The influence of Systems Thinking on Agile practices

If one should do a word cloud out of the Agile literature one word that would be big and bold is 'pragmatic'. It is the essence of Agility, which is a response to prescriptive methodologies elaborated far from the reality of the field. However pragmatism, extracting intelligent practices from experience, has a limit: these practices are contextual. For this reason some Agile practitioners and thought leaders would like to see one or more theories underlying Agile practices. Not that there is a need to justify Agile, countless success stories are sufficient, but because with theories comes understanding and the ability to apply them differently in different environments. 

Two related bodies of knowledge, System thinking and Complexity sciences are the main candidates for providing theories behind the success of Agility. In this blog I will give an overview of Systems thinking as taught by Russel Ackoff, A.K.A. the Dean of Systems Thinking, and its applications at the level of organisations. There are other schools of Systems Thinking but I have chosen to present the work of Russel Ackoff because firstly he was the first to apply Systems Thinking to Management,  and secondly the more recent schools of Systems Thinking tend to have a 'new age' aspect certainly very valuable at the individual level but not necessarily relevant when it comes to Management.

Russel_Lincoln_Ackoff

Since Descartes the most common approach to understand systems is to break them down in smaller parts easier to apprehend. This is the analytic approach. While this approach has proven itself valuable, it has an inherent limitation: it ignores the properties of the whole that the parts do not have. It is also argued that the analytic approach produces knowledge (how?) not understanding (why?). Systems thinking is the opposite but nonetheless complementary approach, a holistic approach. A system is a view of the world, composed of at least two interacting parts, which has properties the parts do not have. Different classifications of systems exist, Russell Ackoff proposes the following one based on choice:

  • if parts and system as a whole do not display choices it is a mechanical system, an example is a car.
  • If some parts do display choices but not the whole then it is an ecological system, an example is Nature.

  • If parts do not display choices while the whole does, it is an animate (relating to animal life) system, an example is Humans. These systems are also often called organic systems but it is a slight misuse of the word as plants are organisms that do not display choice.

  • if both parts and the whole do display choices it is a social system, an example is an organisation.

For long organisations have been studied and run like mechanical systems (taylorism) for profit generation and then like organic systems aiming for growth, profit becoming a mean. This model still prevails. In both cases parts or subsystems are mechanical, they do not display choice. This is not deliberate but a consequence of the analytic mindset and its underlying assumption that parts of a whole are independent. It is not even a bad idea, humanism put aside, in an environment where members of an organisation are expendable and where there are more opportunities for growth and profit than the organisation can handle.

However it is not the environment of organisations nowadays, these models are obsolete and system thinkers see this mismatch between model and reality has the source of many organisation’s issues. Typically mechanical parts perfectly fit together. Just like Sales and IT ? Not really.

Indeed when both the parts of a system and the system itself can display choices, the complexity of the system raises drastically. This is when systems thinking really comes into play. If analytic thinking is the reason for many system failures, then analytic thinking should not be used to address these failures. As Einstein stated: "We can't solve problems by using the same kind of thinking we used when we created them.” The answer is systems thinking, which is a synthetic rather than analytic process and explained as follow by Russel Ackoff:

"1) Identify the whole of which the system to be explained is a part 

2) Explain the behavior or properties of the containing system 

3) Dis-aggregate the containing system so as to identify the role or function within it of the system to be explained."

Stage 3 implies the study of the interactions of the system under study with the other sub-systems. This is a key concept of systems thinking. The interactions between the parts of a system matter much more than the individual action of its parts because its emergent properties derive from these interactions. Emergent properties are properties of the system but not of its parts. For instance, consciousness is an emergent property of the brain but a neuron on its own is not conscious. In other words, a system is not the sum of its parts but the product of their interactions. Systems thinkers do not solve problems with solutions likely to create other problems but dissolve problems by redesigning interactions and systems. "Design is to systems thinking what Research is to analytic thinking" says Russel Ackoff.

Based on this approach, Russel Ackoff developed a new kind of planning called Interactive planning, which is  based on the idea that the future of organisations depends as much on the actions of organisations in their environment than on the effects of environment upon organisations. Interactive planning has two phases: idealization and realization. Idealization consists of identifying two elements, the ‘mess’ and an ‘idealized design’. Systems need to adapt to their environment to survive, therefore systems contain the seed of their own destruction. Should they freeze in a changing environment, they would die. Formulating the "mess" is about finding this seed. This seed is likely to be not only one problem but a set of problems. Problems rarely appear in isolation. Once the mess is formulated, systems thinkers know what not to do, what they want to avoid. However, effective management is directed at what one wants, not avoiding what one does not want, thus the need for an "idealized design". It consists of imagining that the systems you want to improve has been destroyed and that you can redesign it with only three constraints:

1) it has to be technologically feasible,

2) operationally viable,

3) able to continuously learn and adapt.

The realization phase is about filling the gap, getting closer and closer to the idealized design. Interactive planning is continuous and iterative. Working backward from where you want to be to where you are, reduces the number of variables to be considered, it is the the best way to ‘beat the system’. (e.g. if Wimbledon starts with 65 players, how many games will be played?...think about it...If you took the problem from the front you are still searching, if you took the problem backward:1 winner, thus 64 losers, thus 64 games).

To endeavor that kind of transformation in organisations requires a new kind of leaders. Leaders aware that organisations are social systems. Leaders who take into consideration not only the objectives of their organisation but also those of its employees and environment. Leaders who are creative because structural issues are not dissolved with a quick fix. Leaders who have sufficient courage and influence to redesign organisations. Leaders who manage the interactions between subordinate teams or employees and not their actions. Leaders who encourage and facilitate an increase in competences at the level of both employees and organisation in a win-win situation.

 

This is obviously not a demonstration that Systems thinking underlies all Agile practices. Nonetheless we can recognize in Systems thinking many Agile key concepts such as iteration, continuous planning, importance of people, non-controlling leader and importance of interactions, to name a few. Hence, it is not pushing inductive reasoning too far to assume that what Russel Ackoff showed at the level of an organisation is also true at the level of a team. Also I am not sure if approaching problems backward is directly related to systems thinking although encouraged by Russel Ackoff but it seems somehow related to TDD

Now if you wish to appreciate the power of theories and see systems thinking applied to other fields than management, for instance the work of Marvin Minsky on cognition is fascinating.