Local Optima Versus Global Optima: Secure Overall Value

Local optimums are whatever is best for the performance of an individual part. Global optimums are what is best for the performance of the system as a whole. Adding up a series of local optima does not automatically result in a global optimum. Imagine the design of a car. The local optimum would be hiring the most capable experts to make the very best component they could for each part of the car. But this would add up to a mess of a vehicle, since everybody was concentrated on the local optimum, not the performance of the whole.

This is the idea of local versus global optimum in complex systems. Local optimums are not only ineffective, but sometimes harmful. This is manifest throughout industries: engineering, project, product, design, accounting, sales. Sales closes a deal. Product scopes the project and sends it to the design team for wireframing. Design passes their sketches to engineers to build it. Everyone is jolly, until engineering’s capacity is filled and sending more work their way won’t result in any more production. But sales, product, and design teams operate under the “work hard” principle. This is, in effect, a local optimum.

So they start new projects. More work is sent down the pipeline. Now the engineers are overloaded, and this cycle repeats itself. This exacerbates the cycle further. Meanwhile, downstream work centers are waiting because their capacity is not fulfilled. Thus they start projects of their own. Engineering is drowning in overhead. Everyone works late, works weekends, and eats lunch at their desks to keep up. Hence a “busy” company isn’t necessarily doing good. The only way it is possible to stay efficient is for everyone to optimize both their individual and the team’s productivity, not merely their station. Else you risk the entire pipeline.

The concept of local and global optima is fundamental to optimization, decision-making, and many natural and artificial systems. The local is the best solution within a limited area and subset of possibilities. It’s rarely the overall best. The global is the best possible solution taking into account all the possibilities. You want the global optima whenever it is accessible. Even though local optima are easier to find, and global optima often demand innovation, exploration, and risk-taking. The challenge is that many decision-makers converge on local optimums, mistaking them for the best possible solutions. They exploit readily available and existing solutions rather than exploring the better and unknown ones.

Real life implications of local versus global optima:

• Business: the local is improving existing products or services, whilst the global is innovating with an entirely novel business model or expanding into a new market; Kodak is a notebook case of a company that chose local optima—their film products—and not global optima—they failed to shift to digital photography—resulting in the collapse of the company;

• Career: the local is remaining in a comfortable and secure job with steady pay, whilst the global is pursuing a career shift or extra development to unlock your full potential; key examples are leaving a job to start a business or acquire skills for a dream role;

• Fitness: the local is following a basic fitness routine to maintain health, whilst the global is implementing a comprehensive lifestyle change to optimize your well-being; this is the transition from casual exercise to a structured training program;

• Innovation: the local is using existing technology for incremental benefits, while the global is improving and developing disruptive innovations; electric vehicles were a representation of a global optima while optimizing the internal combustion engine was a local optima;

• Relationships: the local is maintaining surface-level harmony, whilst the global is assessing underlying issues for a deeper, more stable, lasting connection.

How you might use local versus global optima as a mental model: (1) first, overcome the local optima by asking “What am I giving up by staying here?” or “What could I be doing differently?” to challenge yourself into exploring untried ways, even if they result in temporary setbacks; (2) avoid prioritizing convenience at the expense of greater rewards, such as by allocating time to learn skills in lieu of pursuing easy tasks; (3) ask others to see if you overlooked ideas or opportunities for global optima; (4) evaluate the scale of your decisions when you make them, assessing whether they are local- or global-level solutions; (5) avoid being content with incremental gains when larger, long-term improvements are possible. Examples: building foundational skills instead of niche know-how, investing in education now for higher earning potential later, expanding into an untried market instead of just optimizing operations and marketing campaigns as a business, abandoning a stagnant and comfortable strategy to explore untapped demand.

Thought-provoking insights. “Don’t climb a tree until you are sure it is the right one.” and “Make sure your ladder is leaning against the right building before you start climbing.” reflect the risk of pursuing local optima without understanding the broader implications. “Sometimes, the best way forward is a step back.” highlights that escaping a local optimum often means revisiting past decisions or abandoning progress. “The greatest enemy of progress is not failure, but comfort.” underlines how staying in a local optimum is easy but limits growth. Don’t settle for local optima. Seek broader opportunities. Pursue great, not good.