Complicated vs. Complex, Part II: Solving the World’s Most Difficult Challenges


This page in:

Confronting the hardest problems on the planet requires humility to admit that we don’t know many answers when we start and sometimes we don’t even know the right problem to work on. If we address symptoms rather than root causes, we can exacerbate conditions. Penalizing teachers for example for not coming to school may ignore very real issues related to over-crowded classrooms, transport or meager wages for educators. If you start with the wrong problem, you’ll certainly propose the wrong solution.
What would it take to accept that most of the problems we encounter in development require listening better to end-users, learning about technical and political obstacles, and an ability to course correct when conditions change? That requires flexibility, faster response times, and treating beneficiaries as partners in solving complex problems.
I recently blogged on the difference between complicated and complex systems: the importance of identifying each to solve problems and particularly to scale solutions. What follows is a brief description of what makes a system complicated or complex and why it matters. 

  Complicated Complex
In planning
  • Describe What; dictate How
  • Focus on details
  • Coordinate everything centrally
  • Deliberate tradeoffs
  • Solution is often reached through a series of algorithms
  • Describe What but not How
  • Only key details—the fewer, the better
  • Limit central coordination to what’s absolutely necessary
  • Tradeoffs not always foreseeable, and they can shift over time
Goal Optimal solution Good enough to learn from and adjust
Focus on All the details Potential side effects
During execution
  • Make sure plan is adhered to
  • Adjust to make things more efficient
  • Compliance 
  • Measure results against all desired outcomes
  • Don’t get attached to any particular course of action
  • Adjust constantly and learn

Scaling-up what works may have more to do with process expertise – figuring out what problems need to be solved in a given environment – than blueprints based on expert knowledge. Expert-driven organizations are too often hammers looking for nails rather than solution seekers looking to partner with local experts to solve local problems.
The risk with over-resourcing complicated problem-solving techniques in complex environments is that solutions feel forced, don’t resonate with end-users, and more importantly don’t solve underlying problems. A glaring example is the construction of the Choluteca Bridge in Honduras. Although the bridge was constructed by the U.S. Army Corps of Engineers in the 1930s and remains technically sound, the road it was connected to moved in 1998 after Hurricane Mitch; today, it’s a bridge to nowhere. Though structurally flawless, without attention to shifting realities on the ground, it serves no purpose.
And often solutions even when they do work, don’t scale. We’ve seen mobile telephony and micro-lending spread like wildfire while toilets and basic sanitation continue to be unavailable for more than a billion people. We can’t impose solutions on people. They have to want them and often demonstrate their demand through markets and other instruments.
But the biggest lesson is that of differentiating between solutions that can be replicated easily and usefully (say building a road) and those that require an understanding of local conditions and require an experimental approach driven by data, feedback, learning, and adaptation.

We are at such a fork in the road. The hardest problems of the world - climate change, food security, youth employment are complex and require data-driven experimentation and adaptation to solve. Solutions may not always scale. But the processes by which we experiment, learn, and adapt can scale. That’s the challenge for global development institutions in the 21st century. 


Aleem Walji

Director, Innovation Labs

Join the Conversation

Fadi El-Eter
November 13, 2013

Hi Aleem,
For some reason, and from a project management perspective, I feel that complicated systems are managed by waterfall and complex systems are managed using Agile. This is very obvious when you read the point "Adjust constantly and learn"
I'm not sure whether Agile has proved itself (yet) to solve complex projects.

November 14, 2013

Dear Aleem,
thanks for the great post!
I agree with your stements in quadrants. But how will you set up the budget in your project if you only describe what you want to do but not how?
It would be great if we could focus only on the effects of an intervention but as long as you work with tax money the supreme audit institution of the respective donor country will have its say.
I believe we shouldn´t really worry about activities. We should be measured against achievements. But I would like to see the face of a federal auditor if he sees the following budget:
1. Personal Cost: 5.000.000
2. Awareness raised: 2.000.000
3. Change in behaviour: 4.000.000
4. Change in Law: 500.000
5. Improved Livelihood: 2.000.000
6. Contingency: 500.000
I´m exagerating. But do you get my point?

November 17, 2013

Great article! I completely agree. It goes back to Amartya Sen's capability approach and the importance of working on solutions that are valued by the community - and you cant do that unless you constantly engage, evaluate and adapt your model to local feedback. Those are the solutions that work and are sustainable.

December 27, 2013

Yes, this is a very descriptive article on analysis to a problem. Many times we try with logic to fix a problem. We do it everyday without thinking. But, when there is a problem possibly undefined to the eye, we have to ensure that the information we are collecting is indeed measurable. If we do this, our analysis will show us the weak and strong areas, as well as areas that need improvement.
As for end from end users is the one part of the puzzle that many corporations do not want to fully utilize, fearing that they are giving in...or giving the user to much inclusion. But if they use the feedback from the users correctly, then that can be added to the data collection which will then give you the optimum result.
Those are my thoughts for this Friday. Have a great weekend everyone!

Dr. Erwin Rooze
December 30, 2013

The blog is correct in stating that when one identifies the symptoms rather than root causes as the problem, the risk is extremely high that one proposes the wrong solution. Since I earned my Ph.D. on the subject of formulating strategic problems, I can comfortably state that the blog should indeed have been about complex problem situations (which are complex but static) and wicked problem situations (which are complex and dynamic). Needless to say that a full quadrant also has the simple and static problematic situation, and the simple and dynamic problem situation. They all require different methodologies. In the blog the static dimension is depicted as the engineering types of problems (building a road / bridge) and the dynamic dimension as social types of problems. However, it is not by definition that an engineering type of problem is a complex/static problem, and a social type of problem is a wicked (complex/dynamic) problem. I believe you are absolutely correct that when dealing with wicked problems one should interact intensively with the changing problem situation and use an approach driven by experimentation, learning, and adaptation. The words used suggest an evolutionary approach. This is at least the notion I used in my Ph.D. research. We must however not forget that lately the notion of “super wicked” problems was invented for problems like climate change, since they have a number of additional features that make them harder to address: (a) Time is running out; (b) There is no central authority; (c) Those seeking to solve the problem are also causing it; (d) Policies seem to discount the future irrationally (see Levin, Kelly; Cashore, Benjamin; Bernstein, Steven; Auld, Graeme (23 May 2012). "Overcoming the tragedy of super wicked problems: constraining our future selves to ameliorate global climate change". Policy Sciences 45 (2): 123–152). One of the pressing questions is if they ca n be dealt with by just "muddling through" (Lindblom) or need a grand solution design as well."