Nonlinearity and complexity would be nice to ignore, but they impact learning experience development projects. This section gets to the meat of the challenges today. Though you may be tempted to skip this section, consider that complexity is a large part of why Lean-Agile works well in our modern courseware projects.

By analogy, if you needed to create eLearning or mobile learning but had no authoring tools besides Microsoft Word and PowerPoint, you would face many hurdles. Taking some time to investigate appropriate authoring tool options and selecting a good one can reduce your frustrations and help you reliably deliver training products on time. Similarly, taking some time to investigate how complexity impacts training projects today and exploring how you can to respond differently can mitigate risk and reduce frustration.

Before we look at the relevance of complex systems to learning and development, let’s clarify a few terms.

When discussing complexity it is helpful to also discuss systems and their boundaries. A system is a set of interacting or interdependent component parts forming a complex/intricate whole.[1] Every system is delineated by its spatial and temporal boundaries, surrounded and influenced by its environment, described by its structure and purpose and expressed in its functioning.[1] The term system may also refer to a set of rules that governs structure and/or behavior.[1] We may choose the system perspective to consider the courseware, the buyer’s team, the producer’s team or a particular technology component.

Our discussion of systems necessarily includes complex systems. The study of complex systems represents a new approach to science that investigates how relationships between parts give rise to the collective behaviors of a system and how the system interacts and forms relationships with its environment.[2] It is useful to represent such a [complex] system as a network where the nodes represent the components and the links their interactions.[3]

Modern scientific study of complex systems is relatively young in comparison to conventional fields of science with simple system assumptions, such as physics and chemistry.[3]

Complexity science says that we’re surrounded by systems and that these systems are complex and constantly adapting to their environment—called complex adaptive systems.[4]

Now that we have the terms clarified, let’s look at the relevance of complex systems to learning and development.

LX development teams and groups of buyer stakeholders are complex adaptive systems with elements, in this case people, aggregated into systems, in this case teams or groups, that demonstrate complex behavior and adapt to a changing environment around them.

For much of our history, we have viewed the world and universe around us as a linear environment where simple cause and effect rules apply. Mankind has believed for most of our history that everything could be predicted and controlled.

We have often used our preferred linear mindset, looking for causality—that connects one process (the cause) with another (the effect)[5]—in our collective efforts to better understand and predict the behaviors of the systems around us to aid our planning, problem solving, and decision making. We often rely on decomposing or breaking systems down into smaller components to understand the pieces apart from the whole. We call this approach systems analysis. This is closely related to requirements analysis.[6]

Slowly, we have begun to better understand complex systems and the phenomena associated with complex systems. As theories about complex systems have begun to emerge, people from many disciplines have discovered similar findings.

The search for simple solutions to complex problems is a consequence of the inability to effectively deal with complexity.
— The late Russell Ackoff

The challenge in gaining a more full understanding of system behavior by the traditional decomposition approach with complex adaptive systems is that the components interact differently.

Complex systems may have the following features:[7]

  • Possess a high amount of structure or information, often across multiple temporal and spatial scales.[8]

  • Complex systems may have nonlinear interactions.[7] Small perturbations may cause a large effect, a proportional effect, or even no effect at all.[7] Effects may be indirect rather than linear cause and effect.

  • Easy to see in hindsight, but very difficult to see in foresight.[9]

  • Complex systems have a large numbers of components, or autonomous processing nodes, [10] often called agents, that interact and adapt or learn.[11]

  • Dynamic systems of interacting agents.[12]

  • The agents change the state of their environment by their actions.[10]

  • Complex systems may have a high degree of adaptive capacity, giving them resilience in the face of perturbation.[4] The agents as well as the system are adaptive.[11]

  • Complex systems may have cascading failures.[7]

  • Due to the strong coupling between components in complex systems, a failure in one or more components can lead to cascading failures.[7]

  • Complex systems may be open systems.[7] It may be difficult or impossible to define system boundaries.[11]

  • Complex systems have a history [11] or a memory.[7] They evolve and their past is co-responsible for their present behavior.[11]

  • Complex systems may be nested.[7] The agents of a complex system may themselves be complex systems.[7]

  • Complex systems may have networks of agents with many local interactions.[7] Interactions are primarily with immediate neighbors and the nature of the influence is modulated.[11]

  • Complex systems may produce emergent behaviors.[7]

  • In complex systems, relationships contain feedback loops.[7] Both negative (damping) and positive (amplifying) feedback are always found in complex systems.[7] This is known as recurrency.[11]

We recognize more nonlinearities in today’s world, including in learning and development. Unexpected events, volatility in requirements, technical complexity, social complexity, organizational culture and interdependence between parts of a courseware project have all added uncertainty to projects, and they tend to cost more and take longer. Bottlenecks that aren’t noticed for too long impact costs and schedule.

Overemphasis on efficiency can make our teams and processes more fragile to random stresses that occur unpredictably. When applying Lean, be careful not to reduce organizational structure to the point where it cannot absorb unexpected shocks and survive. Buffers, reserves, and redundancy help us avoid fragile structures. Our projects tend to get as weak as the weakest link in the chain. The Theory of Constraints buffers the primary constraint.

Parent organizations and buyer stakeholders still desire predictions, but often predictability is elusive in complex projects.

Complexity theory predicts that we cannot rely on predictions.[13]
— Jurgen Appelo

Lean-Agile provides an approach, rather like an exploratory model, that breaks the project size down into smaller pieces, experimenting, and gaining feedback and adjusting earlier in the project lifecycle, which reduces uncertainty each iteration, increasing our chances of a successful project. Feedback changes the system. Visible Kanban boards allow you to see bottlenecks as soon as they appear, reducing delays, which in systems dynamics tend to cause overshoots in our responses if we’re acting on outdated information. Daily synchronization meetings provide information exchange and collaboration instead of standard status reporting meetings. A Lean-Agile team’s size caps out at 7-9 people ideally, reducing the impact of size effects. Smaller experiments by more teams produce smaller harm when they fail than do huge experiments that can become too big to fail.

Feedback is an important part of Lean-Agile used in teams that are complex adaptive systems. Each team member observes the daily flow of information and context on the Kanban board and responds appropriately. Their individual responsive actions consequently feed back and influence the next set of contexts for the team. Damping or balancing feedback tends to help settle on the desired target change, while amplifying or runaway feedback tends to reinforce the trend, sometimes leading to nonlinear snowballing and exponential changes. Delays in feedback can cause actions that cycle between overshoots and undershoots to the target change. Daily standups and Kanban boards making information visible to the entire team help mitigate delays in feedback.

Attempts to decompose a complex system into its component agents and observe the one agent’s behavior in an effort to predict the system’s behavior are obstructed by the tendency for emergent system behaviors. Emergence is a process whereby larger entities, patterns, and regularities arise through interactions among smaller or simpler entities that themselves do not exhibit such properties.[14] The emergent property itself may be either very predictable or unpredictable and unprecedented, and represent a new level of the system’s evolution.[14]

For LX development teams, small issues may accumulate and may cause large effects or even blowup the project/program. Or the accumulation may have no effect at all. Each person in a development team may not have all information about the project, although Lean-Agile mitigates this somewhat. People communication and collaboration leads to unexpected results.

Even as learning and development professionals, we strongly encourage gaining a deeper understanding of these complex adaptive systems because it will illuminate some of why you have experienced what you have experienced. It also can help you to see why Lean-Agile principles can help address our challenges going forward.

We also recommend Nicolas Taleb’s books about the effects of uncertainty, volatility and randomness. His titles are Black Swan, Antifragile and Fooled by Randomness. His books are dense with original thinking that can also apply to learning experience development. The experimental and adjustable nature of Agile fits the optionality that Taleb recommends.

We humans work hard to develop mental models or maps to navigate our world. Be careful not to confuse the map for the actual terrain or the model for the reality. Our maps and models are incomplete simplifications to help us find our way. Thinking about complex systems depends on context.

Multiple weak models can make just as much sense as one strong model (and its certainly better than no models). In the end, all models fail.[15]
— Jurgen Appelo

So does our adaptation of Agile principles address complexity? Lean-Agile:

  • Addresses complexity with complexity

  • Uses multiple models, rather than over dependence on only one model like Scrum or ADDIE alone

  • Depends on context

  • Allows for subjectivity by system participants and the impact of their actions on the system

The principles of warfare that we mentioned at the beginning of this book make no mention of "best practices". David Snowden’s Cynefin Framework identifies best practices as only appropriate for his "obvious" domain which would include simple courseware projects. Emergent practice is more appropriate for warfare (complex and chaotic domains) and for complex courseware projects. Lean-Agile supports emergent practice. Adapt a principle to your situation rather than copy someone else’s method(s) without changing them.

Your organization may still crave predictability. You will have to work within that challenge.

Be careful applying any one expert’s opinion or mental model as a universal solution in our complex world. Apply correct principles into current methods. Adjust your methods as circumstances change. In the end, we point back to small scale experiments, which cause little harm when they fail, using PDCA to continually improve. The plan part of Agile is both backward-looking, last iteration, and forward-looking, imagining an ideal future state. Find what works for you in your context and apply more of what works and less of what doesn’t, just like the children’s game of red-light, green-light.

This also means that teams not accustomed to identifying and quickly and often applying lessons learned tend to need coaching to realize that the adjustments to the feedback are where most of the gains of Agile come from.

Our purpose for adding a section about complex systems in a book about learning experience development is to highlight the need to understand indirect effects when planning and executing courseware projects/programs. Start your problem solving using the Cynefin framework to help you realize where your project fits.

Applying the Cynefin Framework for Training Projects/Programs
  • If cause and effect is obvious to all then you’re in the obvious domain where best practices can work well as you sense, categorize, and respond.[16]

  • If the relationship between cause and effect requires analysis or investigation and/or expert knowledge then you’re in the complicated domain where good practice can work as you sense, analyze and respond.[16]

  • If the relationship between cause and effect can only be perceived in retrospect, but not in advance, then you’re in the complex domain where emergent practice will see you through as you probe, sense and respond.[16]

  • If there is no relationship between cause and effect at the systems level, then you’re in the chaotic domain and only novel practice will see you through as you act, sense and respond.[16]

Solving problems within complex systems can be challenging and difficult. Assuming linear cause and effect is only appropriate in the obvious and complicated domains. As you influence your organizational culture towards adapting to complexity, Lean-Agile will help.


1. Reproduced from Wikipedia article https://en.wikipedia.org/wiki/System under a Creative Commons Attribution-ShareAlike 3.0 license.
2. Reproduced from https://en.wikipedia.org/wiki/Complex_systems under a Creative Commons Attribution-ShareAlike 3.0 license.
3. Reproduced from https://en.wikipedia.org/wiki/Complex_system under a Creative Commons Attribution-ShareAlike 3.0 license.
4. Reproduced from Wikipedia article "Complex adaptive system," Wikipedia, The Free Encyclopedia, https://en.wikipedia.org/wiki/Complex_adaptive_system (accessed January 14, 2016) under a Creative Commons Attribution-ShareAlike 3.0 license
5. Reproduced from Wikipedia article https://en.wikipedia.org/wiki/Causality under a Creative Commons Attribution-ShareAlike 3.0 license.
6. Reproduced from Wikipedia article https://en.wikipedia.org/wiki/Systems_analysis under a Creative Commons Attribution-ShareAlike 3.0 license.
7. Reproduced from Wikipedia article https://en.wikipedia.org/wiki/Complex_system under a Creative Commons Attribution-ShareAlike 3.0 license.
8. Reproduced from Scholarpedia article http://www.scholarpedia.org/article/Complexity under a Creative Commons Attribution-ShareAlike 3.0 license.
10. Reproduced from Wikipedia article https://en.wikipedia.org/wiki/Distributed_artificial_intelligence, under a Creative Commons Attribution-ShareAlike 3.0 license.
11. Reproduced from Wikipedia article https://en.wikipedia.org/wiki/Complex_adaptive_system, under a Creative Commons Attribution-ShareAlike 3.0 license.
12. Reproduced from Wikipedia article https://en.wikipedia.org/wiki/Computational_economics under a Creative Commons Attribution-ShareAlike 3.0 license.
13. Reproduced with permission from Jurgen Appelo from http://www.slideshare.net/jurgenappelo/complexity-thinking/67-Complexity_theory_predicts_that_we. See his book Management 3.0: Leading Agile Developers, Developing Agile Leaders (Addison-Wesley Signature Series), 2011, by Jurgen Appelo for more about how Agile helps complexity.
14. Reproduced from Wikipedia article https://en.wikipedia.org/wiki/Emergence under a Creative Commons Attribution-ShareAlike 3.0 license.
15. Reproduced with permission from Jurgen Appelo from http://www.slideshare.net/jurgenappelo/complexity-thinking/94-A_key_point_of_complexity. See his book Management 3.0: Leading Agile Developers, Developing Agile Leaders (Addison-Wesley Signature Series), 2011, by Jurgen Appelo for more about how Agile helps complexity.
16. Reproduced from Wikipedia article https://en.wikipedia.org/wiki/Cynefin_Framework under a Creative Commons Attribution-ShareAlike 3.0 license.
Image

Line By Line

Here a Little, There a Little, Layer by Layer.

Back to Overview