Organizational Cognition Revisited

[n.b. This post got completely out of control (2200 words!) as I tied together way too many ideas with too little explanation as I relied on heavy hyperlinking. Skip to the end for a couple actionable takeaways.]

I recently read Jeff Hawkins’s book A Thousand Brains after Ben Thompson referenced it in a Stratechery update with this quote:

The brain creates a predictive model. This just means that the brain continuously predicts what its inputs will be. Prediction isn’t something that the brain does every now and then; it is an intrinsic property that never stops, and it serves an essential role in learning. When the brain’s predictions are verified, that means the brain’s model of the world is accurate. A mis-prediction causes you to attend to the error and update the model.

I appreciated Hawkins’s first book On Intelligence and I had even bought A Thousand Brains when it came out, but had lost it in the morass of books in Kindle.

The critical piece Hawkins adds to his cognitive theory in this book is that the brain is built around “reference frames”, aka the mental model or sense making we have of the world. In other words, he posits that thinking itself at the neuron level is about navigating with respect to a mental map we have of the environment and objects around us. We are constantly refining those mental maps (stored on a per object, and per domain, basis by individual “cortical columns”, clusters of neurons) by predicting what we will sense next as we navigate using that map, and update the map in response to any surprises.

In other words, humans never experience incoming sensory data as is, but are instead always interpreting that data relative to our previous experiences. This means we can never have an unbiased or objective experience, one where we experience something with no preconceptions. The only exceptions I can think of are a newborn with no previous experiences, or a highly trained meditator who has learned to be completely present to their current experience. This is relevant to my previous post, where Tim Urban presents a well-reasoned case based on what he knows, but is completely limited by his own experience as a wealthy white man and the experiences of other wealthy white men (college professors); they literally don’t see other people’s experiences because those experiences aren’t represented on their mental map.

Note that we don’t just have a single mental model or reference map. Hawkins estimates that we have thousands of cortical columns in our brains, each of which can hold reference frames for multiple objects and multiple situations; hence the title “A Thousand Brains”. He theorizes that they are each trying to match the incoming sensory data, and then “vote” on which existing model seems most relevant; this matches my theory of cognitive subroutines where we can regularly mis-apply previous models without even noticing that we are using them out of context. Only by consciously paying attention using System 2 thinking can we notice those mistakes and then re-train the brain through deliberate practice. We can also address the biases of the previous paragraph by increasing our diversity of experiences and people so that we have more maps with which to interpret our experiences.

This cognitive theory makes intuitive sense to me, as they match how I experience the world, and align with how I work with clients as a coach. My clients are often stuck because they are using a mental model that no longer applies to their current job, and I help them develop a new mental model of their job, and then offer experiments for them to test that new mental model. Once those experiments confirm a new possibility, we develop practices that allow them to retrain their brain to apply it through unconscious ease (aka System 1) rather than conscious effort (aka System 2).

We can apply these cognitive theories to organizations as well. I had dabbled around in the idea of organizational cognition back in 2008, building on the ideas of Edwin Hutchins in his book Cognition in the Wild where he observed how a ship crew collectively navigates even though no one single person has all the information. Using the Jeff Hawkins theory, the crew is building a shared reference frame of where the ship is, making predictions based on that frame, and taking new observations to test that prediction. A crew is more effective when they have the psychological safety to speak up when new data doesn’t align with the expected prediction; rather than dismiss it, they discuss it and build a new shared understanding of the situation. The alternative is described in the opening chapter of L. David Marquet’s book Leadership is Language where the crew of the container ship El Faro did not update their shared understanding of the situation as a hurricane approached, and all aboard died.

I was reminded of these musings recently by a Commoncog post on Operational Excellence where he repeated a principle of Statistical Process Control (SPC) theory: “Management is prediction”. Since I had just read the Ben Thompson piece citing Jeff Hawkins, I made the connection that if thinking is prediction, and management is prediction, then management is a form of organizational cognition, describing how leaders make predictions about critical company functions, and how they update their models in response to what actually happens in reality.

In this context, effective organizational leadership requires building a consistent reference frame for people in the organization so they are using the same mental map to interpret incoming signals. They are trained on it through repeated scenarios and deliberate practice, so that their brains learn to unconsciously navigate using that reference frame; everybody looking at the data is making similar predictions of what will happen next because they are using the same reference frame. CEO coach Matt Mochary even recommends that every new hire shadow existing leaders to learn how to do their job, similar to the “watch one, do one, teach one” method of teaching surgery. An outsider might call it “drinking the Kool-Aid”, but agreeing on what’s important to pay attention to and what can be ignored is a necessary part of forming an efficient collective brain.

This also explains why Patrick Lencioni claims in his book The Advantage that the most effective leadership teams 1) Create Clarity, 2) Overcommunicate Clarity, and 3) Reinforce Clarity (by making all processes like hiring and promotions consistent with the communicated priorities of the organization). Through the lens of this theory, such leaders intentionally create a common reference frame in people’s brains, which means that they are collectively using the same reference frame to understand the incoming data, and can therefore decide on a response faster, because they need to spend less time making sense of the situation. Organizations that react faster have an advantage over other organizations, which might explain Lencioni’s claim that “The single greatest advantage any company can achieve is organizational health. … when it is whole, consistent, and complete, that is, when its management, operations, strategy, and culture fit together and make sense.”

The Hawkins theory also points to why psychological safety is key to effective teams, as demonstrated by Google’s research. If the situation is changing such that the organization’s reference frame needs to be updated, the incoming data will not match the prediction, and the first person to notice it will likely be a low-level person on the front lines. An environment of psychological safety is necessary for their observation to be listened to, and not dismissed by others as “you don’t understand how things work around here”. Treating each person as an individual “cortical column” with its own model of what’s going on, and including all of their inputs, leads to better decision making, because you’re more likely to get early warning signals of mismatches between predictions and reality. If all such observations are filtered through a person that “knows better”, then those signals may be prematurely dismissed.

Connecting these different areas of cognitive theory, organizational behavior, and leadership is all very interesting, but what are the practical implications besides reinforcing the importance of psychological safety?

One of my takeaways is that this theory may help explain the value of overcommunication to an organization. In the brain, neurons are constantly sending signals to each other as they process new information from moving through the world, and they are densely interconnected both near and far to exchange those signals. In hierarchical organizations, communication is limited to increase efficiency, but this reduces the opportunities for both bottoms-up learning (when reality isn’t matching leadership assumptions) and top-down context setting (aka reference frame building). Creating more (and random) connections between different parts of the organization or between executives and front-line workers may help the organization “think” better. This would come at the potential cost of efficiency or productivity, but my belief is that organizational alignment adds significant value in increasing opportunities and shared context building. The best leaders I’ve seen are constantly seeking out viewpoints from different parts of the organization to test their understanding and improve their decision making.

The SPC principle that kicked off this whole cascade of connections for me was “Management is prediction” with the implication that “in order to run your business properly, you need to be able to predict — within limits — the effects your actions have on business outcomes”. If we take that seriously, then developing a shared understanding of how our actions affect critical metrics is the key to effective organizational cognition. This is why companies like Amazon have Weekly Business Reviews as a forum to build that shared understanding, and offers insight on why that might be a valuable practice for other organizations to adopt.

My own experience with this was in the “Revenue Force” meeting at Google, described in this New York Times article. I started at Google on the revenue forecasting team; that forecast was used by the leadership team to decide how much capital they would have available to invest in various areas. I developed a keen intuition of how the myriad moving parts (pretty much everything in the world can affect Google’s revenue) translated into results. For most of my ten years at Google, I participated in a weekly revenue meeting with Google leaders to discuss what was happening with revenue. In terms of this theory, that meeting is where we developed a consistent reference frame to understand revenue, as we analyzed where reality deviated from the finance team’s forecasts and determined whether the forecast was missing something and needed to be updated. What I particularly loved about that meeting is that all input was welcome; a junior finance analyst could point something out to the SVP of Ads and be taken seriously if they added insight.

In fact, that experience was what started me thinking about organizational cognition in the first place; I was sitting in these meetings and developing my understanding of how Google leaders thought about revenue, which enabled me to get up to speed quicker as an analyst. Using cognition as a framework might offer new possibilities for onboarding, in more explicitly designing ways to develop the understanding of critical metrics and the common ground of principles to collectively make decisions. Given the emphasis on prediction in this cognitive theory, perhaps designing a structured set of predictive exercises against previous data might be a way to develop that understanding.

To sum up this beast of a post, here are two actionable takeaways and something to think about:

  1. Seek out alternative interpretations or reference frames. Charlie Munger once said “I never allow myself to have an opinion on anything that I don’t know the other side’s argument better than they do.” Or, as I wrote last year, realize that you don’t have the answer, just your answer based on your life experiences. When people complain about somebody acting in a way they don’t understand, I ask them “What must be true in their world, but not yours, that would explain their actions?” as a way to help them build new reference frames.
  2. Be thoughtful about what actions you consistently repeat, as actions reinforce the perspective that generated them. Doing something repeatedly makes it happen more efficiently as an unconscious habit, and is a vote for your identity; this applies both as an individual and as an organization aka What You Do is Who You Are.
  3. As a leader, consider the tension between efficiency and learning. Building a common reference frame where people across the organization interpret events consistently leads to efficient decision making and quicker action. At the same time, too rigid a reference frame leads to dismissing evidence that the world is changing and that leads to ruin; ideally, you can build psychological safety in the organization so that people can safely point out different interpretations and possibilities for the leadership to update their mental models. I’ve explored this tension before in Learn and Latch, and it also shows up in Built to Last as “Preserve the Core, but Stimulate Progress“. Being aware of this tension as a leader (and in one’s own thinking) is key to balancing the desire for efficiency and productivity (which can lead to rigid thinking) with the need to keep learning.

2 thoughts on “Organizational Cognition Revisited

Leave a Reply

Your email address will not be published. Required fields are marked *