Introduction

This Expert Lens utilizes complexity science to help better understand how to intervene with systems in a structured, yet non-linear way.

ToCs describe change processes within social systems. Social systems often are complex: they are dynamic, constantly changing. Have many interdependencies such as interacting actors and a myriad of influencing contextual factors. And they comprise of non-linear processes that result in emergent changes.

Complexity science provides methods and concepts that help navigate the complex landscape and thereby help social organisations achieve system change. Several of these we’ll discuss as part of this expert lens:

• Building on emergence and self-organisation
• Learning and adapting as we go
• ‘Loose’ Theory of Change, including ToC as a Compass and Development Evaluation
• Complexity aware monitoring

Systems thinking

Systems thinking emphasizes the connections between different parts of the system and the notion of a system as a holistic whole.

“A system is an entity with interrelated and interdependent parts; it is defined by its boundaries and it is more than the sum of its parts (subsystem). Systems often have multiple levels and actors. They have the capacity to change, to adapt when it is necessary in response to internal or external stimulus. Change in one part of the system may cause the whole system to change.”Based on https://en.wikipedia.org/wiki/Systems_theory.

Complex systems

Complex systems can be considered a more specific form of systems thinking. Complex systems are non-linear, entangled, wandering messes that do not lend themselves easily to traditional analysis and action.

“How a complex system reacts to changes in one part is not predictable but often shows itself in rather unexpected ways. It cannot be understood nor planned in a linear manner.”H. Baser & P. Morgan (2008), Capacity, Change and Performance Study Report. ECDPM Discussion Paper 59B.

Traditional Planning Fails

We have seen an increasing influence of complexity thinking on social change strategies. This came up as a result of the growing notion that the linear positivist approachesA positivist scientific approach adheres to the view that only “factual” knowledge gained through observation (the senses), including measurement, is trustworthy. In positivism studies the role of the researcher is limited to data collection and interpretation through objective approach and the research findings are usually observable and quantifiable. Source: http://research-methodology.net/research-philosophy/positivism/. to the planning of interventions do not represent well the systems of change in highly complex contexts and systems.See for example:
– B. Ramalingam & H. Jones with T. Reba & J. Young (2008) ‘Exploring the science of Complexity Ideas and implications for development and humanitarian efforts’, ODI: Working Paper 285.
– www.thebrokeronline.eu Interconnectedness and interdependence cause greater unpredictability and uncertainty. In the extreme, standard operating procedures, best practices and grand designs become irrelevant, counterproductive or downright damaging. Instead, complexity theory:

• provides a set of lenses with which to look at a complex world,
• helps pose questions to better understand the dynamics of real world systems in complex contexts and,
• helps generate insights as to how these dynamics can be ‘sensed’ and ‘navigated’.

Intervention Principles

The following principles can help to effectively intervene in complex environments:Adopted from B. Ramalingam, M. Laric & J. Primrose (2014) ‘From best practice to best fit: Understanding and navigating wicked problems in international development’, ODI: Working Paper.

• accommodating multiple alternative perspectives rather than specific best practices
• looking for multiple intervention points and moving among them dynamically
• working through group interaction and iteration rather than ‘back office’ designs
• generating ownership of problem formulation and transparency through participation of stakeholders, especially front-line staff and end users
• developing coherent visual representations of the problem that enable systematic and group-based exploration of solutions
• concentrating on flexibility rather than predictability of solutions.

Key Lenses with which to Look at a Complex World

When developing ToCs it is important to consider three sets of complexity science concepts. These are relevant to complex change processes and their ToCs as they shape the way we think about how change happens, help us to understand how to eventually negotiate complex realities and change trajectories. These three sets are: complexity and systems, change and agency.This section is based on the following paper, which provides a more detailed discussion and examples: B. Ramalingam & H. Jones with T. Reba & J. Young (2008) ‘Exploring the science of Complexity Ideas and implications for development and humanitarian efforts’, ODI: Working Paper 285.

Complexity and systems

1. Systems are characterized by interconnected and interdependent elements and dimensions.
2. Feedback loops promote and inhibit change within systems; they crucially shape how change happens within systems. It is the manner in which changes influence (feed-back) upon other processes in the change of a system either positive or negative.
3. Emergence exists when the overall system is showing collective behaviour not shown by the individual adaptive agents.Y. Bar-Yam (2011), New England Complex Systems Institute: http://necsi.edu/guide/concepts/emergence.html. The whole is different from the sum of the parts. Although the collective behaviour results from interaction of its parts, the complex interdependence often makes it unpredictable. Simple rules of interactions can generate complex characteristics and behaviours.

Complexity and change: the phenomena through which complexity manifests itself

1. Relationships between elements are frequently nonlinear. That is, change is frequently disproportionate and unpredictable in comparison to what initially triggered it to change. Interconnectedness and interdependent can cause initial small changes to have great effects (the butterfly that laps its wings leading to a Tsunami is an example for this). Thresholds may apply, where change only comes about once a contributing factor crosses a certain threshold level, triggering a substantial response. Similarly, accumulation may apply, where key aspects accumulate and deplete over long periods of time. Therefore, changes made many years ago may still affect outputs and outcomes, and changes now may take years before they translate in outputs and outcomes. Clear causal relationships often cannot be traced because of multiple influences. In a ToC process these multiple influences and elements therefore need to be understood through analysis, drawing insights from earlier experiences and by challenging linearity in the underlying assumptions, etc.
2. Sensitivity to initial conditions highlights how small difference in the initial state of the system can lead to massive differences later (for example, butterfly effects and bifurcations are two ways in which systems can change drastically over time).
   Perfect planning implies perfect knowledge, but in an uncertain world the map of the future cannot be drawn. For this reason planning should abandon prescriptive, goal-oriented decision-making and prediction about future states. Instead it needs to focus on understanding the dynamics of change and promote reflective monitoring, adjustment and a collective learning approach in constant dialogue with stakeholders.
3. Phase space (the ‘space of the possible’) helps to build a picture of the dimensions of the system, and their potential for change over time. The phase space of a system is literally the set of all the possible states, or phases, that the system can occupy. It does not seek to establish known relationships between selected variables, but instead attempts to shed light on the overall shape of the system by looking at the patterns apparent when looking across all of the key dimensions.
4. ‘Attractors’ and the ‘edge of chaos’ describe the pattern of order underlying the seemingly random behaviour exhibited by certain complex systems.

Complexity and agency: the notion of adaptive agents and how behaviour is manifested in systems

1. A system changes because a part of it changes, causing a reaction by the entire system. This can be based on actions of so-called adaptive agents that react to the system and to each other. This leads to self-organisation and co-evolution. These phenomena imply that for complex systems, traditional stakeholder analysis may not help to gain an understanding of the behaviour of the overall system.
2. Co-evolution describes how adaptive agents are mutually interdependent in their evolution. They influence and adapt to each other as well as to the overall system.
3. Self-organisation is a form of emergence where macro-scale patterns of behaviour occur as a result of the interactions of adaptive agents who act according to their own goals and aims, and based on their limited information and perspective on the situation.

Building on Emergence and Self-Organisation

Even the most seemingly unmovable and unchangeable systems can change from a major disruption. A disruption sets in motion a process of self-organisation, which in the end will cause the system to find a new equilibrium. This new balance is the result of emergence.

An example was the manner in which a major systemic change such as the Arab Spring started with upheaval and disruption in the countries in Northern Africa and the Arab World. Seemingly unmovable and unchangeable political systems were in a change process that was forced by agents from within the system.

In social change processes we would like to promote emergence by creating conditions that favour this process. The practices involved in engaging with emergence are broadly related to three iterative phases in emergence, each of which will inform the ToC:P. Holman (2010), Engaging Emergence: Turning Upheaval into Opportunity. San Francisco: Berrett-Koehler Publishers, pp.18.

1. Phase 1: Preparing for a system change
2. Phase 2: Hosting the system in its change process
3. Phase 3: Engaging with the system in its change process

Phase 1: Preparing for a system change

Preparing for emergence requires:

• Accepting that we don’t know and understand everything, but that we should be very curious to understand as much as possible. In doing this we can use many participatory analysis methods such as rich picture, systems diagrams, mapping of resources, participatory learning and action (PLA), value chain mapping etc.
• Choosing possibility: to be open to and to sense the (new) opportunities for changes. Useful methods include Future Search, Appreciative Inquiry and Open Space Technology.
• Following where the (life) energy of the system is going, recognize it and trying to give it space. Where are the hopes, aspirations, and visions pointing? What drives or motivates the people, what change is needed? Methods that can be used are again Appreciative Inquiry, Outcome mapping, etc. Storytelling and active listening play an important role in all of these methods, as are dialogue techniques and conflict handling.

Preparing for Critical JuncturesGreen. D. (2016) ‘How change happens’ Oxford: Oxford University Press.

Crises and disruptive events often present themselves unexpectedly. ‘Critical junctures’, such as wars or political and economic crises, may especially present themselves in complex environments. Under these events a system that seemed unchangeable, suddenly becomes susceptible to change, and may make sudden, unforeseeable jumps. There may be years without change and the organisation may seem ineffective, but these critical junctures may offer a window of opportunity for change, if you are ready to rapidly capitalize on it once it presents itself. It is then that you may be able to create the momentum, turn it into a tipping point and implement the desired change.

Therefore, it is important to be prepared to take advantage of the “window of opportunity” for change these events present. In fact, some ToCs are completely built around positioning oneself to capitalize on these disruptive events once they occur. These ToCs attempt to keep alternatives alive and available, as well as build trust and connections among key individuals, so that when the opportunity presents itself, the conditions are in place to rapidly take collective action and create momentum.

The Theme for Opportunities & Threats can be used to present such potential future opportunities, and to monitor them.

Phase 2: Hosting the system in its change process

Hosting the change process is to create a welcoming environment in which people really feel that their contributions matter. It creates focus in the intentions of all involved: what is it that really matters to us, what would we like to maintain and what would we like to change? It also involves creating a space that is open to diversity; diversity of people, of opinions, of experiences. For this process to be as inclusive as possible we need to make sure that all those who ARE INJ.M. Roberts (2004: 5). Alliances, Coalitions and Partnerships: Building Collaborative Organisations. New Society Publishers. – those with Authority, Resources, Expertise, Information and Need – are present and are welcome to participate actively.

Phase 3: Engaging with the system in its change process

In the engaging process we use several steps: we inquire appreciatively, we reflect, we connect, we listen, we are open to what emerges and we will act /react accordingly.

Many meetings of organisations and stakeholders in the context of developing ToCs already strive to follow a development process that includes these three phases. All three phases are followed in an iterative process and happen either at the same time, in sequence or without sequence at all.

Learning and Adapting As We Go

To operate under a situation of complexity requires an interwoven approach of thought and action, learning and adapting as we go.D. Green (2016: 7). How change happens, Oxford: Oxford University Press. http://how-change-happens.com/ The purpose of initial study is to enable us to place our bets intelligently. Crucial decisions come after that, as we act, observe the results, and adjust according to what we learn.

An organisation likely will benefit from multiple strategies, rather than a single linear approach. Failure, iteration and adaptation should be viewed as expected and necessary, rather than a regrettable lapse.

ToCs in Relationship to Systems and Complexity Thinking

ToCs are partly a response to the recognition that other, more linear, approaches such as the Logical Framework, insufficiently account for the dynamics (linkages, diversity of actors, interdependencies, multiple relevant levels of analysis, etc.) within a change process that result from their systemic and complex nature. ToCs offer a means of reflecting about and clarifying how we think the system will change; what factors may trigger the change, and which persons or organisations can act as change agents.

In any change process we have an implicit ToC. For complex and system level changes especially, it is important to make the implicit explicit. This enables reflection and learning. In any case, expecting the unexpected to happen is inherent to working in complex systems level change processes.

ToCs can help to understand how change in complex systems emerges. They help understand what feedback loops exist and help find ways to promote emergence toward a certain direction. Various tools and approaches can help with this.

System loops diagrams

For example, methods like scenario planning and system loops diagrams can help to further our understanding of feedback loops.

Image by J. Hsueh, Fisheries systems map, Webinar on systems change by the <a href=”http://www.academyforchange.org/” target=”_blank” rel=”noopener”>Academy for Systemic Change</a>, September 30th 2013-9-30.

Cynefin Framework and Four Quadrant model

The Cynefin Framework can enable us gain a more explicit understanding of the level of complexity. The Cynefin Framework distinguishes between different levels of complexity, each of which requires different change strategies and different ToC development processes. That is, the ToC development process needs to account for the unpredictability of how actions affect the change process (e.g., using complexity-aware monitoring). And the extent to which assumptions can be based on previous experiences is more limited, since small differences may result in very different cause-effect relationships.

Simple, in which the relationship between cause and effect is obvious to all, the approach is to Sense – Categorize – Respond and we can apply best practice.

Complicated, in which the relationship between cause and effect requires analysis or some other form of investigation and/or the application of expert knowledge, the approach is to Sense – Analyze – Respond and we can apply good practice.

Complex, in which the relationship between cause and effect can only be perceived in retrospect, but not in advance, the approach is to Probe – Sense – Respond and we can sense emergent practice.

Chaotic, in which there is no relationship between cause and effect at systems level, the approach is to Act – Sense – Respond and we can discover novel practice.

The fifth domain is Disorder, which is the state of not knowing what type of causality exists, in which state people will revert to their own comfort zone in making a decision.

Other approaches that help understand complexity include theories about understanding institutions and institutional change. Institutions are mechanisms maintaining and promoting the stability of systems, and thus have an important bearing on change processes.

Systems change

Kania et al. (2018)J. Kania, M. Kramer & P. Senge (2018) The water of systems change. FSG: Reimagining social change. Available here. offer an actionable framework for systems change. Systems change entails a holistic perspective, shifting conditions that are holding a social or environmental problem in place while taking into account their interdependencies. These conditions exist at three levels, the explicit, semi-explicit and the implicit.

Changing these six interdependent conditions applies not only to the external system but also to the very same conditions for systems change within your own organisation, such as your own mental models about evaluation and investment in learning.

‘Loose’ Theory of Change & Theory of Change as a Compass

The “context” might be understood just enough to approach the development of a ToC as the development of a grand design. However, complex contexts may be ill understood, unpredictable and constantly changing. In these contexts, instead of a blueprint or “roadmap”, the ToC is more appropriate to serve as a “compass for helping us find our way through the fog of complex systems, discovering a path as we go along”.Green, D. (2015) ‘Where Have We Got to on Theories of Change? Passing Fad or Paradigm Shift?’ From Poverty to Power Blog, 16 April. “This doesn’t mean ditching planning processes altogether, but recognising that often, plans reflect best guesses about the future (and about the past too) and will likely shift over time”.Valters, C. (2015) ‘Theories of Change: time for a radical approach to learning in development’ Overseas Development Institute. The ToC then becomes a “loose” ToC.

“Loose” Theory of Change

There is always a risk of trying to develop a perfect ToC from the onset. This makes understanding and usage of the ToC sometimes complex. Instead, to deal with emergent issues and unforeseen risks that may come about, a loose ToC constantly goes through adaptation.

A “loose” ToC thus is broadly translated as moving beyond static theories and looking at the unpredictability or “looseness” of things happening. By “loose” it means that the expected outcomes are stated from the onset of the initiative but the outputs that are necessary for the outcomes may change. In this regard, the ToC attempts at being responsive and adaptive to certain new insights and unforeseen circumstances. The expected outcomes that are stated from the onset effectively function as a compass directing toward the change process you seek to activate. These expected outcomes or objectives will often be stated more broadly to allow for emergent outcomes. A consequence by design of a loose ToC is that there’s uncertainty (i.e. openness) about the organisation’s Theory of Action.

The “looseness” of the change process in a loose ToC approach can vary. One may state the full change process from the onset or only some of the outcomes that come more at the end of the change process. In the case of the latter, not only the Theory of Action but also the change process are loosely defined.

Focusing and Scoping

Related to loose ToCs is what Funnel and RogersS.C. Funnell & P.J. Rogers (2011) Purposeful program theory: Effective use of theories of change and logic models. San Francisco: Jossey-Bass. Pp. 163-176. refer to as focusing and scoping. Complex environments can increase fluctuations in an organisation’s sphere of control and influence. Therefore, the scope of a ToC needs to be revised frequently, and so do the priorities within the scope that the organisation focuses on. Hence, the boundaries of the scope and focus should be frequently reconsidered and become more fluid to accommodate emergent needs, opportunities, and outcomes.

Working with a “Loose” ToC Approach in Evaluation

In most instances evaluation requires testable hypotheses to be stated upfront specifying what will happen. A flexible ToC design therefore presents a major challenge for theory based evaluation and learning. How do you approach this when faced with such then?

Rick DaviesDavies, R. (2016) ‘Evaluating the impact of flexible development interventions’ Methods Lab. presents some guidelines. He acknowledges that data availability is particularly challenging for evaluation of a loose ToC. Substantially more data is needed because data is needed for all events that could be making some form of causal contribution, and not only for a sub-set of combinations defined by a specific prior hypothesis.

Developmental Evaluation

In circumstances of complexity and uncertainty, Development Evaluation might also be a good approach to evaluation. Developmental evaluation facilitates real-time, or close to real-time feedback thus facilitating continuous learning and adjustment. At the core of developmental evaluation is experimentation with rapid, real time interactions that generate learning, and to follow up to the emergent and dynamic reality with radical adaptation and innovation. Adaptation and innovation can take the form of new projects, programs, products, organisational changes, policy reforms, and system interventions.Patton, M. Q. (2010) ‘Developmental Evaluation. Applying Complexity Concepts to Enhance Innovation and Use’, New York: Guilford Press.

More information about this evaluation approach can be found at BetterEvaluation.

Further Reading

• Davies, R. (2016) ‘Evaluating the impact of flexible development interventions’ Methods Lab.
• Valters, C. (2015) ‘Theories of Change: time for a radical approach to learning in development’ Overseas Development Institute.

Complexity Aware Monitoring

Complexity aware monitoringThis section is adapted from USAID (2016) ‘Complexity-Aware Monitoring’, USAID Discussion Note. is a monitoring approach for complex contexts where cause and effect relationships are poorly understood. It builds upon three key principles to monitor the emergent and dynamic aspects of strategies and projects:

1. Synchronize monitoring with the pace of change
   For monitoring to be useful, its frequency needs to match the pace at which the context changes and the pace of project adaptation. In highly dynamic contexts you need to collect information more frequently and possibly even on an ongoing basis. In such fast-paced circumstances leading indicators can act as an early warning system. Under slowly evolving circumstances, on the other hand, monitoring is less frequently needed to be able to act on information on time. While lagging indicators may inform you too late to be able to act in fast-paced circumstances, in relatively static contexts they can offer more accurate and definitive information than leading indicators.
2. Attend to three blind spots
   Given their primary focus, monitoring systems often have tree blind spots. First, they often focus primarily on intended outcomes. This makes monitoring systems blind to a broader range of outcomes associated with the intervention or system (intended, unintended, positive or negative). Second, they focus on intervention(s) as the dominant causal factor. This makes them blind to alternative causes from other actors and factors. And third, they focus on clear and pre-defined causal pathways, making monitoring systems blind to the full range of non-linear pathways of contribution. Monitoring systems in complex environments, however, can’t afford these blind spots and have to track a fuller range of outcomes, causal factors, and pathways of contribution.
3. Attend to relationships, perspectives, and boundaries
   A monitoring system in a complex context needs to track the structures, processes, and exchanges linking actors and factors within a system. It needs to provide information on the different perspectives within a system. And it needs to provide information that is useful for the consideration of what is in and what is outside the system.

Below are five methodologies suggested for complexity aware monitoring. Each of which – some more than others – follows the three key principles.

Sentinel Indicators

A sentinel indicator is like the thermometer of a system. It’s a proxy indicator for the overall “health” of the whole system and captures the essence of the change process. It thus takes a holistic perspective; change anywhere in the system will affect the sentinel indicator. It is easily communicated but provides incomplete information, which therefore needs to be followed up with further observation or probes to learn about the underlying details.

Stakeholder Feedback

This refers to participatory methods, such as citizen report cards, community scorecards, client surveys or other forms of collecting opinions. These can source the diversity of stakeholder perspectives.

Process Monitoring of Impacts

Process Monitoring of Impacts looks at how outputs are used by targeted beneficiaries or partners to produce the first level of results (outcomes).It can also be used at other levels of the ToC but it seems most valuable for the level between outputs and outcomes. It focuses on relationships between outputs and outcomes by describing the processes by which partners or beneficiaries (are expected to) use the outputs. Do they use them as was expected and does this result in the outcomes we expected? It is important to also be attentive to emergent processes. One advantage of Process Monitoring of Impacts is that it tracks the occurrence of impact-producing processes long before changes would be apparent in the corresponding outcome indicator.

Most Significant Change

This is a participatory monitoring and evaluation technique that involves the collection and analysis of stories describing the most important project outcomes. Instead of measuring indicators, the method collects and analyzes qualitative data on broadly defined “domains of change”. Story collectors ask open questions about the most significant change that took place for participants, and why they consider it significant. Phases that are part of this technique include: story verification, groups selecting the most representative stories while making selection criteria explicit, as well as meta-analysis of the selected stories and of the applied selection criteria.

Outcome Harvesting

Outcome harvesting
For more on outcome harvesting:

– R. Wilson-Grau (BetterEvaluation), Outcome Harvesting.

– R Paz-Ybarnegaray & B. Douthwaite (2016) ‘Outcome Evidencing: A Method for Enabling and Evaluating Program Intervention in Complex Systems’, American Journal of Evaluation. puts more emphasis than Most Significant Change on verification and on identifying and describing contribution. It is a participatory monitoring and evaluation method. Primary intended users play an important role in determining what to monitor. Key informants include change agents, primary intended users, as well as knowledgeable independent sources.

As point of departure it takes a change that has taken place in an outcome. Each outcome description explains how a specific change agent contributed to changes in the behaviour of particular individuals, groups, organisations or institutions. It further describes what changed in these social actors’ actions, relationships, policies or practices. After collecting evidence of outcomes (and with an open mind), it seeks to find out the causes (inputs, activities, outputs and other outcomes) of that change.

Success Case Method

The purpose of Success Case Method is not to examine the average performance – rather, by identifying and examining the extreme cases, it asks: ‘When the program works, how well does it work? What is working, and what is not?’ It is useful for documenting stories of impact and for understanding the factors that help or hinder impact. It is particularly useful for uncovering the contextual forces that influence impact. Also, it facilitates adaptive management in response to contextual changes or new understandings.

An example of an application can be found on the blog below by Liz McGuinness. First, participants were interviewed to identify success cases. The second step was to interview the participants and their supervisors in order to learn more about how they were successful, and which conditions enabled or inhibited this success. The same methodology could be used to identify unsuccessful cases in order to learn why and under what conditions, they were not successful. Brinkerhoff who developed this approach delineates five steps in the Success Case Method.

* BetterEvaluation blog ‘Lessons from a trial of the Success Case Method’.

* BetterEvaluation article on ‘Success Case Method’.

Further Reading

• USAID (2016) ‘Complexity-Aware Monitoring’, USAID Discussion Note.

More about Complexity Science for ToC Thinking

Various literature sources discuss concepts from complexity science that offer food for thought as you seek to understand and create change in social systems. BetterEvaluation offers links to several of such papers. Eguren applies several of these concepts to offer advice to facilitators of workshops about ToC thinking.

• I.R. Eguren, Process Oriented Theory of Change Facilitation: Surfing the Waves of Complexity.
• BetterEvaluation on Complexity.
• TED Talk by Eric Berlow: ‘How complexity leads to simplicity’.