is about the replacement rather than the removal.

Organisations that distribute authority durably write down three things.

Roles come first. Each one carries a purpose, a domain it decides on alone, and accountabilities others can expect. A role is smaller and more mobile than a job, one person holds several, and it changes when the need changes. That is the difference between a role and a job description.

Then the decision rule. Knowing who decides, on what scope, and through which process when a topic exceeds a single role. Consent decision-making answers that without requiring unanimity and without sending everything back to the CEO.

Finally the space where the structure evolves. A regular governance meeting is where a team creates a role, amends another, or processes a tension about how they work. Without that space the org chart freezes and the gap with real work widens month after month.

Sociocracy and Holacracy supply exactly those mechanisms. You can adopt them whole or borrow the pieces that fit.

	Liberated company as it gets told	Explicit governance
What you remove	Hierarchy layers, control, status symbols	Hierarchy layers, control, status symbols
What replaces it	Trust and culture	Written roles, a decision rule, governance meetings
Who decides what	Worked out case by case	Written in the role, readable by anyone
When it gets stuck	Arbitration goes back to the leader	The tension is processed in a governance meeting
Survives the founder leaving	Poorly	Well, the rules outlive the people
What a new joiner sees	A culture to decode	A map of roles to read

The EVEA cooperative is a good illustration. The company grew from 60 people in 2020 to 140 in 2023, spread across three sites, with more than eight competency areas and cross-cutting assignments. Shared governance was already in place. What was missing was the map, since a hand-maintained slide deck no longer kept up. Making the structure readable in real time is what turned autonomy into something usable, including for new joiners and in front of clients. Their full story is in the EVEA case study.

Where to start

Weil and Dubey recommend gradual experimentation over a big-bang switch. The sequence below follows their points of vigilance.

1. Map what already exists before changing anything. Writing down the roles people genuinely hold usually reveals accountabilities nobody owns and topics three people each believe they lead.
2. Name the non-negotiable zones: safety, client commitments, budget, legal obligations. Stating clearly what stays closed is what makes everything you open credible.
3. Pick a pilot team of 5 to 40 people, the range where self-organisation works according to the study, and give it a written decision scope.
4. Set a decision rule and apply it to real topics in the first week.
5. Open a monthly governance meeting so roles evolve instead of ossifying.
6. Train and back the middle managers first, since they are the ones losing the most footing.
7. Measure average decision lead time, tensions processed, turnover and perceived workload. And talk about it publicly only once you have results, to avoid expectations nobody can meet.

The guide to transitioning toward horizontal management walks through that progression, and the article on defining roles gives the writing template.

Frequently asked questions

What is a liberated company?

A liberated company is an organisation that removes its middle management layers and its control systems so teams decide for themselves how they do their work. Isaac Getz and Brian M. Carney popularised the concept in 2009 with Freedom, Inc., using the French foundry FAVI as the founding case. Getz also calls the model the F-form, short for freedom-form organisation.

What are the disadvantages of a liberated company?

Field research points to five recurring risks. Informal power replaces formal power, and nobody can challenge it. The leader's figure grows stronger rather than fading, because every arbitration returns to them. Middle managers lose their role without receiving another one. Workload and psychosocial risk rise when new responsibility arrives without extra resources. And transparency can turn into peer surveillance that weighs more than the hierarchy it replaced.

Why do liberated companies fail?

Failure rarely comes from the principle of autonomy and usually from how it is rolled out. Thierry Weil and Anne-Sophie Dubey observed leaders who announce letting go while keeping control, big-bang switches instead of experiments, managers left without support, limits of autonomy that were never made explicit, and external communication creating expectations nobody can meet. An organisation that removes its structure without writing roles or a decision rule ends up with a vacuum, and the vacuum fills itself.

Is Zappos still a liberated company?

Zappos adopted Holacracy in January 2014 and offered a buyout in March 2015 to anyone unwilling to commit, which 18% of employees accepted. The company moved away from the model over the following years and now keeps only parts of it. Medium abandoned Holacracy in 2016 and GitHub reversed its manager-free structure, which makes Zappos the most documented case in a broader pattern rather than an isolated one.

What is the difference between a liberated company and Holacracy?

The liberated company describes a philosophy and a leadership posture, with no prescribed process. Holacracy is a written constitution defining roles, circles, governance meetings and the decision method in detail. The first says what to remove, the second says what to install instead. Many organisations borrow the purpose of one and the mechanics of the other.

How many companies have actually succeeded with this model?

Very few have been studied properly. Economist Thomas Coutrot noted in 2018 how little rigorous evaluation exists, and the same handful of companies gets cited repeatedly. The most substantial evidence base is the 2020 La Fabrique de l'industrie study covering ten organisations between 50 and 1,300 employees, which found real gains alongside real failure modes.

What to take away

The liberated company asked the right question. An organisation where people decide about their own work moves faster, filters less information and keeps its people longer, and French field research confirms that on the cases it studied. The answer it offers stays incomplete, because it specifies what to remove and leaves culture to handle the rest.

Organisations where autonomy still holds ten years later all did the same thing. They wrote their roles, set a decision rule, and kept a space for the structure to evolve. That work is less spectacular than scrapping the time clocks, and it is the part that outlives the founder.

That is what Rolebase makes concrete: a living org chart where every role carries its purpose and accountabilities, structured meetings, and governance decisions kept where the team can find them. Explore what Rolebase does, or create your organisation for free up to five active members.