THE IDEA
Two ways to survive
When something threatens a system, there are two fundamentally different survival strategies. Robustness says: build the walls higher. Make the system so strong that the threat can’t get through. Resilience says: let the threat in, bend with it, and reorganise.
A fortress is robust. Thick walls, narrow windows, a moat. It resists attack by preventing penetration. It works brilliantly against threats it was designed for. But it can’t move, can’t adapt, and if the threat exceeds what the walls can handle, the failure is catastrophic.
A reed is resilient. The wind pushes it flat. When the wind passes, it springs back. It doesn’t resist the force - it absorbs it and recovers. It survives not because it’s strong enough to resist, but because it’s flexible enough to yield without breaking.
Neither strategy is better. They’re suited to different conditions. When threats are known, bounded, and predictable, robustness is efficient. Build for the known threat. When threats are unknown, varied, and surprising, resilience is essential. You can’t build walls against things you can’t predict. You need the capacity to absorb, adapt, and recover from whatever arrives.
IN PRACTICE
Fortresses and reeds
A data centre invests heavily in uninterruptible power supplies, redundant cooling, and physical security. These are robustness measures - they prevent known threats (power outages, overheating, break-ins) from reaching the servers. The strategy works because the threats are well-understood and bounded. But when a novel threat appears - a new type of cyberattack, a supply chain disruption for replacement parts - the fortress strategy has no answer. The system needs resilience: the ability to detect the novel threat and improvise a response.
A large corporation builds elaborate compliance systems to prevent regulatory violations. Robust: detailed procedures, approval chains, training programmes. It works for known regulations. But when a new regulation arrives or an old one is interpreted differently, the system freezes. The procedures don’t cover this case. The approval chains don’t know who decides. The organisation that was robust against known risks is paralysed by an unfamiliar one. A more resilient approach would maintain fewer rigid procedures and more adaptive capacity - people with judgement, access to information, and the authority to act.
A person builds their financial life around a single high-paying job. Robust in one sense: high income, growing savings, clear trajectory. But entirely dependent on that one source. When the job disappears, there’s no alternative income, no portable skills outside the specialism, no network in other industries. A more resilient financial life might earn less but depend on multiple income streams, diverse skills, and a wide network. Less optimal in stable times. Far more survivable in unstable ones.
WORKING WITH THIS
Choosing your strategy
The first question is: how predictable is your environment? If the threats you face are known and bounded, invest in robustness. Build strong defences against the specific risks you can identify. This is the right strategy for engineering systems, regulatory compliance, and any domain where the failure modes are well-understood.
If your environment is unpredictable - and most human, social, and market environments are - invest in resilience. Build the capacity to detect, absorb, and recover from threats you can’t specify in advance. This means diversity (multiple approaches, not one), modularity (parts that can fail independently), feedback (early warning signals), and slack (resources that aren’t committed to current operations).
Most real-world systems need both. Robust against the known threats, resilient against the unknown ones. The mistake is over-investing in one at the expense of the other. Too much robustness creates brittleness - the system handles everything it was designed for and shatters at the first thing it wasn’t. Too much resilience without any robustness means absorbing shocks that could have been simply prevented. The art is getting the balance right for your context.
THE INSIGHT
The line to remember
Build walls against what you can predict. Build flexibility against what you can’t. The mistake is building only one.
RECOGNITION
When this is in play
You’re choosing between robustness and resilience when a system handles familiar challenges beautifully but freezes at unfamiliar ones - that’s robustness without resilience. When a system absorbs every shock but could have prevented most of them - that’s resilience without robustness. When an efficiency drive removes slack, redundancy, or flexibility in the name of optimisation - that’s trading resilience for short-term performance. When someone says “we didn’t plan for that” about something that was inherently unplannable - that’s a sign the system needed resilience, not a better plan.