Complexity in Organizations.

Most managers never discuss the future. They are too much focused on reducing the complexity of managing today. The implicit vision of the future they create in their mind tends to be set and unchanging. Human beings are subject to bias and imperfect reasoning about uncertainty. They tend to misperceive events that are quite unlikely and to ignore or stress other possible unpleasant outcomes.
Today complexity has proved to be a primary cause of failure, for both private and public organizations. Understanding complexity is not a matter of reducing or ignoring, it. It is, rather, important to deal with management complexity and unpredictability, and to foster a learning-oriented strategic Planning & Control.
This does not imply the need to draw up detailed and long term plans.
It, rather, requires an ability to identify and understand:

• strategic goals and objectives;
• how to link short and long term objectives;
• performance drivers, (e.g. delivery delay, production flexibility, service reliability, quality/price ratio);
• tangible and intangible strategic assets (e.g. production capacity, personnel, financial resources, customer base, knowledge, product portfolio) affecting performance drivers;
• the current and desired level of such assets, and interdependencies between them;
• how to affect the dynamics of strategic assets through policy levers;
• how and when to measure achieved results;
• feedback loops between causes and effects, delays and non-linearities;
• the role and possible impact of significant external factors, and the relevant system’s boundaries.

Often spreadsheets and mathematical simulation tools based on an optimization approach are used to forecast the financial implications of plans. Quite often, such analyses are based on simplistic and misleading hypotheses that can lead decision makers to dangerous conclusions. In fact, they do not make explicit interdependencies between relevant variables, delays, non-linearities and policy levers.

Incremental and bureaucratic views are also very common to planning in Public Administrations, where goals and objectives are often the result of muddling through, rather than of a communication and deep learning process. Very often, in such contexts political goals and managerial objectives are not aligned. Planned activities are confused with expected results, performance indicators are too much linked to the accomplishment of activities rather than to the outcomes they are going to generate.

This perspective is often the cause of contrasts between the political and managerial level, between managers operating in different responsibility areas, and in different levels inside a same department. A defensive or even obstructive behavior is a very common reaction to the use of planning, goal setting and performance evaluation.

Our approach to planning the future.

Our approach can substantially help decision makers to better sketch strategic plans in a learning-oriented perspective, to face dynamic complexity.
System Dynamics models allows decision makers to link strategy to action, to better perceive interdependencies between business units and functions, between the firm and its relevant environment, to understand the crucial role of strategic resources on company performance and lifelong existence.

The System Dynamics Methodology

According to System Dynamics principles, dynamic models are based on a feedback view of business systems, seen as a closed boundary, i.e. embodying all main relevant variables related to the problem being investigated. In a dynamic model key-resources (whose monitoring on a strategic perspective over time is crucial) are represented as level variables and their inflows and outflows are shown as rate variables.

Levels are pieces of information concerning system conditions at a given time and they are a result of an accumulation process triggered by rates. Levels cannot be directly affected, as they represent resources currently available. They can be modified only through rate variables.

Rate variables describe how and why current decisions are made by the system’s key-actors and allow one to figure out system dynamics, according to a given set of selected policies. Rates include four main components:

1. the goal of decision maker(s);
2. the observed condition, mainly in accordance with the related level variable and of eventual delays in information gathering;
3. a discrepancy between goal and observed condition;
4. the desired action based on the discrepancy, in order to achieve the goal.

Moreover, input variables represent external constraints or even policy levers on which key-actors may operate, in order to affect – through rates – levels.

Consequently, a dynamic simulation model is based on explicit statements of policies underlying the decision making process, according to conditions (information on levels, time delays and external input constraints) arising within the system. In accordance with the systems feedback view, decision making is seen as a continuous process of converting information into signals which feed actions oriented to change levels, i.e. to affect key-strategic resources to be monitored over time.

Such a conversion process is not always clear and explicit in organizations; it may also often involve many decision makers. Making decision processes more explicit through dynamic modelling and improving them over time may substantially help people to better understand a same objective reality concerning day-to-day problems’ structure, which they usually perceive differently according to their several mental models. Such a learning process leads to improve mental models and helps to achieve a common shared view of reality.

The quality of policy making depends on information taken into consideration by key-actors and on the way they convert it into action, according to a given set of explicit or implicit “rules”. Concerning this, it has been observed that decision making problems are not usually related to a lack of information, but to a proper selection of it, which depends on the quality of mental models, i.e. on the way people frame systems in which they currently operate.

According to this perspective, decision making may also be seen as a process of filtering different sets of internal and external pieces of information, related to level and input variables. Such organizational and cognitive filters can be represented as concentric circles: the more they are positioned at the core of decision function, the more they affect it. It follows that, of the many pieces of information directed to a given decision making unit, only a few of them reach its core.

A proper consideration of such filters is crucial in order to build a dynamic model which could better depict concrete patterns of reality.
It is possible to distinguish two main kinds of feedback loops:

1. positive (or reinforcing) loops;
2. negative (or balancing) loops.

A positive loop describes a virtuous circle or a vicious circle concerning a growth or a declining process. It is characterized by a sequence of interactive relationships between several parameters, all varying in a same direction.

For example, a virtuous circle can be found in advertising, which could be an engine of a system growth, as it increases sales and consequently scale economies, which allow more advertising which increases sales once again, etc. Another example could be found in a vicious circle concerning poor quality, which decreases sales, which in turn reduces incentives to workers, leading to a decreasing morale and to an ever decreasing product quality, etc.

Positive loops dominance in any socio-economical system (and consequently also within a firm) is not endless; both virtuous and vicious circles may be counterbalanced on a longer perspective by negative loops, which are a source of the system’s stability.For example, the first positive loop commented on above could be counterbalanced by a market saturation.
A system feedback view through dynamic modelling may help an entrepreneur and other key-actors in the company to understand causes of system behaviour and to figure out different possible “futures” of the system itself, according to a given set of policies adopted.

More particularly, they can support decision makers in a timely perception of rising negative feedback loops which make growth slower, giving rise to a resource waste for the company. Such an awareness of system dynamics may allow the company to better explore the opportunity to pursue new policies, levering on new input variables in order to reinforce the same or to grow other positive loops.

Likewise, dynamic models may also help decision makers to understand how to neutralize vicious circles, i.e. which policy levers to act on in order to promote negative loops that could lead to stability.