Guidelines

Here are a few words of wisdom developed after numerous years of enduring the pain from not following my own advice.

Relationship diagrams, or maps, are created from Elements which have Relationships among them, some of which create Loops, along with notes which describe the Elements, Relationships and Loops so others know what you meant or were thinking when you created them.

• Element: a non directional noun or noun phrase representing some entity in the world and is of one of three forms:
  • Stock: an accumulation which increases or decreases over time based on relationships
  • Constant: doesn't change during the period the model considers
  • Variable: computed from Stocks, Constants and other Variables

• Relationship: an arrow representing the manner in which an influenced Element changes based on an increase in the influencing Element. A Relationship has the following influences on the Elements:
  • Adds to/Same: indicated as blue arrow meaning Adds to for Stocks and Same direction for Variables.
  • Subtracts from/Opposite: indicated as a red arrow meaning Subtracts from for Stocks and Opposite direction for Variables.
  • Delay: indicated by || on the relationship meaning there's a delay before the influenced element realizes the influence.
  • Note: the S and O notation developed by Peter Senge in The Fifth Discipline ensures that at least one relationship in a loop is not valid as represented. I strongly recommend not using it.

• Loop: a set of relationships from one Element back to the same Element containing at least one Stock. The following types of Loops exist:
  • Reinforcing: promotes exponential growth, which may be virtuous or vicious and has zero or an even number of Subtracts from/Opposite relationships in the loop.
  • Balancing: promotes goal seeking, stabilization or stagnation and has an odd number of Subtracts from/Opposite relationships in the loop.
  • Sequence: label the loop in terms of what it promotes and use Bn (Balancing Loop #) and Rn (Reinforcing Loop #) in a single sequence so others understand in what order to interpret the loops.

• Notes: the rest of the world wasn't with you when you thought your way through the relationships so whatever you produce is seldom as intuitive for others as you might think. Provide notes to help them understand. Kumu enables you to add notes on Elements, Relationships and Loops, and I recommend the use of all of them. I have continued to find it quite surprising, after time, to understand some of my own models when I don't follow this advice.

The process for developing a relationship map is far more straightforward than you might initially think.

1. Situation: start with the definition of problem/situation that is to be understood, not the cause, explaining what you want to understand how to deal with. This will be your guide as to what's relevant and what's not. When one endeavors to map systems they typically end up with things that look like the Afghanistan Strategy Model and don't serve to promote understanding.
   • The first things on the diagram could be the Element for the problem or situation, what is to be influenced, and and Element for the outcome or goal. Dive in, don’t worry about it being perfect or correct - Remember, there are not “right answers” rather “usable models.”) It will sort itself out as entities are added and moved around. It's an iterative process.

2. Behavior Over Time: attempt to describe or sketch how the situation has developed over time as this will provide a sense of the time frame for your model. How long does the pattern you suspect is present take to play out?
3. Influences: beginning with the situation, ask what it most influences or what most influences it. Create the Element, and, if it is a stock, then set the type (in Kumu) to stock. Define what you think the relationship is and set its type (in Kumu) as Adds to/Same or Subtracts from/Opposite.
   • Eventually, it may be desirable to quantify the adds & subtracts but much can be accomplished with directional relationships.

4. Loop: when you identify a loop label and sequence it as described above.
5. Repeat: repeat steps 3 and 4 until all relevant relationships have been identified. This might requires multiple conversations with others who know something about this particular system.
6. Present: share the relationship model (including your thinking process) with others in the hope that they may be aware of relevant relationships you didn't think of. And when they do, be curious rather than defensive or protective of your precious model! The worst thing you can do is argue with them.

While the guidelines for developing relationship maps are relatively few and straightforward you will find they create a very powerful tool set for simplifying what appears at first to be very complicated interactions. No matter how complicated the map appears it's only some combination of reinforcing and balancing structures. The interactions are complex, not complicated. If you develop and label the relationship map well it should tell a story.

Many thanks to Jeannine Siviy, Steven Byers, Franco Pittoni for feedback on this piece.

19.07.12 Steven Byers comment after I asked him how he differentiated complex and complicated.

I think I first made this distinction when I studied Heifetz and adaptive leadership, and the distinction he and his colleagues make between technical problems and complex adaptive challenges. Problems that are mainly technical in nature can be very complicated, but someone knows the “answer”. You might need an expert, but the knowledge is available. In the realm of the complex, the answer is not known and likely the challenge is not well understood or even defined. In the realm of the complex, emergence happens.

I’m also influenced in my thinking by the Cynefin framework.

If there’s a lot of dots to connect, it’s complicated. If there’s a lot of dots and they are moving and merging and changing, it’s complex.

For me, all systems involving humans are complex. When I think of “complicated” my mind goes to mechanical systems.

In response I sent him a link to Knowledge Management: Bah Humbug paper and asked him about the Extending the Concept segment.