Operational Excellence is the Pursuit of ‘Knowledge’ - Commoncog

It turns out that operational excellence results from the pursuit of a certain form of knowledge. This is Part 3 of the Becoming Data Driven series, and the result of a deep dive into the field of statistical process control.

This is a companion discussion topic for the original entry at https://commoncog.com/operational-rigour-is-the-pursuit-of-knowledge/

Stratechery’s Ben Thompson recent reminded me of Jeff Hawkins’s theory of intelligence that he describes in A Thousand Brains: A New Theory of Intelligence:
“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 read Hawkins’s first book On Intelligence, but got distracted after starting A Thousand Brains, but moved it back to the top of my reading queue after seeing that excerpt.

Thompson was using the Hawkins theory to talk about ChatGPT, but it also applies here. If thinking is prediction, and management is prediction, what you are describing with SPC is a form of organizational cognition, the way in which the organization thinks by describing a standard way to make predictions about critical company functions, and a way to update those models in response to reality.

Organizational cognition is one of those areas I always meant to expand upon further (Edwin Hutchins’s book Cognition in the Wild is a seminal text describing how a ship crew collectively navigates even though no one single person has all the data) but got distracted e.g. Cognitive Theories of Corporations – Eric Nehrlich, Unrepentant Generalist looks like my last post explicitly on the topic.

1 Like

I’d never thought of linking the SPC mantra of ‘management is prediction’ to the notion of ‘our brains are prediction machines’, but now that you’ve pointed it out, I really like it.

One observation on organisational cognition that I really like is Gary Klein’s “organisational cognition is easier to study than individual cognition, because you can observe organisational cognition in action.” And it immediately came to mind when reading your blog post.

Good connection.

1 Like

Ooh, I hadn’t seen that Gary Klein quote! I like it, and not just because I’m a huge fan of Gary Klein’s NDM. It’s always been a little surprising to me how people don’t see the cognition system of organizations but I guess my brain is wired differently.

1 Like

A lot of parallels here with Deutsch/critical rationalists, good stuff

1 Like

Ooh, could you expand on this a little more, @uberstuber?

One implication of the process control approach is that it cannot work for the latter sort of processes.

In No Silver Bullet , Fred Brooks points out that the complexity, conformance requirements, changeability, and invisibility of software results in inherent and essential variation that cannot be removed.

I’m skeptical whenever I hear absolute statements stating that something is impossible, particularly when a natural law isn’t being violated. Sure, businesses that rely more heavily on creativity and human collaboration can have higher variation and be more difficult to predict, but I disagree that it’s binary.

In systems-thinking terms, there’s nothing fundamentally different between a system that produces a physical product via some manufacturing process vs. one that produces a non-physical product via some creative human process. The differences are largely in the input metrics used to model the systems. In the latter case, input metrics may be more complex and less causal because of the larger creative/human element. In an SPC chart, this may show up as a wider confidence band between the upper and lower bounds.

In this case, input metrics should be chosen more thoughtfully to avoid unintended Nth-order effects (ie. Goodhart’s Law). For example “cycle time” (time elapsed between starting and delivering a unit of value) is probably a better measure than “number of commits/lines of code”. The DORA report offers several other empirically-tested metrics.

In short: Yes, I firmly believe that SPC can be applied outside of manufacturing as long as you understand the nuances and relationships in the system that you’re modeling.

1 Like