Resilience Over Maximum Optimization
Trade a little efficiency for systems that survive unpredictable shocks
- Difficulty
- Advanced
- Time to result
- ~months to results
- Steps
- 5
- Confidence
- 98%
This framework treats efficiency and resilience as competing design goals rather than assuming optimization is universally beneficial. Highly interconnected systems can turn a gust of wind, supplier failure, software fault, or mistaken intervention into disproportionate damage. Practitioners identify concentrated dependencies, estimate the full cost of failure, and deliberately purchase slack through buffers, redundancy, alternative routes, modularity, and reversible decisions. These safeguards may reduce short-term margins or speed, but they protect long-term continuity and make adaptation easier when causal conditions change. The framework does not reject optimization; it limits optimization at points where a small disruption could cascade through the entire system.
Origin
Extracted from Masters in Business. Brian Klass used the Suez Canal blockage, Mao's Four Pests campaign, and the fragility of interconnected systems to argue that optimization carries an often-overlooked resilience cost.
Core principles
- 01Optimization and resilience frequently trade off.
- 02Efficient systems can amplify small disruptions.
- 03Redundancy has a visible short-term cost and an invisible protective value.
- 04Interconnectivity increases the reach of local failures.
- 05Uncertainty favors optionality and slack.
How to run it
- 1
Map critical dependencies
Document the suppliers, routes, people, systems, and assumptions required for continued operation. Mark dependencies whose failure would halt the whole system.
Pro tip Include dependencies shared by several apparently independent processes.
Watch out Interconnected components can conceal a common point of failure.
- 2
Price failure honestly
Estimate lost revenue, recovery time, reputational damage, safety consequences, and downstream disruption. Compare these costs with the expense of protection.
Pro tip Model prolonged and correlated failures, not only average downtime.
Watch out Short-term accounting often understates the value of resilience.
- 3
Buy targeted slack
Add inventory, capacity, time, capital, or staffing buffers where failure would be most damaging. Concentrate redundancy rather than spreading it indiscriminately.
Pro tip Start with the dependency having the highest combination of impact and irreversibility.
Watch out Redundancy without independence may provide only an illusion of protection.
- 4
Create alternatives
Maintain substitute suppliers, routes, tools, or operating modes that can take over during disruption. Exercise them before an emergency.
Pro tip Use genuinely different failure domains for primary and backup options.
Watch out An untested fallback is not a reliable fallback.
- 5
Stress-test and adapt
Simulate abnormal shocks and examine how failures propagate. Revise optimization targets when testing reveals cascading risk.
Pro tip Test combinations of unlikely events rather than one isolated failure at a time.
Watch out Past reliability does not prove the causal environment will remain stable.
In the wild
A gust of wind turned one container ship sideways in the Suez Canal. Because global supply chains were highly optimized around that route, the local obstruction propagated across industries and reportedly caused tens of billions of dollars in economic damage.
→ A single physical blockage exposed the systemic cost of concentrated, tightly optimized logistics.
A manufacturer pays slightly more to qualify a second supplier in another region and holds two weeks of critical inventory. When the primary supplier closes unexpectedly, production shifts to the alternative rather than stopping.
→ Higher routine costs prevent a much larger interruption and protect customer commitments.
Common mistakes
Optimizing only for average conditions
Average performance hides tail risks and correlated failures. A system can look excellent until one shock disables everything.
Treating redundancy as waste
Unused capacity appears inefficient during calm periods, but its purpose is to preserve continuity under abnormal conditions.
Building identical backups
Two options exposed to the same failure mode do not create genuine resilience. Alternatives must be independent enough to fail differently.
Is it for you?
Best for
It is best for leaders designing supply chains, infrastructure, organizations, and portfolios exposed to uncertain disruptions.
Not ideal for
It is not ideal where failure has negligible consequences and efficiency can be restored immediately at low cost.
From the transcript
“there's a trade off very often between optimization and resilience.”
“they come at a cost. They do create less resilience”
“put a little bit more into resilience and a little bit less in optimization.”
From the episode
Cause and Effect on a Global Scale