THE IDEA
Two different kinds of hard
There’s a word people use for anything that’s difficult, and it hides one of the most important distinctions in systems thinking. “Complicated” and “complex” sound like synonyms. They’re not. Confusing them is the source of enormous waste, frustration, and failure.
A complicated system has many parts, but they interact in predictable ways. A jet engine has thousands of components, but an expert can understand every one. Given the same inputs, it produces the same outputs. You can take it apart, study each piece, put it back together, and it works the same way. Complicated systems reward expertise, analysis, and planning. If you understand the blueprint, you can control the outcome.
A complex system is different in kind, not just degree. It’s made up of many agents that interact, adapt, and learn. Their behaviour creates patterns that can’t be predicted from knowledge of the individual parts. A rainforest, a city, a family, a market - these aren’t just complicated. They have properties that emerge from interaction, that change when you observe them, that resist the very idea of a blueprint. Complex systems don’t reward better analysis. They reward better attention, faster learning, and a tolerance for being surprised.
IN PRACTICE
The wrong toolkit for the job
A hospital builds a detailed protocol for surgical procedures - step-by-step, highly specified, consistently followed. Outcomes improve dramatically. This is a complicated problem solved with a complicated solution. The same hospital then builds a detailed protocol for patient recovery at home - step-by-step, highly specified. It barely works. Recovery at home involves family dynamics, mental health, diet, motivation, access to support - a complex system that doesn’t respond to detailed scripts. The toolkit that saved lives in surgery fails in the living room.
A city government commissions a masterplan for a new neighbourhood. The roads, utilities, and building specifications are complicated - and the planners get them right. But the plan also specifies where the “vibrant community hub” will be, where the “thriving market” will emerge, and how “diverse social mixing” will happen. These are complex outcomes that can’t be engineered by drawing them on a map. Ten years later, the buildings are exactly as planned. The community is nothing like the brochure.
Someone decides to learn a language. They buy the most comprehensive textbook, create a study schedule, work through every grammar exercise. This is treating language like a complicated problem - master the components and the whole follows. Months later, they can conjugate every verb and can’t hold a conversation. Language in use is complex - it depends on context, improvisation, culture, and the unpredictable responses of other people. The textbook addressed the complicated part. The complex part needed immersion.
WORKING WITH THIS
Asking the right question first
Before choosing a strategy, ask: is this complicated or complex? The answer determines almost everything about how to proceed.
If it’s complicated, invest in expertise, analysis, and planning. Break the problem into parts, understand each one, and assemble the solution. Bring in specialists. Build detailed plans. Expect that good analysis will produce good outcomes.
If it’s complex, invest in experiments, learning, and adaptation. Don’t try to blueprint the outcome. Instead, run small probes, watch what happens, and amplify what works. Bring in diverse perspectives rather than narrow expertise. Build the capacity to respond quickly rather than the capacity to predict accurately. Accept that you’ll be wrong often, and design your approach so that being wrong is cheap and being right can be scaled.
The biggest failures happen when people treat complex problems as merely complicated - when they bring a blueprint to a situation that needs a compass. The tell-tale sign: a plan that’s incredibly detailed, took months to produce, and assumes the world will hold still while it’s executed.
THE INSIGHT
The line to remember
A complicated problem rewards better planning. A complex one rewards faster learning. Using the wrong approach doesn’t just waste time - it makes things worse.
RECOGNITION
When this is in play
You’re facing this distinction when a detailed plan keeps needing revision before it’s even finished. When experts disagree not about the answer but about what the question is. When the same approach produces wildly different results in different contexts. When someone says “we just need to understand it better” about a situation that gets less predictable the more you study it. When the problem changes shape in response to your attempt to solve it - that’s complexity announcing itself.