This is a high-level overview to the domain of complexity management, which we will be going over in more detail in following articles. We can understand complexity management as the application of complexity theory to the practice of management. Thus, it draws upon the key insights and ideas from complexity science and uses them to try and help us manage complex organizations. Examples of these complex organizations would include cities, international politics, multinational corporations, global logistics networks or healthcare systems. These are all complex organizations due to their nature of having many parts that are highly interconnected, interdependent and autonomous.
We can and do go on using our traditional industrial age management approach to try and management these organizations, but as we will be discussing in a future module, the basic principles underlying our traditional management approach were designed for dealing with relatively simple systems. That is, organizations that have a limited number of components, that interact in a simple linear fashion at a low level of interconnectivity and where we can constrain those components.
A factory would be a classical example of this. Thus, it should be not surprising that the industrial age management approach was designed to manage relatively simple closed systems (like factories) because that was exactly what we had to do a hundred or two hundred years ago, when this approach was formalized. But today, in post-industrial economies, managing factories, as important as it is, is really the least of our concerns. We face much more complex challenges such as trying to get our healthcare systems working, enabling effective forms of governance, or collaboration between companies across large supply chains.
These are all very different forms of organizations that present a different set of challenges and require a more complex approach to management that is aligned with their core attributes. And we already see lots of innovation in management theory to meet this. We have seen the rise of a new set of ideas around agile and lean organizations, many different ideas around flipping the hierarchy upside down and creating networked organizations, and lots of innovation in creating more collaborative organizations. But most of these new management theories can be best understood within the context of complexity theory, as they all fit within the complexity framework. This article then tries to outline what that framework is. That is to say, the key concepts within complexity theory and how they apply to management. Although complexity theory consists of a whole zoo of new ideas, we will structure it around a few central concepts namely, systems thinking, nonlinearity, networks and adaptation and evolution.
Firstly, complexity management is based on the systems thinking paradigm. Put simply, systems theory or systems thinking is a holistic way of looking at the world. This paradigm posits that the parts to a system can only be properly understood, and thus managed, when taken in relation to the whole system. Whereas our traditional analytic approach takes a system and breaks it down, focusing on the individual parts of the organization, systems thinking looks at the relationships between the parts and the context or environment within which something exists. Whereas our traditional analytical approach is focused on closed well-bounded organizations, systems theory is instead looking more at open systems. That is to say, organizations that have so much interaction and exchange with their environment that we can’t model them as being closed. And this is the nature of complex organizations. They are really open networks instead of closed well bounded formal organizations.
Think about a healthcare system. It is a complex network of many different organizations and individuals – from primary care practitioners to hospital administration, to government agencies to universities and so on. All of them are interdependent in affecting the outcome to that system and the boundaries are fuzzy or even nonexistent. For example, would we include the food companies in this organization, as what they produce clearly has a direct impact on people’s health? Would we include the clean air lobby group as air quality would also have an effect on people’s health? Another example would be global supply chains, where we have to get many independent organizations across the value network to collaborate for the whole supply chain to function effectively. This supply chain has no real boundaries to it and no one is in control of the whole organization. The same would be true for international politics and many more forms of complex organization. This is the nature of complex organizations. There is no real boundary to them and this makes them quite different in nature to our traditional well-bounded organizations. Therefore, because they have no real boundary, it is more relevant to talk about them as systems, or networks of connections.
So the question is then how do we actually manage this form of open organization where we don’t really have control over the members, but still, they are interdependent and we need to get a functioning global outcome from the system for the end user? The problem being that our traditional reductionist approach is predicated upon actually being able to directly itemize and control the members of the organization. Our traditional approach works by creating a boundary around the organization and a hierarchy within the organization, where top management then decides what is best for the entire organization and the members are constrained and coordinated towards that predefined outcome. In such a way, we get global coordination. Sufficed to say this does not work in open organizations. Global macro level patterns in complex systems are emergent phenomena of local level interactions that give rise to self-organization. Thus, instead of directly aligning the actions of the members towards the desired global outcome, we create the context or platform within which the members can interact so as to coordinate locally and, out of this, we will get the emergence of some global organization. For managers, this means creating the context that facilitates the process of self-organization to take place. We can’t directly control the outcome to the system, but we can influence the initial conditions. We do this by creating a conducive context that represents an attractor towards coordination and cooperation between members. The big idea here is that of collaboration. We no longer have control but we can enable the context and conditions for collaboration.
As we turn up the connectivity, the parts to the organization become more interdependent and this is something we are currently witnessing around the world with the rise of IT and globalization. Interdependence creates nonlinearity. When we put two or more things together and they become interdependent, they can work together constructively or, inversely, they can counteract each other. Creating then a combined organization that is more or less than the sum of its parts – 1 and 1 stops adding up to two and linear thinking starts to break down.
It should come as no surprise the fact that our traditional management methods are very much based on linear thinking. We look for linear cause and effect interactions to describe events. And, in reality, we will often go on using simple linear cause and effect descriptions of events when they don’t really work, simply so that we don’t have to deal with the complexity behind a situation. We go on talking about GDP when we all know this is a very blunt linear metric that hides the complex set of interacting variables to a society’s overall quality of life. We go on looking at maps of the world divided up into nation states when this simple linear model hides the much more complex set of networks that make up our global economy and society. Linear thinking works well in simple environments and it is a necessary competency to being an effective manager. But it is not sufficed. In complex environments, we need to be able to recognize that it can hinder us in dealing with the real dynamics of the situation. Thus, in such situations one has to be prepared to switch to a more complex nonlinear form of reasoning.
Complexity management would posit that most complex phenomena that we encounter are the product of a number of different variables interacting in a specific nonlinear fashion. They are amplifying or dampening each other to give us the overall emergent outcome. Because no one single thing causes the outcome, we can not simply solve the problem by affecting one input variable. We need to identify and effect the multiple relevant factors. So, for example, let’s apply some nonlinear thinking to our current challenge of global terrorism, which is clearly a very complex problem.
A linear cause and effect approach would be to simply exert superior military forces against the enemy, and that cause would effect the desired outcome that we wish, the eradication of terrorism. We might note that in this approach, we are looking at the problem as independent from us and other things. It simply exists out there and we can just go and solve it with a single cause.
Now, if we treated this phenomenon as complex and nonlinear, we would be looking less at the actual phenomenon itself and more at the network of interdependencies that are generating it, including a recognition of our own interdependence in this issue. That is to say, what role do our actions play in causing this problem. We would be also remembering that events in time feedback on themselves, current events are path-dependent, meaning they are typically conditioned by feedback from historical events. Using this nonlinear paradigm, we would have to look at how the nexus of education interacts with religion and global culture. You would have to look at the socio-political dynamics in the region and of course how economics and corruption interact with poverty, food prices and so on. All of which interact in a constructive or destructive fashion to produce this emergent outcome.
This should help to illustrate why it is called complexity management because, before anything, it is about accepting that not all things are simple. Some things are, but some things are truly complex, and trying to treat them as if they are not, does not change the situation. You really have to roll up your sleeves and dig into all the specific interacting parts but also understand the system as a whole. This is the only way you really solve difficult problems. There is no silver bullet. But on the other hand, if you do actually understand how things work and you are prepared to do all the hard work, then complexity theory does provide us with a framework that actually makes it possible to solve problems that otherwise are virtually impossible.
Network theory is another central part of complexity science, as it deals with the highly interconnected architecture of real-world complex systems, such as transportation networks, financial markets or ecosystems. With the rise of the Internet, we are seeing the birth of new forms of network organizations and the so-called access economy. Open platforms like Airbnb, The App Store, Uber and many others have shot to fame taking over or creating whole new industries. The networked platform model is proving highly scalable and they are currently disrupting many industries.
With the reduction in transaction cost that IT enables, these networked organizations are able to harness new value sources and access whole new markets that were previously not possible within the closed formal organization, whose cost of coordination was too high to reach out to the mass of people. Such networked platforms are instead able to harness the small, but when combined, vast productive capabilities of the so-called crowd, or long tail. The mass of people that were previously not productive enough to organize formally, are now able to set up their own networks of collaboration, and we are seeing a new mode of production within society, sometimes call peer-production or mass collaboration. Classical examples being Wikipedia or the Linux foundation. A much more dynamic swarm-like form of organization.
The access economy is a major paradigm shift from ownership of products to access to services. In the same way, hyperconnectivity is unlocking a vast amount of untapped productivity of the crowd on the long tail. It also has the potential to do the same to virtually every product around us. We can think of all the products around us as both things and the function they perform. The industrial economic model was all about ownership of things. That monolithic ownership worked to lock up the functionality of the product, meaning that the function was only accessed a fraction of the time. A good example being the typical car that is used on average less than approximately 5% of the time. With the reduction in transaction costs and the proliferation of networks, we can now think of products as services and economies not as being about the buying and selling of products but instead about access to value. The access economy requires us to change our management approach – to think in terms of access, connectivity and networks. Luckuly, network theory give us the language to do that.
Adaptation and Evolution
With fast-paced technology innovation, globalization and changes in our supporting ecosystem, the post-industrial world is no longer a stable predictable environment it once might have been. It is marked by what business leaders call VUCA, standing for volatility, uncertainty, complexity, and ambiguity.
Our traditional centralized organizations are normalized for stable and predictable environments and are inept at dealing with fast-paced change, or nonlinear radical events. But fast-paced change and radical events are becoming the new norm as the world becomes more interconnected and dynamic. And this is congruent with what complexity theory tells us about the capacity for nonlinear systems to generate black swan events and the butterfly effect. As we see more indicators becoming exponential, we also see a shift from normal distributions to long tail distributions – that generate many more events that are several standard deviations from the norm, such as financial crisis or extreme weather phenomena. We currently see a lot of creative destruction as companies in the S&P 500 are lasting shorter and shorter periods of time. The primary focus of organizations within this disruptive environment should be survival through building adaptive capacity to enable their evolution.
This VUCA world requires a recalibration of many dimensions to our approach to management – a fundamental shift from resisting change to adapting to it. This requires a new set of capabilities surrounding adaptive capacity, the capacity for the organization to evolve new solutions in response to their changing environment. Adaptation involves a recognition of uncertainty. That is, that we can not fully know future outcomes. This idea of uncertainty runs very much contrary to our traditional management approach. In fact, it is so alien that in economics we typically call it radical uncertainty. But radical uncertainty is really just normal old uncertainty. It just looks radical because our standard analytical approach leads us to think that the future should be knowable. We conceive of the future in terms of some linear transformation of the past, what is called ergodicity – we take a sample from the past and use it to compute future outcome probabilities. Most of our business analyses and, particularly our financial analyses, are dependent upon this idea of ergodicity, that the future is knowable.
But when things get more nonlinear and complex, ergodicity doesn’t really hold. As the system evolves, we get nonlinear interactions between events that have not even emerged yet. Try turning the clock back twenty or thirty years and predicting how the Internet would unfold; how we would get the emergence of smartphones; that would give rise to the app economy; that would enable big data and all the future possibilities that will emerge out of that. This is not really possible because of the nonlinear interactions and the emergent evolutionary dynamic to the system’s development. In such a case, our strategy needs to change from one of trying to predict the future to adapting to it as it unfolds. Adaptation means being open to uncertainty and maintaining a diversity of states in order to be able to respond to a variety of possible outcomes that are as yet unknown. It requires an evolutionary approach that involves experimenting, testing, and rapid iteration. In a word, it requires business agility. We are currently seeing how the idea of enterprise agility has gone from nowhere to being identified as one of the few top-level strategic enterprise capabilities. And we will be discussing agility and evolution in following articles as we look at the VUCA framework.