Open Undercurrent

Let’s talk about transparency. It’s simultaneously the most powerful and most scary of the 21st-Century Org Design principles – it asks those “in power” to put a huge amount of trust in their teams, exposes issues inherent to the business to employees and maybe the world, and tugs at deep-seated desires for privacy and safety. But it also provides a critical ingredient for self-organization: casual enforcement of advantageous behaviors. Organizations usually handle “enforcement” through lots of leadership presentations, “core values”, mission statements, and aggressive review processes. It’d be better if we didn’t have to do any of that, and instead trusted each other to do the right thing.

In the last two months, we’ve become radically more transparent. We opened up Undercurrent’s books to our Members, and developed an opt-in salary transparency system in use by around 60% of the team. Today we’re going even further by opening up our key metrics to the world.

Let’s start with the top line, and work our way down!

Continue reading Open Undercurrent

The Legacy Organizations We Designed 100 Years Ago Are Broken

In the previous post, I outlined a few of the reasons why organizations are forces for evil. With all of those reasons in mind, I’ll ask you now to consider that organizations are our only hope, and the forces that act on them are becoming more favorable by the day.

Before we get into why this is, we should have a collective understanding of what an organization is.

Continue reading The Legacy Organizations We Designed 100 Years Ago Are Broken

The Organization is Broken

I’ve spent the last couple months developing a new piece of work that’s probably 50% complete: an organizational pattern language. I presented the preamble to that work up at #planningness last week, and the talk seemed to go over well. What follows are the speaker notes for the first portion of my deck, where I present the bear case for organizations – that they are powers for evil, despair, and inequality. I’ve gotten a few speaking requests related to this deck, and to be quite clear, I’m more than happy to deliver this talk again to a crowd.

divergence of wages and productivity

Since the early-mid ’70s, wages and productivity – when measured in terms of widgets produced per hour – have been diverging. Wages are generally flat, while workers are close to 2x as productive today as they were 30-40 years ago. This divergence is coincident with the start of the so-called “Third Industrial Revolution,” marked by the introduction of the Programmable Logic Controller (1969), allowing factory-owners to install ever-more-capable robotic production equipment. Things really started to get interesting with the commercialization of the personal computer and the advent of the supercomputer (1976-forward), allowing white-collar work to feel the impact of Moore’s law.

Continue reading The Organization is Broken

Clarity Hurts

Most approaches to organizational design prize clarity of operations above all else. And this seems, a priori, like the right way to go. A robust job description is helpful. Knowing what to do in every situation makes it easier to get things done. Well-defined roles help separate you from your work, enabling personal development.

In a traditional organizational design, every operation and division is intentionally designed using data about past performance and long-range projections for future needs. The clarity at the outset, detailing process designs, job descriptions, meeting rhythms and outputs, RACI/DICE models, etc., makes the new plan sensible by every employee; it’s like great wayfinding in an airport. That said, as new data is introduced, this clarity becomes less valuable, especially with more rigid, centralized governance methods.

Newer methods, like Holacracy, drive organizations to continually refine their design to a point of perfect clarity where each role-filler and team knows exactly what to do – and has the opportunity to change their processes as new data is introduced. Over time, this leads to extraordinarily clear, likely well-organized operations. However, the cost of the ongoing, evolutionary improvement is a potentially disruptive loss of pattern at the outset.1 And things that get more clear over time might not be good, after all.

Clarity makes you a cog

The following explains the allure of clarity pretty well, if a little bleakly. It’s from Sugata Mitra, referring to one of the monuments of the British Empire, the “Bureaucratic Administrative Machine.” I found the text in the Brynjolfsson & McAffee’s Second Machine Age:

They created a global computer made up of people. It’s still with us today. It’s called the bureaucratic administrative machine. In order to have that machine running, you need lots and lots of people. They made a machine to produce those people: the school. The schools would produce the people who would then become parts of the bureaucratic administrative machine. They must know three things: They must have good handwriting, because the data is handwritten; they must be able to read; and they must be able to do multiplication, division, addition and subtraction in their head. They must be so identical that you could pick one up from New Zealand and ship them to Canada and he would be instantly functional.2

I’ve heard this exact idea mentioned in the discussion of the design of several 100,000-person-plus organizations: that members of functions should be immediately transplantable between regions and divisions. This is not inherently bad – some would argue that this is a necessary feature of large organizations – but it speaks to a centralized authority on role, task, process and output that does naturally limit the potential for self-organization, agility, emergence, and empowerment at the edge. Clarity to the level of hot-swappability, then, may be a bad thing for the organization of the future.

And while clarity does help (see the first paragraph), there are diminishing returns. A too-clear job, ATMO,3 where the upper and lower bounds of your authority are clearly defined and rigidly held, is an important part of a shitty, small existence.4 The more disturbing outcome: pure, task-level clarity in your job description means that your work is almost immediately computerizable.

To seal this “cog-making” argument, I’ll offer a segment from Good City Form, by Kevin Lynch,5 about the sensibility of cities. Pattern, transparency, legibility, and structure play an essential part of making cities functional and “good”, but

None of these characteristics, however important, are absolute desiderata—qualities to be maximized. No one would want to live in an infinitely vivid place, where everything is patently connected to everything else. We do not seek an absolute one-to-one correspondence between form and society; we don’t wish to live in a goldfish bowl; we would be overwhelmed by a multiplicity of evocative signs. Human cognition has its limits, and the process of cognition is of greater value than the resulting mental structure. There are pleasures (and there is food for development) in puzzles, ambiguities, and mysteries. We want definable elements rather than defined ones, complex connections, regions remaining to be explored, and some freedom to camouflage. Privacy—the ability to deny information about personal beliefs and actions—is a sensitive issue and a shield against tyranny.


So there are two important qualifications to the ideal of good sense: first, that there are limits at which individuals may wish to deny further knowledge of their affairs, or beyond which the human mind is overloaded, and second, that a settlement should permit an unfolding creation of meaning, that is, a simple and patent first order structure which allows a more extensive ordering as it is fully experienced, and which encourages the construction of new meanings, through which the inhabitant makes the world his own.

Again, wow. Unfoldingness is a really cool idea.

Clarity stunts your growth

Clarity is great for a mechanized, efficient working environment, but it stands in the way of the rapid growth of the skills that will be valuable in the post-industrial economy. If you’re building an organization for the Second Machine Age, those skills include a deep capacity for: large-frame pattern recognition; ideation; and complex forms of communication.6

The children in Mitra’s studies form teams, use technology to search broadly for relevant information, discuss what they’re learning with each other, and eventually come up with new (to them) ideas that very often turn out to be correct. In other words, they acquire and demonstrate the skills of ideation, broad-frame pattern recognition, and complex communication. So the “self-organizing learning environments” (SOLEs) Mitra observed seem to be teaching the skills that will give them advantages over digital labor.

We probably shouldn’t be too surprised by this; SOLEs have been around for a while, and have produced many people who have excelled at racing with machines…Montessori classrooms emphasize self-directed learning, hands-on engagement with a wide variety of materials (including plants and animals), and have a largely unstructured school day. And in recent years they’ve produced alumni including the founders of Google (Larry Page and Sergey Brin), Amazon (Jeff Bezos), and Wikipedia (Jimmy Wales).

By my reading, a company is better off building a Montessori-inspired training program, or viewing their entire working experience as “training”, than developing a rigidly structured curriculum for their leaders and new employees.

So, to be clear: less clarity means more individual growth, more prep for the post-digital economy.

Clarity creates positive pressure toward the status quo

Having more numerous and lengthier policies (in service of increased clarity), and particularly makes everything harder to change, and requires that everyone who wants to make a change be perfectly versed in what they’re changing. We see this in law, where the highly networked, richly documented nature of legislative and regulatory precedent results in an entire industry dedicated to changing the rules around our lives.

At Undercurrent, we have a collaboratively improved set of rules around our operations. Each team sets their own rules – using Holacracy – and almost everything is changeable, including our employee agreement. Our governance record contains just over 6,000 words. For comparison, the US constitution spans just over 4,400 words. And while the act of writing down and consenting to our own “laws” has had a huge impact on our business, it’s debatable how much our overage makes our document better than the oldest and shortest constitution in the world. But ultimately, length isn’t the enemy: it’s the pressure to avoid changes once something has been comprehensively defined. I’ll reference one of my favorite passages from Clay Shirky:

Over and over again, what we see in interactive environments is that if something looks too good, people won’t touch it. And you can do this in your own kitchen: if you go and spend half a day arranging every single thing, laying everything out just so, like it’s ready for a magazine shoot, and then you send someone into your kitchen, they will not pick up a knife. They will not help you cook. They will not touch anything, because the perfection of the kitchen says, “You don’t belong here.” On the other hand, if your kitchen looks like my kitchen on average, where the recently washed dishes have not yet been put away, and there’s some stuff around, a guest will come in and feel right at home opening up the refrigerator and helping you prep. The messiness, the openness, the human characteristic tells people it’s okay.7

The longer our governance record becomes, the less likely we are to change individual pieces of it.


If you work at an organization that allows for self-governance like Undercurrent:

  1. Carefully monitor the expansion of your governance “codebase”.
  2. Work toward legible human language even over comprehensive legalese (when we write rules, we tend to use words that almost never appear in conversation, like “shall”).
  3. If you use a process for editing your governance, incorporate a “simplification round” into it, where the intent is to eliminate rules that aren’t being used.
  4. Limit the rules at the edge of the organization (where the customers are) in an almost ruthless manner. For example, consider Nordstrom’s only and perhaps apocryphal rule: “Use best judgment in all situations”.

If you work at a legacy organization, my only recommendation is to write down the rules that actually apply to the day-to-day operations of your team. Even if  – no, especially if – those rules are objectionable/make you cringe.


My colleague, Jordan, frequently says that “our enemy isn’t hierarchy, it’s chaos.” In most corporate environments that I observe, it seems like there’s a lot of clarity. Office spaces have nameplates. People have titles. The employee handbook is comprehensively robust and violating the rules gets you fired. Processes are well-defined and decision models are known and “enforced”. Values are encoded and probably on the back of ID cards. And yet, how stuff actually gets done isn’t written anywhere. Nobody can recite the values. Teams don’t know their purpose. Some people are overstepping their remit, and some people are coasting by. The actual processes of the company aren’t very clear at all. I’m not arguing in favor of chaos, nor complete ambiguity. I’m arguing against the desire to encode everything. To make everything perfectly clear. To build an exquisite organization, one where everything has been painstakingly designed. That’s not a world I want to inhabit, and I’d argue it’s one that stunts everyone’s development.

Show Up Whole

I spent the better portion of P11 on employee reviews. At Undercurrent, we do official reviews every four months, with the intent of doing them all in one week while we’re “closed” for renovations. There’s no better gift than being in these reviews, and hearing about everyone’s experiences with the company that we’ve built. Lots of good, lots of challenges, and quite a few revelations that over the past four months, people grew professionally and personally faster than they ever had in their careers.

One thing that’s obvious after doing these reviews is that Undercurrent is a place where by and large people show up as themselves. And as we work to build new ways of organizing and working into client organizations (even smaller, newer ones!), it’s very clear that most workers have a firm partition between their work-selves and their home-selves.

Thinking about this, and fueled by wine and pizza, I did the obvious thing, and took my feels to Twitter:

A quality criticism of this was brought by @malbiniak, that what I’m talking about “seems like an option accessible mostly to people from privilege.”

I think that’s generally correct. The networked, on-demand, everything JIT businesses of the future might be a particularly bad solve for this.2 There are still loads of people working in factories around the world. And more broadly, any organization offering minimum-wage jobs will struggle to allow their employees to show up as 100% of  themselves, all the time, if only because it’s hard and costly to support social norms among a high-turnover working population. I’d argue that they’d be better for it if they did, but that’s an anxiety-inducing proposition for most executives.

This, to me, is the point that I’ve been failing to make understood for the last couple years.

It’s not necessarily that the organizations of our past are ineffective economic engines. In many industries and geographies the established, century-old ways of working and organizing do just fine, at least for the owners of capital.

The problem is that for most people, work sucks. Jobs suck. Bosses suck. Offices suck. Commuting for a reasonable portion of your waking hours, especially by car, is the fucking worst.

That’s what we’re working to fix.

The Main Objective: Legitimacy

When it comes to organizing humans, the only thing that matters is legitimacy.

How Google Works1 devotes a little time to the business of org design, and it’s all well worth your attention. One bit stood out, though. The following is a snippet of an anecdote about a Google Sales team redesign, highlighting a Google re-org principle: “Do all re-orgs in a day.” The idea is that emphasizing speed over perfection leads to higher engagement…not because of the speed, but because legitimacy is the only thing that matters.

The key was doing the re-org quickly, and launching it before it was complete. As a result, the organization design was stronger than initially conceived, and the team was more invested in its success because it helped create the end result. Since there is no perfect organizational design, don’t try to find one. Get as close as you can, and let your Smart Creatives figure out the rest.

That’s *exactly* the idea. A decent organizational design that is created by its inhabitants is better than a perfect organizational design pushed down from senior leadership. Worse is better.2

There’s a fun connection between this organizing principle at Google and a portion of the Counterinsurgency Field Manual:3

The primary objective of any COIN operation is to foster development of effective governance by a legitimate government. Counterinsurgents achieve this objective by the balanced application of both military and nonmilitary means. All governments rule through a combination of consent and coercion. Governments described as “legitimate” rule primarily with the consent of the governed; those described as “illegitimate” tend to rely mainly on coercion. Citizens of the latter obey the state for fear of the consequences of doing otherwise, rather than because they voluntarily accept its rule. A government that derives its powers from the governed tends to be accepted by its citizens as legitimate. It still uses coercion—for example, against criminals—but most of its citizens voluntarily accept its governance.

And this one, too.

The host nation doing something tolerably is normally better than us doing it well

It is just as important to consider who performs an operation as to assess how well it is done. Where the United States is supporting a host nation [HN], long-term success requires establishing viable HN leaders and institutions that can carry on without significant U.S. support. The longer that process takes, the more U.S. public support will wane and the more the local populace will question the legitimacy of their own forces and government. General Creighton Abrams, the U.S. commander in Vietnam in 1971, recognized this fact when he said, “There’s very clear evidence, … in some things, that we helped too much. And we retarded the Vietnamese by doing it. … We can’t run this thing. … They’ve got to run it. The nearer we get to that the better off they are and the better off we are.” T.E. Lawrence made a similar observation while leading the Arab Revolt against the Ottoman Empire in 1917: “Do not try to do too much with your own hands. Better the Arabs do it tolerably than that you do it perfectly. It is their war, and you are to help them, not to win it for them.” However, a key word in Lawrence’s advice is “tolerably.” If the host nation cannot perform tolerably, counterinsurgents supporting it may have to act. Experience, knowledge of the [Area of Operations], and cultural sensitivity are essential to deciding when such action is necessary.[3.

I use that first line – the “tolerably” bit – all the time.

On Transparency and Success

Marco Arment recently posted a year-in-review for Overcast, a podcast app that he designed, built and launched himself. There’s lots of interesting stuff inside the report, especially if you’re interested in the app business.

But what strikes me are the changing definitions of transparency and success, both of which seem to be happening because Internet.1

First, on Transparency:

I’ve decided that the potential educational and market-research benefits to others of adding Overcast to the mix [of small, private companies publishing their revenue figures] will be greater than the risk of people thinking I’m an asshole for doing so.

Most smaller companies aren’t transparent for similarly emotional reasons, but they’ll usually try to rationalize the decision by saying that it’s to throw the competition off their case, or because privacy. These are bullshit reasons for opacity and I’m glad to see that the only thing standing in Marco’s way is that he might look like a jerk.

Second, on Success:

Overall, I’m very satisfied with Overcast’s finances so far. It’s not setting the world on fire, but it’s making good money. For most people, the App Store won’t be a lottery windfall, but making a decent living is within reach for many.

After the self-employment penalties in taxes and benefits, I’m probably coming in under what I could get at a good full-time job in the city, but I don’t have to actually work for someone else on something I don’t care about. I can work in my nice home office, drink my fussy coffee, take a nap after lunch if I want to, and be present for my family as my kid grows up. That’s my definition of success.

The large companies we work with need people like Marco inside their walls.2 But they’ll never be able to; he’s essentially saying that the value of NOT working on something he doesn’t care about is…infinity dollars.

Something to think about.

The Devil’s Bridge

In preparation for my (now annual) European Cycling Vacation, I stumbled across something really fascinating: Devil’s Bridges.1

At the turn of the first Millennium, a series of bridges were built in Europe that were so daring in their construction, and so unbelievable in their span, that they came to be known as Devil’s Bridges; their designers must have made a deal with the Devil in order to pull them off.

And yet, they were made possible – even easy – with the advent of new technology.

Whenever we talk about these companies that seem to do better than their peers, despite being in fast-moving waters, it feels like they’re doing something magic, something that nobody else can do.2

It’s not magic. It’s not their strategy. It’s not their ideas. It’s not even their people.

It’s how they’re organized. It’s the way they work.