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.

Wow.

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.

Actions

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.

Caveats

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.

1. We experienced this at Undercurrent – while our business is today stronger than ever, which is directly attributable to our adoption of Holacracy, we went through an adoption phase where our revenues were severely impacted…there was just a ton to learn WHILE we were trying to run our business at the same time. ↩
2. Watch Sugata Mitra’s TED Talk, “Build a School in the Cloud,” in full on TED.com ↩
3. ATMO = “According to my opinion“ ↩
4. For a good time, check out this job description for a World Wide Naming Coordinator and Strategist. My favorite of the accountabilities: “Serves as a non-technical administrator of the online HP Name Tool to track and manage process.” This from a company that offers a HP EliteBook Folio 9480m Notebook PC ↩
5. The quotes are from page 143 and 146, respectively. This book – Good City Form – is extraordinarily relevant to org design work. Do pick it up. ↩
6. Page 194-197, Erik Brynjolfsson and Andrew McAfee, The Second Machine Age, 2014 ↩
7. Clay Shirky, in an interview with David Cushman from 2008 ↩