Shared Mental Models for Boardrooms

How to keep stakeholders on the same page in times of exponential change

Board members have a duty to keep themselves informed about matters put before them for decision. This duty cannot be complied with unless board members have relevant material upon which to base their decisions. Management provides the Board with information they consider relevant though a dossier (commonly called “board pack”) ~7 days before a board meeting.

The board pack is a key source of information for board members, but also a source of concern for all stakeholders, especially management teams. According to 2017 research by UK governance consultancy Board Intelligence:

• 50+% of the respondents said their pack was 200-plus pages and in extreme examples nudged 1,000 pages
• Directors spent an average of four hours reading a board pack (up 30% from 2011)
• Directors digested 30 pages per hour of information, which implies that majority of the board pack goes unread

Management teams spend ~3–4 weeks preparing a board pack. Management complains about not receiving clear guidance on what to present. Boards complain about the board pack containing too much information and not enough relevant information. To complicate matters further, board and management incentives are not always aligned. For example, growth initiatives (such as an acquisition) generally improve a manager’s career prospects but can destroy shareholder capital.

Even if incentives are completely aligned, agenda is clearly set, templates are agreed upon and board pack quality is excellent, the following challenges remain inevitable:

1. Management’s curse of knowledge: managers spend more time with a problem which causes them to assume many facts as highly obvious. As a result, Management teams struggle to provide boards with sufficient context.
2. Board’s cognitive biases: board members have to contend with a number of biases including Framing Bias, Bandwagon Bias and Pro-innovation bias. As a result, boards struggle to provide meaningful insights.
3. Information’s shelf life: based on my experience of serving on five company boards, I found that Management presented new information on the day of meeting ~80% of the time. This is because content in board papers and pre-agreed templates go obsolete quickly.

Considering pace of change will only increase further, companies need a system that is simple and can evolve as new information emerges. While telepathy through neural implants could solve this problem, the exercise of building shared mental models appears far more feasible. In this post I will:

1. Define what shared mental models are
2. Describe a process to create a shared mental model for boardrooms
3. Share an example of a shared mental model I use for my company NetEquity

What is a Shared Mental Model?

Mental models are hypothesized knowledge structures used to reason about the world, to make inferences based on available information, and to make predictions about future states (Held et al., 2006). The general concept of mental models was extended to teams based on the notion that team members have elements of their individual mental models in common, the shared mental models (Cooke et al., 2003). Academic research points to four types of Shared Mental Models.

Academic research can be painful to internalize. Video game analysis on the other hand offers a more fun way to understand mental models. On August 4 2019, Arslan Ash from Pakistan won the Tekken world championship. I consider myself as a pretty good Tekken player but when I heard game analyses on YouTube about Arslan’s game play I realized how weak my mental model was. It also hit me that Arslan had internalized the complete mental model of Tekken.

The efficacy of shared mental models was confirmed by a 2017 study (Kim, et al., 2017) that took actual League of Legends players in groups of five. This study found that tacit communication where players make sense of other members’ activities and change their own behaviors accordingly was important to victory. This is important when working on frantic, fast paced tasks such as a competitive League of Legends match where there is little chance to communicate with written or even spoken words.

Simply put, companies with shared mental models will win

How to create Shared Mental Models in boardrooms?

Almost all video games come with an engaging tutorial. This tutorial allows a player to become familiar with a game’s environment. A tutorial is not action packed by design because to go fast, you must go slow initially. This concept is best captured by how Josh Waitzkin mastered chess. When Josh started learning chess he only played with a pawn and a king until he had a complete and clear mental model of possible moves. Once he mastered pawn and king moves, he added more pieces (see this video to understand better).

The “pawn-king” mental model

To use the chess analogy further, board packs are akin to management teams taking a chess board to a group of people who know how to play chess but haven’t necessarily mastered how each piece moves. Management teams can spend hours explaining possible permutations OR they can pause and go back to basics i.e. the pawn and king moves. Needless to say a group of chess players with a shared mental model will come up with a better strategy than a group which hasn’t internalized the real significance of each piece.

Mikhail Tal’s mega complex chess board — 1959

Pairing the “pawn-king” mental model with the “cone of possibilities” creates the necessary ingredients to build a shared mental model in a boardroom. Board members are intimately familiar with financial models used to create a cone of possibilities each year. Unfortunately, most financial models are similar to the complex chess-board shown above as opposed to the “pawn-king” board. This is because models suffer from scope creep. It is exceedingly common for a CFO to use the same model for a budgeting exercise as he would to plan an expansion. A model without a purpose clearly defined, serves no purpose.

Implication: boardrooms need a “pawn-king” version of a financial model to build a shared mental model.

“Cone of possibilities” mental model

How to build a “pawn-king” financial model?

Imagine a single sheet financial model that uses ~25 rows and ~13 columns only. This model must be designed in a way that it creates user-engagement. Simply put, the model users must play with a model for a while to start forming some type of mental model. Adding buttons and graphs will usually do the trick. Key is to avoid temptation to add more complexity. When in doubt, it’s important to remember that complexity is the enemy of execution.

An over simplified financial model that ensures a board director will “play” with it can create the right kind of “tutorial.” To prepare such a model, management teams must understand the concept of Computational Kindness (read Algorithms to Live By if you haven’t). Here is what this idea means:

• Imagine asking your friend “what do you want for dinner”, usual answer “I’m easy, whatever you want”. This creates unnecessary delays
• A computationally kind approach is to give a few choices such as “hey, would you like burgers or desi for dinner?”

By limiting choices you are exercising computational kindness. A computationally kind “pawn-king” financial model can help board members co-create a cone of possibility. Such an exercise creates a shared mental model. Shared mental models can move boardrooms away from mis-aligned conviction and towards informed intuition. By aligning a model’s design with approval stage, the cone of possibility can be painted with high specificity.

The best part about a clearly painted cone of possibility is a healthy atmosphere in boardrooms. I have observed that unexpected outcomes make shareholders upset. When a wide range of outcomes are equally and clearly understood, board-management relations can thrive.

NetEquity shared mental model

NetEquity is a start-up that wants to make Internet for all a reality by 2028. Creating a shared mental model for NetEquity required creating three separate mental models:

1. Initial conditions or problem statement
2. Technical insight
3. Solution

The initial conditions model

• 5 million cell towers provide mobile Internet coverage to 90% of the human population, yet 50% are not online
• Only 10% cell towers are fiberized and almost all of them are in urban areas
• Increasing network capacity in rural areas through “fiberization” is not financially feasible for the telecom industry

The technical insight model

• 80+% of cell towers are connected to the electrical grid
• The electrical grid’s topology is very similar to a telecom network
• Following the electrical grid to deploy fiber is a viable strategy

The solution

• Fiberizing rural towers will require bringing predictable, non-telecom cash flows to finance fiber builds
• Unlocking capital will require bankable contracts which will require aligning incentives and establishing trust

The “pawn-king” spreadsheet

Note: this spreadsheet is now obsolete and does not represent NetEquity’s actual model which financed a fiber network like a true public utility.

• Fits easily on a 13.5 inch laptop screen
• Creates the necessary conditions for a shared mental model to emerge