Complex Problems

I define a "simple problem" as a problem for which logical reason can be used to find a solution that eliminates or reduces all symptoms of the problem. For example, if Joe is running low on gas, Joe can go to closest gas station and fill up the tank. Dust off hands, problem solved.

But what if the problem is that Joe is habitually running low on gas? And when the problem is examined, we find that there are several contributing factors:
  • Joe has a gas guzzler car
  • His job doesn't pay very much
  • He is paid once a month and towards the end of the pay cycle, he starts running low on cash making it impossible to fill up
  • He doesn't live close to work so he has to drive a lot
  • He has a family and needs to take care of his children in the evenings so his wife can study for her college classes
  • He lives in a city without many employment options, so he's lucky to have the job that he does
Now a perfect solution is harder to come by. If Joe gets a new fuel efficient car, he might not be able to afford the payments. If Joe decides to walk or ride a bicycle, his commute time is going to go way up and it will impact his wife's ability to continue to go to college (which if completed will help Joe with his income situation). A higher paying job would fix a lot of things, but one may simply not be available in his city and moving may not be financially feasible. It seems like any potential solution that resolves some aspect of Joe's problem leaves others unaffected, or even worsened.

That's a complex problem: a problem for which an ideal solution that addresses all symptoms or aspects of the problem is not possible.

In reality, there are not many simple problems or, more accurately, simple problems don't last long in the world. People tend to fix them quickly since the solution is easily identified and implemented. As a result, most problems are of the complex and intractable variety, especially if the problem has been around for awhile. I scoff at people when they talk about how simple it would be to fix some complicated government problem like, well take your pick: high crime rates, the national debt, abortion, trade deficit, terrorism. "I'll tell you how we improve crime rates, we just need to throw these guys in jail!" Sure, right. Because that's worked so well in the past, and there's so much available room in prisons, and prisons are so cost-effective. High crime rates, like the other problems, is not a simple problem with a simple answer. It's one outcome of a series of interdependent factors, causes, and downstream effects. Solutions that resolve or adequately address all aspects of the "problem" most likely do not exist.

My day job is primarily concerned with addressing complex problems. The stakes of the problems I deal with aren't as high as with political problems, but the intractable nature of the problems is the same. And in their own way, the problems I work on have political angles as well, although much less toxic and polarizing than in government. I have a way I've developed of working on these problems. It hinges on the idea that a perfect solution that resolves everything does not exist, so you have to decide what you're going to fix and what you're going to let slide. In other words, it forces you to make hard choices, and more importantly, it forces stakeholders in the problem to make those hard choices with you. Another thing I like about this approach is that it is very rational and minimizes emotion from the process as much as possible. Here are the steps:
  1. Situation Analysis. First, you need to understand and describe the problem and its contributing factors and effects in total and in detail: A good situation analysis should describe the problem and its history. It should define all the factors that are contributing; the issues that have resulted; and the interactions and dependencies between the factors. You may be the person responsible for the situation analysis, but the analysis should be the result of conversations with all the major stakeholders. Everybody affected by the problem should be represented to make sure the analysis is comprehensive. This shouldn't be an overly contentious step since all you're doing is describing the various facets of the problem and everybody should be getting ample airtime.
  2. Prioritization of Objectives. Second, you must set a stake in the ground regarding the most crucial things that a solution must achieve. You also need to define what objectives are less important and acceptable to not resolve so long as you fix the crucial ones. This will be the basis of how to decide which objectives should prevail when they conflict with other objectives, which they always do in a complex problem. There are a number of ways to do this, but in all cases you want to articulate all reasonable objectives and provide a way of comparing their relative importance by either ranking them or categorizing them (e.g. must-do, should-do, would-be-nice-to-do, etc.). This could be the most vigorously debated part of the process. If you're presenting this to stakeholders, you want them to be engaged in the prioritization scheme and implementation because ultimately they must accept the final prioritization. Approval for the prioritization depends on the organization. It could be a single person (the CEO, for example), or it could be a handful of selected stakeholders, or it could be a simple majority of all stakeholders (rare). The one thing you want to avoid is requiring a large number of people to all agree. Somebody, or some limited number of somebodies, has to have authority to make a decision in the event that universal agreement on priorities is not possible. Requiring everybody to sign off is a recipe for gridlock. Regardless of how the prioritization is ratified, in the end all stakeholders must accept the prioritization decision and be accountable to making it successful, whether they agree with it or not. That's just the way teams have to work.
  3. Solution Alternatives. If you've done the first two steps well, this step is usually surprisingly easy. You propose and describe potential solutions to the problem. By this point in the process, you will probably have received a number of potential solutions already from your stakeholders. Presenting more than one is usually preferred for both pragmatic and political reasons. A "solution" may be a set of actions that individually address multiple aspects of the problem; or it may be a strategy or policy that can be applied to various decisions to be made in managing the problem day-to-day; or it may be a high-level project plan. In any case, each potential solution should be assessed by how well it achieves the prioritized objectives in step 2. In most cases, you'll have a "recommended solution" which is the one that best achieves the prioritized objectives. If you've truly gotten buy-in on the prioritization, then the presentation of potential solutions should lead inevitably to your recommended solution – in other words, the best solution will usually be self-evident once you have defined what you're trying to achieve. If somebody gets wrapped around the axle about how the recommended solution doesn't address one aspect or another, you take them back to the prioritized objectives and point out that their objective wasn't as high priority as the ones you optimized the solution for, and that the prioritization that you used is the agreed-upon model. In some cases that can lead to a re-prioritization, but in most cases it leads to that stakeholder accepting that a solves-everything solution is not possible and that this is as good as can be crafted.
  4. Risks and Mitigations. We've already established that with complex problems, every solution is flawed in some way. Therefore, for each potential solution you should list the ways in which things can go wrong. Because, you know, shit happens. You should have a description of the risk; it's likelihood of occurring; its impact (how it will make things worse); and a mitigation strategy. The mitigation strategy is simply the actions that need to be taken to avoid, eliminate, minimize, or manage the risk identified. In many cases, the best you'll be able to do is manage the risk, which is to say that the risk is almost certainly going to play out and you just have to contain or deal with its impact. The mitigation in this case is how you plan to deal with it.
Since the nature of complex problems is that every solution is flawed, over time there is a strong possibility that the lower-priority factors that the selected solution didn't address grow worse and may need addressing. It is perfectly fine – expected even – to have to solve the problem over again, this time with a different situation to be analyzed and a new set of prioritized objectives. That's just the nature of the beast. Complex problems are usually not resolved conclusively.

I think this is a good generic approach to "complex" problems that have to be addressed by groups of people. Now, while I think it would work just fine theoretically in government, the fact that you're "showing your hand" in terms of which aspects of the problem you're not going to address would be giving political adversaries a lot of ammunition to use against you. So I don't think this would work so well in highly polarized and toxic political environments. But what does?