A Definition of Complexity Theory

Complexity involves a systems approach to the study of the dynamics of environments and organisations (Bertalanffy 1972). Systems are comprised of a number of connected interdependent parts. The systems are adaptive because the behaviour of the individual interacting agents change in response to different events and to what happens during the course of their interactions. They are also non-linear because the same action or condition can have varying effects upon outcomes.

A very useful definition of the meaning of a complex system was provided by Peter Allen (2001: 150) when he said that a complex system was any system that had a capacity to respond to its environment in more than one way. This meant that it was not a mechanical system, with a single trajectory but had the potential to respond in different ways. One of the major differences between complexity and conventional approaches to strategy is the way it understands the key issue of change in environments and organisations (Arthur et al. 1997). This perspective envisages change as a process in which economic and industry environments move between equilibrium points through time in response to shocks caused by factors such as technological change, politics and change in consumer preferences (McMillan 2008). This is a similar perspective to the ‘theory of punctuated equilibrium’ (Gould and Niles 1972). Periodically, relative stability (‘equilibrium’) is punctuated by episodes of radical change after the process of incremental strategy formulation leads an organisation to a point whereby it is out of tune with its environment. This is a phenomenon known as ‘strategic drift’ (Johnson et al. 2011).

The way that complexity views how organisations change is different to the conventional viewpoint (McMillan 2008). Complexity views environments and organisations as dynamic complex adaptive systems which are constantly changing in both incremental and radical ways. They are dynamic because they are interconnected and constantly evolving together rather than in isolation. They are not conceptualised in equilibrium terms and they never stand still.

Complex systems have many interacting elements, agents and sub-groups and complexity analysis focuses upon their interactions and the interdependencies between them (Waldrop 1994: 145). However, the complexity metaphor for change is not punctuated equilibrium but evolution or coevolution. Just as the actions of a bee collecting nectar impact more widely on the ecosystem in which it is nested as it pollinates plants, so the actions of human agents and organisations impact upon their environments even if they are unaware of those impacts. This means that interactions between agents at the micro-level of the firm can lead to the emergence of new patterns of behaviour both within the firm and at the macro-level of the industry environment or the economy. For example, agents responsible for the micro-processor, software, the personal computer, the Internet and worldwide web and digitisation had a transformational influence in this respect. Similarly, interactions between agents in the wider industry and economic environment can lead to the emergence of new patterns of behaviour in those environments and within those firms. The complexity conception of change is, therefore, coevolutionary (Arthur et al. 1997). No element in the system evolves in isolation from the others. The success of the modern smart phone is the result ofthe coevolution ofa broad range offirms (agents) from a wide range of industries (or business ecosystems) including telecommunications, consumer electronics, computing and media to name just a few.

The dynamics driving change in complex systems are not the simple cause and effect dynamics of classical science but centre on the concept of feedback. This can be negative feedback which dampens activity or it can be positive feedback which encourages and amplifies activity. In complex adaptive systems feedback (particularly positive feedback) can develop its own momentum, rapidly multiplying its effects and escalating rapid development and transformation in ways which leave behind the conventional understanding of change. The network effects in relation to e-commerce platforms such as Amazon and Alibaba and social networks such as Facebook and WhatsApp plus the Google search engine, are all examples of this positive feedback.

A further point to consider with regards to feedback in complex adaptive systems is that the effect is often lagged. In other words, changes do not happen instantaneously but occur much later than the point at which feedback becomes available to the system. This further multiplies complexity and unpredictability in systems. An example of this is illustrated in the Gartner Hype Cycle (Fenn and Raskino 2008).

For example, when an innovation breakthrough occurs there is a time lag between initial interest (Technology Trigger - Phase 1) and when mainstream adoption starts to take off (Plateau of Productivity - Phase 5). Although the Internet-of-things has received significant amounts of publicity it has still to reach mainstream adoption. The time lag between Phase 1 and Phase 2 (if feedback is positive) can range anywhere from 2 to 10

Table 3.1 The five phases of the gartner hype cycle (Adapted from Fenn & Raskino: 2008)

1) Technology Trigger

A technology breakthrough occurs triggered by proof of concept. This stimulates media interest although there may not be proven commercial viability.

2) Peak of Inflated Expectations

Publicity leads to hype and a range of successes and failures.

3) Trough of Disillusionment

A shake out occurs as failures increase and interest starts to diminish. However, the survivors improve their products leading to phase four.

4) Slope of


There is a broader comprehension of the benefits of the innovation and second and third generation products emerge.

5) Plateau of Productivity

Mainstream adoption occurs in the form of successful commercialisation and monetisation.

years. If the feedback is negative, the innovation may take longer or may fail to progress in its original format (see Table 3.1).

< Prev   CONTENTS   Source   Next >