The decentralized organization: lessons from 3 German pioneers
Last week I spent a few days in Germany's Rhein-Ruhr area, in the middle of a 39°C heatwave, visiting three pioneering decentralized organizations most people outside Germany have never heard of.
That's the strange thing about Germany. It's a bit of a black box when it comes to new ways of working. A lot is happening behind the language barrier. Very little of it travels beyond the country's borders.
If some of these stories had been Dutch or American, everyone would already know them. Instead, they remain quietly remarkable, known mostly to those who go looking.
Jochen Goeser, one of the coordinators of our German Rebel Cell, is one of those people. He's been tracking these stories for years, and he took me to see them firsthand: a social-care non-profit in Mönchengladbach, a telco brand inside a corporate giant in Cologne, and a sustainable bank in Bochum.
Three very different worlds. One shared lesson. A decentralized organization is not the absence of structure. It's a different, and often more demanding, kind of structure.
Here's what I saw.
AWO mönchengladbach: decentralized organization in social care
If you wanted to name the least likely place for radical self-organization to take root, you might land on the German social economy. It's heavily regulated. Publicly funded. Bound by collective wage agreements. Shaped by a deep culture of caution.
As our host Uwe Bohlen, CEO of AWO Mönchengladbach, put it: it's an industry run by the stopwatch, where care is paid for by the minute and documentation often matters more than the work itself.
Familiar with our Masterclass? 10+ years of research, 200+ pioneering companies visited. Distilled into proven ways to organize differently.
Which is exactly what makes AWO Mönchengladbach so fascinating. With around 300 colleagues across some 35 locations and six organizations (child care, home care, youth work, counselling, integration, and education) they have rebuilt their entire organization from the ground up.
They did it on their own steam. No consultants, barely any outside help. They read, experimented, and learned in public, deeply inspired by Niels Pflaeging and Silke Hermann's BetaCodex approach.
The half-measures trap
The biggest lesson Uwe shared was hard-won. Radical decentralization only works all the way, not halfway. Their first attempts at half-measures didn't carry. Mixing a bit of this model (a piece of Holacracy) with a bit of that one (a piece of Buurtzorg) left everyone unsure where to go.
What finally worked was a clear framework, consistently held.
Three structures, not two
The conceptual heart of their story comes from Mary Parker Follett, writing in 1920s Chicago. Every organization, she argued, has not two structures but three:
the formal structure (leadership of position),
the informal structure (leadership of personality),
the value-creation structure (leadership of function).
The last one is where the real work is done. Where reputation and mastery are the true currency. Most organizations let the formal structure run the show. AWO deliberately moves decisions into the value-creation structure instead.
In practice, that shows up everywhere:
Most decisions are made by consent. Whoever knows best consults the right people, decides, and makes the decision visible.
Bigger, more complex questions go to a mastery circle of no more than eight people with relevant expertise and experience.
There are no team leaders and no consensus-seeking.
Pay follows the role, not the position.
Every cell sees and steers its own economics, and sees that of the other cells, with the goal to compare their performance rather than rank it.
New colleagues are hired through peer recruiting across different fields, precisely so people don't end up hiring others just like themselves.
Uwe was blunt about accountability. Only individuals can carry responsibility. A team that is collectively responsible, he said, is a team where no one feels responsible.
AWO Mönchengladbach is simply one of the best pioneering workplace examples I've seen in years.
Joost discussing decentralized organizations during the trip to Germany.
congstar: decentralized management inside a corporate giant
From the social economy we moved almost to the opposite end of the spectrum. A commodity business inside a corporate giant.
congstar is a fully autonomous daughter of Deutsche Telekom. A German mobile brand serving around seven million customers, built up over nearly two decades from a founding team of about twenty people.
What makes congstar interesting is the question it answers: can an autonomous daughter sitting inside a large corporate organize itself in a genuinely progressive way?
The answer is yes, and built on:
a small set of values (fairness, diversity, community),
a flat structure of just three hierarchy levels,
a mix of stable "streams" and temporary cross-functional teams,
a decision-making logic of consent and a "safe enough to try" mentality.
The prison of politeness
They were also clear about the hard parts. Something I've come across many times before.
A friendly, almost bro-like culture can have a shadow side. Conflict tends to be handled emotionally, or avoided altogether, precisely because people see each other as friends. Lisa Gill calls this the "prison of politeness."
Building the muscle for direct feedback and healthy conflict is one of the solutions. And for many progressive companies it's a constant work in progress, a challenge we hear again and again across the Rebel Cell network.
Is decentralized management good or bad? The honest answer is: it depends on whether you're willing to do this work. Removing hierarchy without building the capacity to disagree well is a recipe for quiet dysfunction.
GLS bank: freedom within a form
The third visit took us to GLS Bank, Germany's pioneering sustainable bank. Founded in Bochum in 1974, shaped to its core by its anthroposophical roots. It grew, improbably, out of an effort to build a school, and the founders have always been slightly reluctant bankers. The bank is the vehicle. The real goal is changing society.
You see that conviction in how seriously they take their principles. GLS publishes every corporate loan it makes, every quarter. That's radical transparency that banking secrecy law technically forbids, so they simply ask borrowers to sign a waiver. As a customer, you can see exactly where your money is at work.
The human-centred philosophy runs deep. People are deliberately called co-workers, not staff. There is a peer trust fund for colleagues in financial difficulty, paid psychological support, and a genuine effort to treat people as whole human beings, even when parting ways. More than half of their leaders are women, the result of a deliberate goal set a few years ago.
The failure that taught them
Their reflections were refreshingly self-critical, and echoed AWO almost word for word.
An early "new ways of working together" experiment removed hierarchy without putting a clear alternative structure in place. It failed. The lesson they drew is one we keep coming back to: horizontal, self-organizing work needs more structure, not less.
The concept that solved it for the bank was framed as "freedom within a form." There is a management system. There are rules. But within that form there is a great deal of freedom, and a habit of stretching those boundaries as far as possible.
It's the same principle EPPO calls "freedom within a framework." Different words. Same insight.
What these decentralized organizations have in common
Three organizations. A social-care non-profit, a telco brand inside Deutsche Telekom, a values-driven bank. Completely different industries, sizes, and origins.
And yet the same thread kept surfacing.
- A clear framework, consistently held. AWO learned that mixing Holacracy with Buurtzorg left everyone confused. GLS learned that removing hierarchy without an alternative structure failed. congstar built a flat structure with three levels, stable streams, and consent-based decisions. In every case, the answer was clarity, not eclecticism.
- Decisions made where the value is actually created. This is Mary Parker Follett's value-creation structure, a century old and still radical. AWO moves decisions into it by design. GLS does it through freedom within a form. congstar does it through cross-functional teams and consent. The formal hierarchy is not where the real work happens, so it should not be where the real decisions happen either.
- The hard parts are real and ongoing. Conflict avoidance. Accountability gaps. Leaders who instinctively step in and steal responsibility from teams. None of these organizations pretend they've solved it. All of them are still working on it.
What strikes me most is where these organizations sit. A regulated social-care non-profit. A subsidiary inside one of Europe's largest telcos. A bank. If a decentralized organization can work here, the "we're too regulated" or "we're too big" excuse gets a lot thinner.
How to build a decentralized organization that actually works
A few practical takeaways I'm leaving Germany with:
- Go all the way, or don't start. Half-measures create confusion. If you're going to move to decentralized decision making, commit to a coherent framework and hold it. Cherry-picking from three models is how you end up with none.
- Design for structure, not against hierarchy. "Freedom within a form" is the operating principle. The form is not the enemy of freedom. It's what makes freedom possible.
- Make decisions visible. Consent-based processes where whoever knows best consults the right people, decides, and then makes the decision transparent. This is AWO's default, and it works because everyone can see what's been chosen and why.
- Build the conflict muscle. Friendly cultures slide into avoidance. Direct feedback is a skill, not a personality trait. Train it, practise it, keep training it.
- Assign individual accountability. A team that is collectively responsible is a team where no one feels responsible. Distribute ownership to people, not to abstractions.
- Let cells see and steer their own economics. Peer-based hiring, peer-based comparison, peer-based learning. When teams can see their own performance and the performance of others, they don't need a manager to tell them what to fix.
None of this is easy. None of it is quick. All of it demands more from people than the org chart they were handed on day one.
Germany's black box is full of stories like these. We'll keep opening it.
Want to know everything we learned after 10+ years of studying progressive organisations? Check out the Masterclass.
Learn directly from pioneering companies Buurtzorg, Haier, and Viisi's own leaders. Walk away certified, with a lifetime network of 1,200+ pioneers in 38+ countries.
