THE IDEA
You can’t just do one thing
You introduce a policy to reduce one problem, and it creates two new ones. You reorganise a team to improve efficiency, and morale collapses. You build a road to ease congestion, and it attracts so much traffic that congestion gets worse.
Unintended consequences are the effects of an action that weren’t part of the plan. They show up in every complex system because systems are interconnected - pulling one thread always moves others. The question isn’t whether there will be unintended consequences. There always will be. The question is whether you’ve thought about where they’re most likely to appear.
This isn’t about being pessimistic. It’s about being realistic. In a simple system, you can predict what will happen when you make a change. In a complex system, you can’t - not fully. There are too many connections, too many feedback loops, too many delays between cause and effect.
IN PRACTICE
The cost that wasn’t in the plan
Hot-desking gets introduced to cut property costs and encourage mixing. The property saving arrives. But people feel unsettled, teams lose their sense of shared space, and the mixing never happens because nobody can find each other. The unintended cost - in productivity, belonging, and people leaving - is larger than the saving.
A performance system rewards individual output. People stop helping each other, because helping a colleague doesn’t show up in the numbers. The informal knowledge sharing, the willingness to cover for someone having a rough week - the things that made the group effective - quietly disappear. The metric goes up. The thing the metric was supposed to measure goes down.
A charity scales up a successful pilot programme. What worked in one community doesn’t transfer to another, because the conditions that made it successful were local and specific. The unintended consequence of scaling isn’t just that it doesn’t work elsewhere - it’s that the original model gets diluted in the attempt to make it universal.
WORKING WITH THIS
Trace the connections first
Before any significant intervention, ask: what else might this change? Not “what could go wrong” - that’s too negative and too vague. Instead, trace the connections. If we change X, what does X connect to? What depends on X being the way it currently is?
Run a pre-mortem. Imagine it’s a year from now and the intervention has produced a consequence nobody wanted. What is it? Where did it come from? This exercise doesn’t guarantee you’ll predict the right thing, but it makes everyone much better at noticing early signs of trouble when they appear.
Start small. The single best protection against dangerous unintended consequences is working at a scale where the consequences are containable. Try the change in one place, one group, one context. Watch what happens. Then decide whether to expand.
THE INSIGHT
The one-thing illusion
In a complex system, you can’t just do one thing. Every action has effects beyond the one you intended.
RECOGNITION
When the solution creates the next problem
When a solution creates a new problem nobody anticipated. When people say “we didn’t see that coming” about something connected to an earlier decision. When a policy produces the opposite of what it was designed to achieve. When success in one area coincides with decline in another, and nobody’s connecting the two.