Making Decisions

Most decisions do not require finding the perfect solution. They require finding a solution that is good enough.

This insight comes from Herbert Simon, who observed that people operate under a pattern referred to as "bounded rationality":

Complete information is unavailable
Time is limited
Cognitive capacity has boundaries

These constraints mean that seeking the optimal solution is often impractical or impossible.

Simon called this approach "Satisficing": a combination of "satisfy" and "suffice." Rather than searching for the best possible option, a person sets a threshold for what counts as acceptable and selects the first option that meets that threshold.

Satisficing is not settling for less. It is recognizing that the cost of continuing to search is often more expensive than the benefit of finding something marginally better.

This map shows how decisions work when perfect information and unlimited time are not available — in other words: this is how most decisions tend to actually work.

Two Modes of Decision-Making

Decision-making seems to happen at two levels of awareness.

Intuitive Decisions

Most decisions are made intuitively: a person notices a problem, they think about how to solve it, make a choice, and take action.

The thresholds are implicit or built-in
The search is unconscious

This works well for routine, low-stakes decisions.

Conscious Decisions

Some decisions benefit from deliberate attention:

Explicitly setting a threshold
Consciously managing a search budget
Systematically evaluating options

This is appropriate for high-stakes, complex, or unfamiliar decisions.

Both modes follow the same underlying structure. As described in Doing Things, this structure is the OODA loop: Observe, Orient, Decide, Act.

The loop can operate at different speeds and levels of consciousness: a snap judgment follows the same pattern as a carefully deliberated choice.

The difference is in how much attention is given to each phase.

The maps contained here suggest a structure for decision making - interestingly, the choice of whether to operate intuitively or deliberately is itself a decision — typically made quickly, based on the stakes involved.

Starting Point

Decision-making features an objective and a current state: a difference between where things are and where an objective or desired outcome exists.

It may be worth mentioning that the Objective lives in the future (or maybe it is not worth mentioning - who knows?!).

Starting point and objective with gap — A starting point and an objective

The Gap

This connects to Accomplishing Objectives: there is a current state and a desired state. The gap is what creates the need for a decision.

Understanding the gap is the first step. The second step is recognizing that not all solutions to that gap are equal.

Solution Quality

Solutions exist on a spectrum of quality. At one end is a theoretical optimal outcome: the best possible solution given perfect information and unlimited resources.

At the other end are solutions that fail to address the problem at all.

Solution quality zones — Solutions arranged by quality: from optimal at the center to failure at the edges

The diagram shows four zones:

Optimized: Near-perfect solutions, close to the theoretical ideal
Acceptable: Solutions that meet the threshold for "good enough"
Settle / Compromise: Solutions below the threshold, taken only when forced
Failure: Solutions that do not solve the problem

The boundary between Acceptable and Settle has been referred to as the "aspiration threshold" in some of Simon's works. This is the line that is drawn to define what counts as "good enough".

Where this threshold is placed determines the strategy. Place it near the center and it is attempting to optimize. Place it further out and it is satisficing.

The critical insight: the Acceptable zone is larger than the Optimized zone. There are many more satisficing solutions than optimal ones. This is why satisficing is often more efficient.

Constraints

The solution space is not unlimited. Constraints define what is actually reachable from the starting position.

Resources and limitations create boundaries — Resources and limitations create a bounded corridor of accessible solutions

Two types of constraints shape what solutions are accessible:

Resources: What is available to work with.; Time, money, skills, energy, information, tools.; Resources determine how far the search can reach and what kinds of solutions can be implemented.
Limitations: What is physically or practically impossible.; Laws of physics, legal constraints, organizational policies, missing capabilities, conditions that cannot be changed.; Limitations define hard boundaries on what solutions can exist.

Together, these constraints create a corridor of accessible solutions.

Only solutions within this corridor can be reached from the current position.

Solutions outside this corridor may be theoretically superior, but they are practically unavailable.

This connects to Doing Things: the Forces in the Context define what Forms are possible. Resources and limitations are the forces acting on the decision — they shape what solutions can fit.

Achievable Space

The corridor of constraints defines the achievable space: the complete set of solutions that could potentially be implemented.

Achievable boundary filled in — The achievable space: solutions within reach given the constraints

Notice what the diagram reveals:

Part of the Acceptable zone falls within the achievable space — satisficing solutions exist
Part of the Optimized zone falls outside the achievable space — some theoretically better solutions are unreachable
Most of the Settle and Failure zones are beyond the achievable boundary

This is why optimization is often irrational. The optimal solution may be theoretically superior, but if it lies outside the achievable space, pursuing it wastes resources that could be spent finding an acceptable solution that is actually reachable.

The rational strategy is to search for solutions in the intersection of Acceptable and Achievable. This is what satisficing does.

Search Radius

Even within the achievable space, every possible solution cannot be examined. Search itself requires resources — primarily Time.

The search radius represents how far can be looked within the available time. It is determined by:

Search budget: Total time available for finding and evaluating solutions before a decision must be made
Assessment time: How long it takes to evaluate each potential solution

The number of solutions that can be examined equals the search budget divided by the assessment time. With 10 hours to decide and each option taking 2 hours to evaluate, at most 5 options can be examined.

This creates a tradeoff: more time per evaluation means fewer options examined. Less time per evaluation means more options surveyed but with less confidence in each assessment.

The search radius is typically shorter than the achievable boundary. This means some achievable solutions will not be discovered within the available time. The search covers a sample of the achievable space, not the entire space.

This is bounded rationality in action: not all options can be known, so the search operates within practical limits and selects from what is found.

Finding Solutions

Within the search radius, actual solutions exist. They are scattered throughout the space, distributed across different quality zones.

Possible solutions distributed across quality zones within the searchable space

The distribution of solutions shows:

Solutions cluster more densely in the middle ranges — there are more acceptable solutions than optimal ones
Several solutions sit in the intersection of Achievable, Searchable, and Acceptable — these are viable candidates
Some solutions fall outside one or more boundaries — theoretically good but practically unreachable, or discoverable but inadequate

With this kind of arrangement, the satisficing strategy becomes clear: search within the radius, evaluate each option against the threshold, and select the first one that lands in the Acceptable zone.

There is no need to find the solution closest to the objective. Any solution that meets the threshold and lies within the constraints will do. Once found, the search can stop.

This is more efficient than optimization, which requires examining all options to ensure the best one has been found. Satisficing allows the search to stop as soon as something good enough turns up.

The Process

The map above shows the geometry of decision-making: the space of possible solutions and the constraints that bound it. The flow below shows the procedure: the steps involved in moving through that space.

The process operates at two levels. The intuitive level is what most decisions feel like: notice a problem, think about it, decide, act. The conscious level is what happens when there is opportunity to slow down and make each step explicit.

The intuitive process: how decision-making feels

At the intuitive level:

Notice a Problem: Recognize the gap between current and desired state
Think about how to solve it: Generate and evaluate possibilties
Make a Decision: Select an option
Take Action: Implement the choice

This is the OODA loop in its most basic form: Observe the problem, Orient to possible solutions, Decide on one, Act on it. The cycle completes quickly, often below the level of conscious awareness.

When a decision warrants more attention, the same structure expands into a more deliberate process:

The conscious process: making implicit steps explicit

The detailed process breaks down as follows:

Notice a Problem expands into:

Identify the objective — what is the desired outcome?
Identify current situation — where is the situation right now?
Assess the gap — how far is the distance between current and desired?

Think about how to solve it expands into:

Assess available resources — what is available to work with?
Assess core limitations — what boundaries cannot be crossed?
Refine objective and set acceptable outcome — where is the threshold?
Determine search budget — how much time is available to look for solutions?
Determine assessment time — how thoroughly will each option be evaluated?

Make a Decision expands into:

Generate possible solutions
Evaluate each solution against the threshold
If threshold is met and search time remains: assess whether continuing the search is worth the cost
If threshold is met: select the solution
If threshold is not met and search time remains: continue searching
If threshold is not met and search time is exhausted: choose among three options

Take Action remains the same: implement the chosen solution and observe the outcome.

The outcome then feeds back into the OODA loop. The result of action becomes the new current situation, potentially creating a new gap and starting the cycle again. This connects to Accomplishing Objectives: each iteration provides information about what is achievable.

Outcomes

When search time is exhausted without finding a solution that meets the threshold, three paths become available:

Settle: Accept the best solution found, even though it falls below the threshold.; This is compromise. Progress is made toward the objective, but not as much as intended.; This often happens when the deadline is firm and some progress is better than none.
Pivot to New Objective: Recognize that the search process revealed the original objective was misconceived or unachievable.; Reframe the problem based on what was learned.; This is not failure — it is learning. The search provided information about what is actually possible, allowing for a more realistic objective to be set.
Accept Failure: Abandon the objective entirely.; This happens when settling would create more problems than it solves, and no better framing of the objective is available.; Sometimes the rational choice is to do nothing and accept the current state.

The distinction between these outcomes matters. Settling is choosing an inadequate solution under pressure. Pivoting is intelligently reformulating based on new information. Accepting failure is recognizing when the cost of any solution exceeds the benefit.

All three outcomes feed back into the OODA loop. Settling creates a new current state that may trigger another decision cycle. Pivoting creates a new objective. Accepting failure means living with the current state — which may itself create new pressures that eventually trigger a different decision.

When to Satisfice

The choice between satisficing and optimizing is itself a decision. Not all situations call for the same approach.

Satisficing is appropriate when:

Time is limited: The opportunity window closes, or the decision loses value with delay.
Stakes are low or reversible: The cost of a suboptimal choice is small, or the decision can be revisited and changed.
Information is expensive: Gathering more data costs more than the potential improvement in outcome.
Evaluation is difficult: The decision involves multiple competing criteria with no clear way to compare them.
Good enough is actually good enough: The difference between acceptable and optimal does not matter for the purposes at hand.

Optimization is appropriate when:

Stakes are high and irreversible: The cost of getting it wrong is severe, and the decision cannot easily be undone.
Small differences compound over time: Initial choices create path dependencies that amplify, making early optimization valuable.
The luxury exists: Time is abundant, information is cheap, and the search process itself has value.
Marginal gains matter competitively: Small edges accumulate or determine success in zero-sum contexts.

Most decisions fall into the satisficing category. The exceptions — truly high-stakes, irreversible decisions where optimization is warranted — are rarer than people assume.

A common mistake is optimizing when satisficing would serve better. This shows up as analysis paralysis: endlessly researching options for a decision that does not warrant the investment. The search cost exceeds the potential benefit.

The reverse mistake — satisficing when optimization is warranted — is less common but more costly. This happens when the irreversible consequences or compounding effects of a decision go unrecognized.

The skill is calibrating the threshold and search budget to the actual decision context. High-stakes decisions warrant tighter thresholds and larger search budgets. Low-stakes decisions warrant looser thresholds and smaller search budgets.

The decision about how to decide is typically made quickly, based on intuitive assessment of stakes and time pressure. This meta-decision does not itself require optimization.

Broader Context

Decision-making sits within the larger cycle of action described in Doing Things. The OODA loop — Observe, Orient, Decide, Act — provides the framework. This map shows the detailed mechanics of the Decide phase.

Observe identifies the gap between current state and objective. Orient assesses resources, limitations, and sets the aspiration threshold. Decide searches for and selects a solution. Act implements the choice and generates an outcome.

That outcome feeds back into Observe, creating a new current state and potentially revealing new gaps. This connects to Accomplishing Objectives: multiple decision cycles accumulate, each one providing information about what is achievable.

Through iteration, boundaries are discovered. Some objectives turn out to be unreachable given current constraints. When this becomes clear, the Pivot option becomes valuable — reformulating the objective based on what the search process revealed.

The speed at which decision cycles are completed affects how quickly learning occurs. Faster iteration means faster discovery of what works and what does not. But speed alone does not guarantee good outcomes. The threshold must be calibrated correctly, and the search must be conducted within the achievable space.

Satisficing is not a compromise with rationality. It is rationality adapted to real constraints. Perfect information is unavailable. Time is finite. Cognitive capacity has limits. Given these conditions, finding a good enough solution quickly is often more rational than searching indefinitely for the best possible solution.

The approach described here — setting thresholds, managing search budgets, recognizing when to pivot — applies whether decisions are made intuitively or deliberately. The structure is the same. The difference is in how much conscious attention is given to each phase.

For routine decisions, the intuitive mode works well. The threshold, search, and evaluation happen below conscious awareness. For complex or high-stakes decisions, making the process explicit allows for better calibration and more effective use of limited resources.

Either way, the goal is the same: find a solution that fits the forces in the context, given the constraints at hand, in the time available.