,

Why Complex Problems are Complex and Hard To Solve

From an early age, I have never liked the observation that something is complex. It usually meant that person is just resigning themselves to never understanding the problem. I couldn’t stand this defeatist attitude and have spent most of my life trying to devise ways to tackle complex problems including the aptly-named “wicked problems.” Even though I may never find the solution to the P versus NP Problem, it has taught me a great deal about problem solving in general.

So, what do we mean when say a problem is complex? According to Dr. Melanie Mitchell, there are nine definitions for complex as used by complexity theorists. These definitions range from “complex as a matter of size” to “complex as a degree of hierarchy” to “complex as a measure of algorithmic information content” (pp. 96-111). I tend to think of complexity in terms of systems theory in which you have a number of discrete components with numerous feedback loops and many variables that are hidden within the system processes.

A good example of a complex system is the American economy. There are many discrete components in the forms of companies, consumers, banks, regulatory agencies, etc. all passing information to each other and reacting to that information. Attempts to model the American economy range from the simple macroeconomic diagrams in textbooks to detailed microeconomic equations that requires years of mathematical study to even understand. Yet these models, no matter how detailed, cannot fully describe and fully predict how the American economy operates.

If you accept my definition of complexity then you can see how the next concept describes why complex problems are hard to solve. We have difficulty in solving complex problems because our observation of the problem is hindered, we cannot fully understand the problem, our decision-making processes are flawed, or we cannot act appropriately in confronting the problem. If any of the difficulties I mentioned sound familiar it is because I am describing the four components of the “OODA Loop.”

The Observe-Orient-Decide-Act Loop (OODA) was created by Colonel John Boyd who was a fighter pilot and scholar in military strategy. This concept has been adopted both by the U.S. military and championed by such business experts as Tom Peters. As the diagram below demonstrates, a person, team, or an organization observes a situation along with other inputs. Based on the observations and several internal factors, the subject attempts to orient themselves or understand the unfolding situation. Based on that understanding, the subject then makes a decision and acts upon that decision. Throughout the OODA Loop, there are several feedback channels that make the entire process nonlinear.

Colonel Boyd explained that the use of the OODA Loop was to travel through the Loop faster than your opponent. You present confusing and ambiguous information to your opponent so that they have difficulty orienting themselves and thus are slower to decide and act. Essentially, you want to go through your own OODA Loop faster than your opponent does so that they start falling behind and then are paralyzed by their inability to analyze the situation. Time is the key factor in OODA Loops.

The OODA Loop is why I think complex problems are so difficult to solve. Consider the five components of the OODA Loop as it applies to your personal abilities or the abilities of your team/organization:

  • Observe: This is the beginning of the Loop and also feeds into another iteration of the Loop. If your observational abilities are hindered or you just cannot observe all parts of the unfolding situation then you are working with incomplete information. History is replete with examples where disasters occurred because of the lack of key information.
  • Orient: This is where you/the team/the organization takes in the new information and pairs it with your previous knowledge, cultural traditions, and other internal factors that influence how you process and analyze information. So, even if you are able to observe the entire unfolding situation, your internal abilities to process and analyze this information can prevent you from fully understanding what is happening.
  • Decide: This relates to your ability to generate hypothesis about the situation and possible responses. There is the common “paralysis by analysis” which hinders decision making because you are still trying to orient yourself to the situation. Or, even if the organization has a good understanding of the situation, decision processes may be so cumbersome that you cannot make a decision in time to act on the situation.
  • Act: You may not have the resources to act promptly and/or appropriately. Your understanding of the situation may have led to a flawed decision that forces an invalid response to the situation. You do not have the proper feedback mechanism built in your action to determine how your act affected the unfolding situation.
  • Feedback: As you go through the OODA Loop, you are constantly generating and receiving feedback from your current iteration and previous iterations. Without good feedback design, your own actions can contribute to the ambiguity of the situation. This is especially true of wicked problems where there is no consensus on the actual shape of the problem and your actions can drastically morph the problem into a completely new problem.

The good news here is that you can also use the OODA Loop to better your abilities to handle complex problems. Use the five components as a checklist for improving your (or your organization’s) processes in handling complex problems.

For example: how well do you observe? How good is your organization at collecting and disseminating information internally? Do your people have the necessary prior knowledge and analysis skills to properly orient themselves when new observations come in? How robust and quick is your team’s decision-making skills? What barriers can you remove so that you can act faster? What can you do to improve your feedback mechanisms?

Government is going to face more complex problems especially in a climate of reduced budgets and increasing responsibilities. All government employees at all levels need to sharpen their problem-solving skills so that we are more innovative and can better tackle the looming wicked problems that face the nation. Whether you accept my suggestion to use the OODA Loop or come up with your own problem solving method, the process of thinking about complex problems is a great way to sharpen your problem solving skills.

Reference:
Mitchell, M. (2009). Complexity: A guided tour. New York: Oxford University Press.

Leave a Comment

16 Comments

Leave a Reply

Jay Johnson

Nice post. It’s interesting to compare Observe – Orient – Decide – Act to Plan – Do – Check – Act that is used a lot in process improvement circles.

Andrew Krzmarzick

Can the first word be “Yearn” as in “earnestly seeking data that makes sense”? I really want this acronym to be YODA. 🙂

Seriously: how do you bake this into culture so that it becomes second nature…like an athlete that practices until muscle memory takes over?

John Bordeaux

I would take slight exception to the complexity definition – the interaction among system agents drives individual agent behavior, observable phenomena, and in some circumstances, the system itself. This leads to emergent properties, visible only when the system is in ‘motion.’ (You can’t autopsy a bird to learn flocking behavior.) This is important because the only true model of such systems is the system itself, as system behavior is predictable only in short cycles – and impossible with large timelines.

With this view, the gentleman who offers Plan-Do-Check-Act as a comparison is in error, in my view. In fact, this is a significant and useful contrast. (Apologies if this is what was meant.) Linear predictive systems offer themselves to PDCA, and process improvement. Complex systems (and complex adaptive systems) do not, the first verb in OODA is essential. Observe rather than Plan. Understand the system’s nature and current state, rather than impose a framework of expectations based on past experience with similar systems. Let the data determine the framework.

@Andrew, this is how fighter pilots are trained, I believe. May be useful to start there.

Also, there’s no such thing as muscle memory. But I’ve gone on long enough. 🙂

Avatar photo Bill Brantley

@Andy – Colonel Boyd often referred to Jedi Knights so it would have been appropriate to try for a YODA acronym.

@John – Good point on emergence. I can’t believe I forgot about it. I also agree with your point about the difference between PDCA and OODA. Colonel Boyd emphasized the non-linear nature of the OODA Loop because it was designed to understand nonlinear situations.

I do have to disagree a bit on muscle memory. It is more that the mind lays down a specific circuit for an action after a long period of practice. You might be interested in Incognito: The Secret Lives of the Brain by David Eagleman.

Avatar photo Bill Brantley

@Andy – As to your question on changing the culture, you may want to read Chet Richard’s Certain to Win: The Strategy of John Boyd, Applied to Business. He writes about the culture necessary to sustain OODA.

There is a lot of stuff by practitioners on Boyd’s concepts but not that much in the academic literature. Given the nature of the theory, you will find a great deal on how to use OODA militarily and for business competition but what attracts me to it is that recent findings in neuroscience can be applied to OODA. There is also some fascinating tie-ins with behavioral game theory to OODA.

John Bordeaux

@Bill – just kindled Incognito, look forward to learning more. The specific circuit is more accurate than muscle memory. Nothing is automatic, not even a golf swing…but the experienced brain establishes a path for SOLVING that swing over time. As usual, we agree. The notion of muscle memory in popular usage infers this is an automatic function. This is the perception I wanted to dispel.

David A. Streat

Excellent post Bill! I agree with you that in order to solve the complex problems that we face today, leaders must change their way of addressing the problem. I would also submit to you that complexity leadeship theory (CLT) may be a way to change how we think about problem solving and decision making. Today problems cannot be solved by the “main boss” making a decision and everyone else following his or her guidance until the answer comes. Today’s leaders have to be aware of their emotional intelligence, how complex adaptive systems (CAS) operate, and how emergent behavior comes out of the networking of groups and organizations.

Kitsy M. Young

Thanks to Col Boyd for giving us the OODA loop. I work in Customer Engagement and our decision-making processes is flawed due to complex hierarchy and ambiguous information that prevents us from obtaining accurate or timely information and content. The OODA will assist me in describing the problem and hopefully cut through the stove-pipe groups withholding valuable content, bring in some light to the problem so we can focus on the solution.

Steve Richardson

Good stuff, Bill! Not only is our economy a complex system, but so is our government. Complex systems feature changing relationships among semi-autonomous agents. Unfortunately, many policies are developed without sufficient understanding of this. We know people are not machines but somehow think organizations can be treated as mechanisms. As you noted, Boyd’s decision cycle work highlighted the strategic import of time. Another consideration that isn’t appreciated by scholars and decision makers is conflict. Anyone working in government knows we seldom have the luxury of unanimous agreement on goals, methods, or priorities.

Avatar photo Bill Brantley

@Steve: Thank you! I do agree that conflict is also another important aspect of dealing with complexity. I have been working with combining OODA with behavioral game theory and this will introduce conflict into the OODA structure.

@Robert: Thank you! Super-complex systems sound like the concept of wicked problems. Are they the same?