Organisational Markov Blankets: The Silo Dilemma

Tom Fisk on Pexels

“Good fences make good neighbours.” — Robert Frost, ‘Mending Wall’, 1914

I’ve been travelling extensively recently, delivering workshops and keynote speeches. Flights offer a unique opportunity to read with minimal digital distractions. During my descent into Frankfurt on a recent flight, I was captivated by the rich network of agricultural boundaries sprawled out below. Coincidentally, I was engrossed in chapter 7 of Mark Solms’ “The Hidden Spring,” which introduced me to Markov blankets.

A Markov blanket, a concept rooted in statistics and machine learning, named after the Russian mathematician Andrey Markov, is similar to the patchwork quilt of fields I observed during my descent into Frankfurt. It is a network of interconnected elements, much like the cells in our bodies. In this network, each element is distinct, yet the elements surrounding it exert influence. The Markov blanket is the boundary separating an element from the rest of the network. It’s akin to a protective shell, enabling us to focus on a specific element without becoming entangled in the complexity of the entire network.

In a self-organizing system, a Markov blanket distinguishes what’s inside from what’s outside, defining the system itself. It consists of two parts: one dependent on external factors (sensory) and one independent (active).

This division creates a feedback loop. External factors impact the internal through sensory components, and the internal can, in turn, affect the external through active components. This mirrors how living organisms use sensory input to make decisions — for example, feeling hot and removing a sweater or perceiving danger and responding accordingly, as our brains rely on sensory input since they cannot “see” directly. A Markov blanket provides insights into how self-organizing systems function and how they bridge the gap between the internal and external worlds.

The Protective Function of Silos

“Silos are a necessary part of large organisations, but they become problematic when they act as barriers to communication and collaboration.” — Stanley McChrystal.

Silos, much like Markov blankets, play a crucial role in large organisations. They serve as functional units that optimise performance by providing protection and focus. However, while they serve an essential function, silos can become problematic when they are barriers to communication and collaboration, preventing a holistic view of the organisation’s goals and objectives.

Born from the agricultural world, they were used to store grain separately to avoid contamination. Organisational silos, in essence, resemble the protective silos used in agriculture to store grain separately. Just as ancient farmers built cylindrical silos to shield their grain from contaminants, large organizations use departmental silos to keep teams focused, efficient, and specialised. Organisational silos compartmentalise information flow in a vast and complex organisational landscape. As industries expanded and became more complex, silos evolved as a necessary construct to manage large-scale operations. Managing a global organisation requires a certain level of compartmentalisation.

Like Markov blankets, silos are not bad; they are designed to optimise performance but become problematic when they optimise suboptimisation. Organisations suffer when delineating between departments leads to a distorted or absent “single version of the truth” (SVOT). When a silo mentality pervades large organisations, it leads to an isolated view of work, a lack of communication and team spirit and ultimately, declining organisational performance.

Optimising organisational Markov blankets becomes a Herculean task in a world of rapid change, with multiple generations in a hybrid workplace amidst volatility, ambiguity and even war. One solution lies in the ‘team of teams’ approach championed by General Stanley McChrystal.

The Team of Teams Approach

“Today’s rapidly changing world, marked by increased speed and dense interdependencies, means that organizations everywhere are now facing dizzying challenges, from global terrorism to health epidemics to supply chain disruption to game-changing technologies. These issues can be solved only by creating sustained organizational adaptability through the establishment of a team of teams.” ― General Stanley McChrystal

To address the challenges posed by organizational silos, Stanley McChrystal introduced the revolutionary ‘team of teams’ concept. This approach is akin to creating a translucent Markov blanket for constant communication, collaboration, and mutual understanding within large organisations. Unlike the silo system, where departments work in isolation, the ‘team of teams’ approach encourages cross-functional collaboration, fostering interconnectivity among various parts of the organisation. It’s about working together and understanding and learning from each other.

Through the lens of a ‘team of teams’ approach, the organisation’s purpose is not lost in the segregation of tasks. Instead, it’s enriched by all teams’ shared mission, vision, agreed values and behaviours.

Beyond the military, former director of Research for Accenture, Paul Nunes found that the top-performing companies used a similar approach to avoid the temptation for executives to form alliances and engage in backroom horse-trading regarding strategy and collaboration. Paul’s team found that high performers create a nested series of teams inside of teams (separated by “leaky” Markov blankets). Top performers shrink the top team to a small number — maybe between three to seven people — who are the primary decision-makers. This small group then receives advice from other teams so that hundreds of people might provide significant input. According to the latest Outthinker research, our future guest and friend of the show, Kaihan Krippendorff found that small, decentralised teams are essential to achieve the speed and agility necessary for innovation in a large company.

In my workshops, I ask an organisation to consider everyone as a sensor. Using the metaphor of the Markov blanket, while they are separated for concentration, they must remain connected for communication. Those developing strategies must know precisely what is happening in “the real world” with suppliers, customers and the industry. An organisation needs a method for its people to share vital information. The most critical data often flows far from the C-suite, and by the time it reaches the strategy deck, it has been filtered, redacted and neutered. As Andy Grove brilliantly said, “Snow melts first at the periphery because that is where it is most exposed”.

THANKS FOR READING

Our final episode of the Paul Nunes series is now live.

https://medium.com/media/2cdefc5968fed86b0d56fc79c1fc0d75/href

We also hosted Gautam Mukunda on his book “Picking Presidents”. Gautam mentioned his Stan McChrystal, which inspired me to write this article.

That episode is here:

https://medium.com/media/702ad064e0d419a6dd54fa7c7545f472/href

Organisational Markov Blankets: The Silo Dilemma was originally published in The Thursday Thought on Medium, where people are continuing the conversation by highlighting and responding to this story.

The post Organisational Markov Blankets: The Silo Dilemma appeared first on The Innovation Show.

Leave a Comment

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