home | strategy development | strategy skills | site index | access keys

Strategy Skills > Structuring the Thinking

Systems thinking

> in practice

A key component of thinking strategically is recognising that issues do not exist in isolation. Holding a mechanistic view of policies as levers that have a focused and direct impact on a situation, without considering the wider implications of an intervention, can be short sighted and potentially disastrous. Strategic thinking requires the inter-related nature of circumstances to be recognised up front rather than relying on a post hoc screening to identify unintended consequences and impacts.

What Is Systems Thinking?

Systems thinking is both a mindset and particular set of tools for identifying and mapping the inter-related nature and complexity of real world situations. It encourages explicit recognition of causes and effects, drivers and impacts, and in so doing helps anticipate the effect a policy intervention is likely to have on variables or issues of interest. Furthermore, the processes of applying systems thinking to a situation is a way of bringing to light the different assumptions held by stakeholders or team members about the way the world works.

When Is It Useful?

Systems thinking is particularly powerful for understanding dynamic complexity, which stems from the relationships between factors in a system. A dynamically complex system cannot simply be broken down into pieces in the same way as a structurally complex system, which derives its complexity simply from the sheer number of factors involved. Where structural complexity can be modelled and managed using databases and spreadsheets, dynamic complexity needs a more organic approach to understand the complex web of influences that often results in various forms of feedback loops. Such loops add a time dimension to system complexity and often magnify or dampen the intended effect of an action in a non-obvious manner.

Influence Diagrams

The core tool in systems thinking is the influence diagram, which captures graphically how each factor or variable in a system influences the others. Arrows are used to indicate the direction of the influence together with a '+' or '-' sign to show whether an increase in the one variable leads to an increase or decrease in the other. A double line across an arrow indicates a delay before the influence is felt.

In the diagram above, an increase in training leads to an immediate increase in costs, but - via a delayed increase in morale which in turn reduces staff turnover and hence recruitment - a delayed reduction in costs. An additional complication is provided by the feedback loop driven by the relationship between recruitment levels and the need to train new staff.

The diagrams help to improve understanding of the drivers of behaviour in the system, and can uncover counter-intuitive effects of interventions. They can show how a change in one factor may have an impact elsewhere or feed back to affect itself, and also how two seemingly independent factors are actually linked.

Influence diagrams are best constructed in a working session with a small number of key people. The sessions are likely to stimulate in depth discussion as each participant's assumptions and views are explored and incorporated into the emerging picture.

Driver Trees

An influence diagram aims to map the relationship between all the variables in a system. However, it is likely that there are one or two key variables of particular strategic interest that need to be either maximised or minimised. Unravelling the influence diagram into a driver tree can be a powerful way of highlighting and communicating the drivers of these key variables, and hence provide insight into the kind of interventions that are needed to impact them.

Unravelling the influence diagram above can help to highlight the drivers of cost. The feedback loops in the system mean that certain variables appear in more than one branch of the tree. Where variables are repeated in this way they are conventionally placed in brackets.

Driver trees raise a number of questions, not least the relative significance of the different branches of the tree in driving the key variable.

Impact Trees

There will be only a limited number of variables within a system that can be directly influenced to act as levers for change. An alternative way to unravel the influence diagram is to highlight the impact that managing these variables will have on the rest of the system. Again using the example influence diagram above, an impact tree can be constructed to more explicitly highlight the consequences of increasing the level of training as described above.

Impact trees provide a causal sequence for understanding how managing one variable is expected to have an impact on another variable of interest. Social Researchers encourage a similarly explicit articulation of how an intervention is expected to have its impact using Theories of Change methodology, outlined in the Magenta Book.

Interpreting Feedback Loops

Constructing an influence diagram will highlight the great number of feedback loops that exist within any complex system. Interpreting these loops is central to understanding the likely behaviour of the system.

Reinforcing Loops 

A dominant reinforcing loop is a self-sustaining process that will lead to either exponential growth or decay. The critical factor is whether the process is proceeding in the desired direction, as once started the process will continue unchecked unless an intervention is made to break the cycle. The rise and decline of neighbourhoods demonstrates the potentially beneficial or destructive power of reinforcing feedback loops.

Balancing Loops

A balancing loop perpetuates the status quo. As one factor changes, other factors exert a balancing influence to return it to original level. This behaviour can either act as barrier to change or a beneficial stabilising mechanism. To drive change any intervention must be influential enough to over-ride the balancing effects.

Balancing Loop with a Delay

A delay in the influence of a balancing effect can produce oscillatory behaviour through repeated over compensation. As the balancing forces act to maintain the status quo, the lack of responsiveness in the system means that corrective action is excessive and the mark is over shot. Aggressive or heavy-handed management of such a system will produce instability. If the system can not be made more responsive the only option is to take change more slowly.

 Reinforcing Loop with Delayed Balance

A reinforcing loop with a delayed balancing influence will demonstrate 's-curve' style growth. The reinforcing loop produces a period of accelerating growth or expansion, which then slows and eventually comes to a halt under the delayed influence of the balancing effect. A classic learning curve follows this pattern.

Sustained growth can not achieved by simply encouraging the reinforcing process, but must be unlocked by removing or weakening the balancing influence that is creating the limitation to sustained growth.

Using Systems Thinking

• Work in groups: developing an influence diagram as a group exercise forces everyone to explicitly list the factors that matter in the system and then decide on the relationships between them.
• Use the influence diagram and tree to identify areas of study at the very beginning of the work and intermittently thereafter for further direction.
• An influence diagram can include both quantitative and qualitative factors and relationships.
• The tree and influence diagram can be used to inform the construction of quantitative models using software such as Vensim (free for personal use), Ithink or Powersim, which can be used to simulate system behaviour. (Note that the model's usefulness will be limited by the difficulty of meaningfully defining a mathematical algorithm for each influence or relationship).
• This approach is best used for designing and testing interventions, rather than designing systems.

Strengths

• Systems thinking can generate new insights into the drivers of a dynamically complex issue.
• The systems approach provides a powerful way for project teams to establish a shared agenda for addressing a problem. It allows development of consensus and ownership, leading to shared commitment to decision making.
• It ensures feedback loops are recognised and incorporated into policy design.
• The systems approach provides a powerful way for project teams to reach a shared understanding of how a system operates.

Weaknesses

• It is very easy to overcomplicate the system map and lose the key insights. It is important to focus on the key feedback loops and cut out the less important links.
• The process can be significantly undermined by team members who:
• dislike the approach and are out to prove it does not work
• are committed to a prior solution or who are fixated on finding "a solution"
• have hidden agendas that they are unwilling to disclose.

References

"Systems Failure" by Jake Chapman (Demos)

Checkland, P "Systems Thinking, Systems Practice", Wiley, 1981

Checkland, P and Scholes,J, "Soft Systems Methodology in Action", Wiley 1990 which provides a thorough update of the methodology together with several extended examples.

"Practical Soft Systems Analysis" by D.Patching, FT Prentice Hall 1990 provides a simple step by step introduction

"The emergent properties of SSM in Use: A symposium by reflective practitioners" by P.Checkland et al, Systemic Practice and Action Research, 13(6) p.799 2000 contains personal accounts of experience in the use of SSM in a wide range of contexts.

The Mind Tools website provides an introduction to system thinking and the behaviour of feedback loops.

Rich Pictures are another creative way of representing systems.

"Systems Thinking: a practice guide" by Business Dynamics, IBM Business Consulting Services (trevor.cooper@uk.ibm.com)

Business Dynamics: Systems Thinking and Modelling for a Complex World . By John Sterman

Systems Thinking, is The Fifth Discipline Fieldbook: Strategies and Tools for Building a Learning Organisation.

^ top