Warren Weaver, an American mathematician known for his work on machine translation and his support for science in the United States, defined the complexity of a system as “the degree of difficulty in predicting the properties of the system if the properties of the system’s parts are given.”
Weaver addressed and handled complexity by dividing it in two general types: disorganized and organized. In disorganized complexity, a system has a very large number of parts (millions or more). So, because of its nature, the interaction between these parts can be treated only using probabilistic and statistical methods to model the real system. By the contrary, in an organized complexity there are a limited number of parts. Correlations between these parts can be treated using specific structured methods like optimization algorithms to model or simulate the behavior of these correlations.
No matter the type of complexity you have to deal with, business organizations apply specific methods and tools to model business processes and solve specific day-to-day issues (making tactical decisions) or establishing long-term roadmaps (making strategic decisions).
Decision makers have to deal with the complexity inherent to problem solving and decision making in order to take advantage of business opportunities and improve business performance on a regular basis. Most decision-making processes involve managing three major issues regarding complexity: its level; its dynamic behavior; and managing big data volumes.
Level of Complexity
The Fact: Business processes are a not just a static set of steps to follow, rather they often change according to new constrains, or needs for improvement. As a consequence, almost every decision-making process has to deal with increasing levels of complexity derived from incremental modifications or functional adds to business models. Often, complexity increases because business processes tend to change, grow, and evolve, not the opposite.
The Act: Avoid using informal methods to discover and deal with the level of complexity of a problem. Establish formal methods for pricing, marketing, measuring competition, selecting a software solution, etc. Apply these methods to set all the necessary parameters, monitor all faces of the analysis, and finally, to measure all the outcomes against the goals. Consider applying alternative solutions in case of any unexpected situation.
Dynamic in Complexity
The Fact: In current business operations, the complexity involved in the decision-making process can change frequently, even during the analysis process. New factors or variables can arise, some other factors suffer modifications due to external and internal influences, and some factors may disappear.
The Act: A decision-making process has to meet flexibility criteria in order to manage changes during the evaluation process as new data is inputted. Here are some things you should consider: develop a plan for taking sequences of decisions and what to do in every possible scenario; and how these decisions are related and affect the others. This can enable the necessary flexibility to make adjustments along the way.
Data Volume Complexity
The Fact: When carrying out an evaluation or analysis, data is the precious raw material. But it can easily turn into part of the problem. High volumes of data carry also the risk of augmented complexity to the analysis process.
A decision support process has to be made based on a defined set of the most reliable data. Sometimes the decision process does not depend on data volumes but certainly on data that is accurate, consistent, and reliable enough to represent at least a sample of the truth. Consider a set of data that guarantees accuracy, timing, and consistency.
The Act: When making a decision support process, it’s necessary to consider having a formal process to conduct it. There are some new approaches that rely on strictly formal processes to address decision-making issues like “decision engineering”. This discipline adopts many other formal engineering practices and incorporates them into a single discipline approach for decision-making evaluations.
Finally, it is important to conclude that no matter the decision support system (DSS) used to conduct the analysis, it is necessary to take into account some important factors: the input information as a valuable resource to initiate, the plan to consider a series of intermediate decisions—actions to take, possible scenarios, the expertise and knowledge of people who may work with the process, and the results based on the goals that are intended to achieve. Always consider that at the end, decision making is essential in every business process.
I welcome your thoughts—please leave a comment below, and I’ll respond as soon as I can.
[…] Warren Weaver, Scientist and Mathematician. […]