Large organizations seem to have a certain persistent dysfunction, at least the ones that I’ve worked in or with. They are places where decisions aren’t made until there has been intensive calculation of the internal political considerations; where going to meetings is equated with getting things done; and where, as one friend described his firm, “great ideas go to die.”
At long last, now we know why. Actually, the insight comes from an article written almost 50 years ago. But because it didn’t appear in the pages of a magazine like strategy+business or in an academic journal dedicated to organizational behavior, it largely escaped the notice of those who run these big, lumbering organizations.
This article, which was written by an academic named Christopher Alexander—trained in architecture and mathematics—was published in Design in 1966. It’s called “A City Is Not a Tree.” Neither, it turns out, is a large, global organization. Let me explain.
Alexander set out to elucidate why artificial cities designed in their totality (such as Levittown in the U.S. or the New Towns in the U.K.) never seemed to function as well as natural cities that evolved organically over time (such as Paris or New York). The answer, he argued, lay in organizing principles and the limits of the human mind.
A tree, in Alexander’s parlance, is a formal method of organization in which any one part is connected to one or, at most, a few other parts. Think of Darwin’s evolutionary tree, rather than that tall, leafy, green plant you find in a park. That approach, used to design utopian cities, is also ostensibly used in the organization charts of large organizations. On paper, it all looks perfectly organized as field offices report up to country-level management and, in turn, to regional headquarters and then to a corporate head office. That’s how it’s all supposed to work.
Yet rarely are these organizations so neat. Instead, a great deal of communication and action occurs over informal networks that do not appear on any official document, just as much of what happens in a city appears to be spontaneous and unplanned by any central authority. These informal networks constitute what Alexander identified as a semi-lattice structure.
Alexander explained the following exercise, which makes the distinction abundantly clear: Make 20 dots in a random pattern on two pieces of paper. The patterns need not match but they can. On the first piece of paper, draw circles around subsets of dots so that none of your circles cross and each subset is self-contained. This is tree-based organization. Take a moment and try it. Now turn to the second piece of paper. Once again, circle subsets of dots, only this time you’re free to have the circles cross. This is a semi-lattice structure.
You could be circling for quite a while. As Alexander noted, “We may see just how much more complex a semi-lattice can be than a tree in the following fact: A tree based on 20 elements can contain at most 19 further subsets of the 20, while a semi-lattice based on the same 20 elements can contain more than 1,000,000 different subsets.”
We design our organizations as trees, but in reality, they are semi-lattices and function accordingly despite our best intentions with reorganizations, matrix reporting structures, and all the other attempts to regulate and standardize interactions at work. Alexander cited brain research that showed the human mind simply can’t grasp all of the complexity of a semi-lattice arrangement in a single design exercise. Our brains evolved to simplify complexity in order to identify threats in the environment. We can’t design organizations that formally incorporate all their complexity.
So what is an executive to do? The first step is to acknowledge and embrace the complexity inherent in our organizations. It isn’t going away. Organizations are not simple, linear systems (trees); they’re complex, adaptive systems (semi-lattices). The beauty of such systems is that they have great capacity for self-organization. They seek order, yet attempts to impose order through control generally fail because they use a tree-based approach.
Create as few rules as absolutely necessary. Instead, seek to establish and continually refine clarity in three critical areas: purpose, values, and performance. Such clarity helps set the conditions for self-organization to achieve desired outcomes in the semi-lattice structure:
• Purpose, in this sense, does not mean the firm’s existential reason for being. It means really knowing who your customer is and what job they’re hiring you to help them do. How are you useful?
• Values are those enduring commitments you make to the way you’ll do business as you pursue that purpose. How are you respecting your stakeholders?
• Performance is how you’ll measure success, and should include measures of performance against purpose (such as customer satisfaction or innovations brought to market, and how your customer measures success) and values, as well as financial metrics. How are you contributing?
Most companies obsess over performance, particularly financial performance. The root cause of performance shortfalls can often be found in the lack of clarity around purpose and values. Are purpose and values regular factors in your decision making and reward allocation? Or does hitting the earnings-per-share each quarter trump all?
Never mistake a map for the terrain. Your organization is not a tree. Expecting it to function as one is a sure path to dysfunction.