Hierarchies vs. Networks

One of my personal passions is the emerging world of cryptocurrency.

I find it fascinating to watch people experiment with the premise that even something as fundamental as our monetary systems does not necessarily need a centralized backer.

Indeed, decentralization seems to be the message of the moment, and cryptocurrency is just the latest expression of a cultural shift away from hierarchical, top-down models toward networked, peer-to-peer systems. We’ve seen it in businesses like Uber and AirBnB; we’ve seen it in open-source software. Why not currency too?

As an organizational innovator, these ideas hold great appeal for me.

But one thing I find to be ironic is that so many of the companies that are promoting these decentralized, peer-to-peer networks are themselves organized as top-down, command-and-control hierarchies.

I had initially wondered if the decentralized vision informing their work would lead them to reshape their internal organizational structures as well. But with the exception of a few outliers, most remain entrenched in a fairly traditional model.

This is, to some extent, quite understandable to me.

Many people fear that if they were to apply a networked, decentralized model to their organizations, they would lose the ability for the company to act as a singular, unified entity, without internal conflicts of interest.

The conventional organization benefits from this singularity in many ways.

As the pioneering economist Ronald Coase observed back in 1937 in his landmark paper “The Nature of the Firm,” there are good reasons we have companies and not simply markets. One of these reasons is the high transaction costs that are associated with independent agents having to negotiate and protect their own interests at every step of the way.

Companies with shareholders to satisfy and targets to meet simply can’t afford to lose the efficiency and coordination that comes from being singular.

That said, there are some pioneers in the space that are experimenting with radically decentralized, networked models for organization, such as Aragon and Colony. In many ways, they aim to replace the traditional, top-down, command-and-control hierarchy with a peer-to-peer network of self-managing individuals who act like sole proprietors of their own businesses. Often, these kinds of pure-network organizations have individuals setting their own compensation, controlling their own profits, and negotiating with the larger network while protecting their own interests. They are economically incentivized to be connected to the network, but they remain fundamentally independent.

Does it work?

Can a completely flat, decentralized network run a business effectively without getting stuck in a gridlock of personal interests, or bogged down in the inefficiencies firms were designed to solve?

I think it’s too early to pass much judgment about whether they will ultimately prove successful. But there’s no doubt they face significant challenges, particularly at scale. I’ve noticed that flat networks often have a hard time acting as a singular entity, even when that would actually be useful, which means their capacity to manage complexity beyond a certain level can become strained.

I don’t believe, however, that the completely flat network is the only way to embrace a decentralized, peer-to-peer ethos in the way we organize. In fact, it is possible to preserve all the advantages that come with having a “firm” — a singular organizational entity — while leaving behind a top-down, command-and-control hierarchy.

I think we can have the best of both worlds: the flexibility and agility of a decentralized, peer-to-peer network and the simplicity and scalability that comes from having a unified organization.

This is a goal of many organizations in the self-management movement, or “teal” organizations using author Frederic Laloux’s evolutionary categorization scheme. Their organizational forms vary in their details, but what they have in common is that they attempt to fundamentally change how we organize, without getting rid of a singular overall entity in the process.

To make this approach work, at scale, there are two fundamental innovations that I believe are essential.

The first is differentiating role and soul.

In other words, ensuring individuals show up not as self-interested parties but as fiduciaries of the organization itself, like trustees of a trust fund. Their job is not to represent their own interests, but to steward their particular roles in the best interests of the organization. In order for this to work, each individual needs to have an agreement with the organization in which they know their personal interests are taken care of, in the form of compensation or profit share. But with this in place, they need to be asked to then set aside their personal interests and instead advocate for the interests of their role and the organization as a whole.

When people are clearly signed-up for representing the organization’s interests vs. their own, this lends itself to a more cohesive organization, with less risk of getting bogged down in navigating internal conflicts of interest between different self-interested individuals. This gets you some of the key benefits of a centralized firm, but without the hierarchical downside — if those roles are themselves organized in a network vs. a command hierarchy.

The key here is that it’s not a network of people; it’s a network of roles, with people serving effectively as fiduciaries for those roles. Every role has its own property, power, agreements with other roles, and so on.

The second key is to organize these roles not just in a flat network, but in a “holarchy.”

In a holarchy, roles are grouped and integrated into larger cohesive functions, which themselves can act just as singular roles within a yet-larger networked system.

A good analogy might be one you’re quite familiar with: the human body. If you observe the body from the outside, it looks and acts like one cohesive entity. But if you look inside, do you find a command and control hierarchy of cells directing other cells, all ultimately directed by a CEO cell? Not really. You find a network of autonomous self-organizing entities distributed throughout the body, each with the capacity to take in messages, process them, and generate output. And each of these entities, say the body’s organs, are themselves made up of yet more granular self-organizing networked entities (cells). The further you drill down, the more peer-to-peer networks you find, but the further you step back, the more unity you see.

The organizational theorist Karl E. Weick once observed: “The real trick in highly reliable systems is somehow to achieve simultaneous centralization and decentralization.”

This is the challenge and the vision that has driven my own work for the past 15 years, as I’ve developed and evolved the self-management system Holacracy. Holacracy seeks to allow organizations to harness the benefits of both centralization and decentralization, without having to choose between them and accept the inevitable trade-offs.

The result is a totally different kind of organization — one that looks and behaves like a single emergent organism, built from the integration of peer-to-peer networked entities serving some piece of the whole system, in a pattern that repeats all the way down.

This is nature’s approach to rocking complexity at scale, and I’ve found it works quite well for organizations too — at least if you get the details right.

Interested in checking out self-management?

Check out this post about two potential approaches.