Systems Thinking is an approach for developing models to promote our understanding of events, patterns of behavior resulting in the events, and even more importantly, the underlying structure responsible for the patterns of behavior. Verifying a model through simulation of the stock and flow model enables one to better understand the implications of the influences and study the patterns of behavior. If we are interested in addressing a particular situation it is only through our understanding of the underlying structure that we will be able to identify the most appropriate leverage points to effect change within the system. Systemic Intervention is conceived as:
- A well defined approach to understanding a situation through systems thinking and simulation,
- Developing and verifying a model of the structure responsible for the situation,
- Developing and verifying an alternative structure which is highly likely to produce the desired situation, and
- Migrating the current structure to the alternative structure.
Is Systemic Intervention Appropriate?
While a Systemic Intervention employing Systems Thinking and System Dynamics is an approach that can provide a very rational view of the situation, as well as the identification of approaches that are highly likely to produce the desired result, it is an approach that requires a substantial investment of effort. The following are some of the signs that indicate a Systemic Intervention is most likely warranted.
- There are multiple perspectives on just what the situation is, and how to deal with it
- Multiple stakeholders with multiple objectives all of whom need to be included in the solution
- Multiple variables that need to be optimized in favor of the whole
- Things seem to oscillate endlessly.
- A previously applied fix seems to overshoot the goal
- A previously applied fix has created problems elsewhere
- Over time there is a tendency to settle for less
- After a fix is applied, in time the problem returns
- The same fix is used repeatedly
- There is a tendency to allow an established standard to slip
- Growth slows over time
- Partners for growth become adversaries
- Limitations experienced are believed to result from insufficient capacity
- There is more than one limit to growth
- Limited resources are shared by others
- Growth leads to decline elsewhere
These are the most often experienced situations indicating a Systemic Intervention is appropriate. These indicators actually represent the behaviors of an interrelated set of system structures referred to as Systems Archetypes.
There is also a need to consider the cost and timeframe. The investment should be commensurate with the potential benefit. There are times where all that is possible will be a causal loop diagram, or a stock and flow diagram, however these could be very useful.
Also, the system to be modeled should be relatively stable during the period modeled. It doesn't seem it would make much sense to model a system which has already substantially changed before you even get it modeled.
A systemic intervention is unwarranted whenever the situation contains no feedback. That is, when the action and the result have no effect on the actors, which is actually very seldom the case.
Systemic Intervention Structure
A systemic intervention is essentially a balancing loop structure as depicted below.
Within this structure the difference between the Current State and desired state result in action intended to move the Current State toward the desired state. Action factor is simply a factor indicating the fraction of the difference between the desired state and the Current State which results in action. See Fig. 2 for a sample run of this structure. Click here to run this model in your web browser.
The question becomes, how do we go about understanding the Current State in a manner that will allow us to define the appropriate action that will actually move the Current State to the desired state.
How we label things influences how we think about things, and how we think about things influences how we label things. As such, there are a few conventions that will contribute to the development of the underlying structure.
- Define the Current State of the situation you're concerned about, not the system. Define the Current State as objectively as possible, without any reference as to what might be the cause or solution of the situation. Include only those elements necessary and sufficient to understand the situation. Attempts to describe "the system" often become overly complicated and bogged down in details that are not relevant to figuring out how to deal with the situation. The effort tends to get lost in process.
- Describe how the Current State has changed over the historical period that's relevant to the situation. That is, if you're describing something that changes quite readily then extensive history might not be necessary. If you're describing something that changes very slowly then an extensive history might be necessary to identify the pattern that has developed over time.
- Define the desired state as a description of what you'd like the future to be, not simply an elimination of the Current State. Simply making a problem go away produces a net result of zero, which isn't a very substantial return on effort is it?
From a systemic perspective the Current State is actually a result of the underlying structure and has arrived at this Current State with some identifiable behavior over time, a behavior which is indicative of the underlying structure. The question then becomes one of how to develop a model of the structure that's responsible for the Current State.
Current State Underlying Structure
Before attempting to deal with a situation it is appropriate to develop an understanding of why the system is behaving the way it is. The underlying structure provides a view of the interactions between the elements of the system which are responsible for producing the patterns of behavior resulting in the current state. Change Management: The Columbo Approach depicts the difficulty associated with the dynamic equilibrium in existing systems.
With a systemic perspective of things I know that I'm not simply looking for a cause of the situation. We're beyond "A" caused "B". The Causal Loop Diagram page contends that there are two basic structures, and more complex structures are composed of some number of these basic structures. So, what we look for are feedback structures among the influences. (In an attempt to keep this particular page under control full examples using the complete systemic intervention methodology will be selectable attachments).
If there are multiple perspectives regarding the structure responsible for producing the pattern of behavior each one must be documented as it is often found that, to some extent, they may all be relevant. Be sure to include only the relevant relationships so the model doesn't become overly complicated. This is generally referred to as identifying the model boundary. Also note that once the underlying structure is developed it may become apparent that additional data should be collected to complete the historical pattern of behavior.
The following pages address the details associated with developing a model of the underlying structure.
Verify Underlying Structure
The problem with systems that contain more than a couple elements is that the behavior of resources which accumulate or decline over time is often beyond our ability to intuit. Systems Thinking Causal Loop Diagrams represent a qualitative view of the system. As far as I can tell the only way to ensure that the structure developed adequately describes the situation is to develop a quantitative view of the system and simulate it.
We've employed Systems Thinking Causal Loop Diagrams to develop the structure, now it is appropriate to employ the rigor of System Dynamics to verify the structure. The intent of simulation is to verify that the structure developed actually produces the behavior over time trends that were produced from the historical data. If it does then there is value in the structure developed. If the simulation doesn't produce the historical patterns of behavior it is most likely that there is a problem with the assumptions embedded in the structure or the structure itself.
The short guide for developing a model is:
- Stocks - identify those entities which represent an amount of something which accumulates or declines over time.
- Flows - identify the inflows and outflows which are responsible for the changes in the stocks
- Parameters & Links - identify the information that is involved in influencing the flows.
For a more detailed description see Create a Model.
Leverage Points are those points within a system where small changes can effect a substantial change in the behavior of the system. At times the leverage points may be obvious, though at times they only become apparent though Sensitivity Analysis. Sensitivity analysis being a process where specific changes are made to certain influences within the model, with all other components held constant, to determine the impact on other elements of the structure. It is this study of the system that promotes the development of our understanding which is so critical to the following steps.
If the desired pattern of behavior is to be different than it has been in the past then the structure of the system must be altered such that it will produce the desired pattern of behavior in the future. It is most appropriate to look for the leverage points within the structure where a small amount of effort will produce a substantial change towards the desired result.
If the current underlying structure happens to be one of the well defined Systems Archetypes then the leverage points are already identified and there are well defined strategies for dealing with each structure. This can save you the effort of developing a strategy from scratch.
Once the alternative structure is developed it must be simulated to determine if it will in fact produce the desired behavior over time. The simulation also provides an identification of all the touch points that should be monitored over time to ensure the system is on track to produce the desired results.
The next step is to develop a plan for the transition from the current structure to the new structure. This is essentially a project management task, though my preference is to perform the transition in the area that will provide the greatest benefit most quickly. This will tend to encourage and reassure those involved that the transition is going to produce the desired result. Note that the alternate structure actually identifies the appropriate points to monitor during the transition to ensure that the transition is on plan.
Systemic Intervention utilizing Systems Thinking, Modeling and Simulation is a powerful approach for understanding why situations are the way they are, and developing an approach for improvement. Systemic Intervention is not an easy approach as it requires a substantial investment of thought and effort, though the results are generally found to be worth more than the investment.
Systems Thinking World Discussions
Systems Thinking World Q&A * Gene Bellinger