Local vs Global Maxima: Why Short-Term Wins Can Trap You

Introduction

Local vs global maxima is a mental model for understanding why short-term wins can trap you. A local maximum is the best nearby result. A global maximum is the best result across the wider landscape.

The problem is that many decisions only look at the next step. If every nearby move seems slightly worse, you may stay where you are even when a much better option exists farther away.

This happens in careers, products, habits, businesses, relationships, investing, and personal strategy. You improve what already works. You polish the current system. You avoid the awkward dip that comes with trying something different. Over time, the current peak becomes a cage.

The local vs global maxima model helps you ask a better question: are you optimizing the best path, or just the best version of the path you are already on?

Aspect	Local Maximum	Global Maximum	Best Use Case
Core idea	Best nearby result	Best overall result	Strategy and tradeoffs
Time horizon	Usually short to medium term	Usually longer term	Long-range decisions
Main risk	Getting trapped by incremental gains	Chasing unrealistic perfection	Choosing where to optimize
Typical move	Improve the current system	Search the wider landscape	Career, product, and life choices
Warning sign	Small changes help, big changes feel scary	Exploration never ends	Balancing exploitation and exploration

What Are Local and Global Maxima?

A maximum is a high point on a landscape of possible outcomes. In mathematics, that landscape may be a graph. In life, the landscape is the set of choices available to you.

A local maximum is a point that is better than the options immediately around it. If you move a little left, right, forward, or backward, things get worse. From close up, it looks like the best place to be.

A global maximum is the best point on the entire landscape. It may be far away from your current position. Getting there may require moving downhill first, crossing uncertainty, learning new skills, or abandoning a strategy that currently performs well.

Imagine hiking through fog. You climb a hill and reach a good viewpoint. Every nearby step down looks worse, so you stop. But beyond the valley there may be a much higher mountain. To reach it, you must first descend.

That is the emotional difficulty of this model. The better long-term path often begins with a short-term loss.

Why Local Maxima Are So Sticky

Local maxima are sticky because they reward reasonable behavior.

If something works, improving it is usually sensible. You get feedback. You reduce risk. You build confidence. You make the next version slightly better than the last one.

This is how skills develop, companies grow, and systems become reliable. Incremental improvement is not the enemy. The danger appears when incremental improvement becomes the only kind of thinking you trust.

Once you are on a local peak, most immediate alternatives look worse. A new job may pay less at first. A better product direction may require rebuilding. A healthier routine may feel uncomfortable. A new business model may reduce revenue before it expands the market.

Your mind compares the messy beginning of the new path with the polished version of the current path. That comparison is unfair, but persuasive.

Local maxima also become part of identity. You are the person with this expertise, this audience, this product, this process, this reputation. Moving away can feel like wasting what you built.

That is why this model pairs naturally with the sunk cost fallacy. Past investment can make the current peak feel more valuable than it really is.

Local Maxima vs Global Maxima in Strategy

The local vs global maxima distinction is especially useful in strategy because strategy is about choosing what not to optimize.

If you optimize everything inside the current system, you may make the system harder to leave. A company can perfect a declining product category. A professional can become excellent at work they no longer want to do. A creator can become trapped by the content their existing audience expects.

The better question is not always "How can I improve this?" Sometimes it is "Is this still the right game?"

Consider a business that sells software to small companies. Its current product has steady revenue. The team can improve onboarding, add features, and reduce support tickets. These are all useful local improvements.

But suppose the market is moving toward integrated platforms. The company may have two options:

• keep improving the current product and defend a shrinking niche
• accept a difficult transition toward a broader platform with a larger future market

The first path may win every quarterly comparison. The second path may look worse for a year. But if the broader platform is the higher peak, the temporary dip is not failure. It is the valley between peaks.

This is why strategic decisions often feel wrong before they become right.

Real-World Examples

Career growth

A person can become very good at a role that no longer fits their future. They earn trust, status, and efficiency. Their current job becomes a local maximum.

Moving into a different field, starting a business, or learning a new technical skill may reduce income or confidence at first. The early stage is awkward. They are no longer the expert.

If they only compare this month with next month, staying wins. If they compare the next decade, switching may be the better path.

The point is not that everyone should leave a stable career. The point is that short-term competence can hide long-term stagnation.

Product design

A team may improve a product by listening closely to current users. They add requested features, refine familiar workflows, and remove friction from existing behavior.

That can be excellent product work. But it can also create a local maximum. Current users often ask for a better version of what they already know. They may not describe the simpler, broader, or more disruptive product that would serve a larger market.

The team must occasionally ask: are we making this product better, or are we making it harder to reimagine?

Personal habits

A routine can be locally optimal. You know how to get through the week. Your calendar works. Your coping mechanisms work well enough. Nothing is broken enough to force change.

But a better life may require a temporary drop in comfort: sleeping earlier, exercising before you enjoy it, ending a commitment, changing your environment, or learning to say no.

The old routine may be the best version of an unhealthy pattern. A better pattern may require a valley.

Investing attention

Attention also has local maxima. You may become efficient at handling messages, meetings, errands, and small tasks. You feel productive because the feedback is immediate.

But the global maximum may be deep work, better relationships, health, or building an asset that compounds. Those options often feel worse at first because they are less immediately rewarding.

The local peak is visible. The global peak needs imagination and patience.

Common Mistakes

The first mistake is assuming every short-term decline is strategic.

Some declines are just declines. Leaving a local maximum only makes sense when there is a plausible higher peak and a credible path toward it. You need evidence, not just restlessness.

The second mistake is chasing the global maximum forever.

Some people use the idea of a better peak to avoid commitment. They keep searching, comparing, and restarting. That is not strategy. It is avoidance wearing a thoughtful coat.

The third mistake is optimizing too early.

If you polish the current path before exploring alternatives, you may invest heavily in the wrong hill. Early in a project, career, or product, exploration is often more valuable than refinement.

The fourth mistake is ignoring switching costs.

Moving from one peak to another has costs: time, money, reputation, emotional strain, and opportunity cost. A global maximum must be meaningfully better, not merely different.

The fifth mistake is asking only people who benefit from the current peak.

Colleagues, customers, friends, and partners may all have incentives to keep you where you are. Their feedback can be useful, but it may reflect the landscape around your current position rather than the full landscape.

How to Apply the Model

Start by naming the current peak. What are you optimizing right now? Revenue, comfort, status, efficiency, safety, approval, speed, certainty, or short-term ease?

Then ask what the wider landscape might contain. Are there options you have stopped considering because the first move looks worse? Are there paths that would require a temporary dip but could create a better long-term position?

A useful practice is to separate local improvement from global search.

Local improvement asks:

• How can I make the current system better?
• What is the next obvious upgrade?
• Where is the friction in the existing path?

Global search asks:

• Is this still the right system?
• What would I choose if I were starting fresh?
• Which better options require an uncomfortable transition?
• What would make the current path obsolete?

Both modes matter. You do not want to abandon every good thing in search of a fantasy. You also do not want to polish a trap.

Use small experiments when possible. If you are considering a career shift, test the new field with a side project, course, contract, or conversation before making a dramatic leap. If a product may need a new direction, prototype the alternative before rebuilding the whole business.

Small experiments reduce the cost of descending from the current peak. They reveal whether the higher mountain is real.

Also define the valley in advance. What short-term pain are you willing to accept? Lower revenue for six months? Beginner status for a year? A slower product cycle while the architecture changes? If the valley is unnamed, it will feel like failure when you enter it.

Finally, set review points. A temporary decline should not become an endless excuse. Decide when you will reassess, what evidence you expect to see, and what would make you change course.

When to Stay on the Local Peak

Sometimes the local maximum is good enough.

You do not need to chase the global maximum in every area of life. Many decisions are not worth endless optimization. A reliable tool, stable routine, decent process, or satisfying role may be perfectly fine.

The model is most useful when the stakes are high, the time horizon is long, or the current path seems increasingly constrained.

Stay on the local peak when the upside of moving is small, the switching cost is high, or the current position supports deeper priorities. Leave it when the current peak is becoming fragile, the wider opportunity is clearly better, and the temporary decline is manageable.

Good judgment is not constant disruption. It is knowing which hills deserve loyalty and which ones only look tall because you are standing on them.

Final Thoughts

Local vs global maxima explains a common trap: the best nearby move can prevent the best overall outcome. Short-term wins are useful, but they can also make a limited path feel safer than it is.

Use the model when you feel stuck despite steady progress. Ask whether you are improving the current hill or searching the full landscape. Then decide whether the valley between peaks is worth crossing.

If you want a deeper framework for using mental models in everyday decisions, 100 Mental Models expands on these ideas in a broader and more practical way.