Archive note: Originally published on LinkedIn on March 29, 2026. This article is preserved here as part of the evolution of our thinking on uncertainty, complexity, and the limits of analysis in organizational decision-making.
Organizations invest heavily in improving decisions.
Very few invest in understanding what kind of decisions they are actually making.
In most organizations, decision-making is structured around a simple assumption: better analysis leads to better decisions. Data is collected, models are refined, scenarios are evaluated. The expectation is clear - if we understand the problem well enough, the right decision will follow.
This logic works in environments where cause and effect are stable.
But not all decision environments are like that.
And this is where the problem begins.
Most organizations do not struggle with decision-making.
They struggle with recognizing what kind of decision they are actually facing.
The Illusion of Analytical Control
A central assumption in classical decision theory is that uncertainty can be reduced through information.
If we collect enough data, improve our models, and refine our forecasts, we should be able to approximate the future closely enough to make optimal decisions.
This assumption holds in environments where uncertainty is a problem of incomplete knowledge.
But as Frank Knight argued over a century ago, not all uncertainty is the same (Knight, 1921).
There is a fundamental difference between:
situations where probabilities can be estimated
and situations where they cannot
Most strategic decisions belong to the second category.
In these situations, more data does not necessarily produce clarity. It often produces competing interpretations.
Yet organizations continue to respond in the same way.
They increase analysis. They add more data.
They refine models that cannot resolve the underlying uncertainty.
What emerges is not control, but the illusion of control.
Complicated Is Not Complex
A critical distinction that organizations consistently underestimate is the difference between complicated and complex systems.
In complicated environments, cause-and-effect relationships exist, even if they require expertise to understand. Engineering problems, financial models, and operational systems fall into this category. With sufficient analysis, they can be understood and managed.
Complex environments operate differently.
In complex systems, cause and effect are not fully visible in advance. They emerge over time through interaction, feedback, and adaptation (Holland, 1992).
This means something very simple - and very uncomfortable:
The future cannot be reliably predicted.
Dave Snowden’s framework captures this distinction clearly. In complicated domains, the correct approach is analysis. In complex domains, the correct approach is experimentation (Snowden & Boone, 2007).
But most organizations do not change their behavior when the context changes.
They apply analysis where experimentation is required.
They seek certainty where only learning is possible.
The Substitution of Uncertainty with Expertise
When organizations encounter uncertainty, they rarely say:
“We do not know.”
Instead, they say:
“Let’s bring in someone who knows.”
This is where expertise becomes a substitute for understanding.
Expertise is powerful. It allows rapid pattern recognition, efficient judgment, and high-quality decisions in stable environments. Gary Klein’s work shows how experts rely on experience to recognize familiar situations and act quickly (Klein, 1998).
But expertise depends on one condition:
The future must resemble the past.
In genuinely novel situations, this condition breaks down.
Patterns no longer repeat. Signals become ambiguous. Familiar solutions stop working.
Yet organizations continue to rely on expertise even when the environment has changed.
What follows is predictable.
Overconfidence increases. Exploration decreases. New situations are interpreted through old models.
And uncertainty is silently replaced with misplaced certainty.
Sensemaking Under Ambiguity
When problems cannot be fully defined in advance, decision-making changes its nature.
It is no longer a process of calculation. It becomes a process of interpretation.
Karl Weick described this as sensemaking - the ongoing effort to construct meaning in ambiguous situations (Weick, 1995).
Sensemaking is not analytical in the traditional sense. It is social.
People interpret events together. They negotiate meaning. They build shared understanding through interaction.
This has a critical implication.
In uncertain environments, decisions do not emerge from data alone.
They emerge from how data is interpreted collectively.
Organizations that rely purely on analysis often miss this entirely.
They invest in information systems, but neglect the processes through which meaning is constructed.
Complexity, Time, and Irreversibility
Another dimension that organizations underestimate is time.
In complex environments, decisions are not isolated events. They are part of evolving systems where early actions shape future possibilities.
Economists describe this as path dependence - the idea that initial decisions constrain future options (Arthur, 1989).
Small actions can produce disproportionately large effects.
Outcomes unfold over time. And decisions cannot always be reversed.
This creates a fundamental challenge.
The quality of a decision cannot always be judged by its immediate result.
A “good” outcome may come from a flawed decision. A “bad” outcome may come from an appropriate one.
Yet organizations continue to evaluate decisions as if cause and effect were immediate and linear.
The Organizational Consequence of Misunderstanding
When organizations fail to distinguish between risk, expertise, and complexity, they behave in predictable ways.
They analyze what cannot be predicted.
They rely on expertise where novelty dominates.
They centralize decisions that require distributed learning.
These responses are not irrational. They are attempts to restore control.
But they produce a deeper structural problem. The organization becomes misaligned with its environment.
Processes slow down situations that require speed.
Hierarchy concentrates decisions that require distribution.
Performance systems reward certainty where uncertainty cannot be eliminated.
Over time, the effects compound.
Decision load increases. Decision gravity intensifies.
And the organization becomes less capable of adapting.
Beyond Classification
The problem is not that organizations lack frameworks.
The problem is that they stop at classification.
Labeling a decision as “strategic” or “complex” does not change how it is actually handled.
What matters is whether the organization can respond differently to different contexts.
Whether it can recognize when analysis is sufficient.
When expertise is relevant. And when uncertainty requires experimentation.
This is not a question of intelligence. It is a question of design.
The problem is not that organizations make poor decisions.
The problem is that they systematically misunderstand the nature of the situations in which decisions are made.
Understanding this distinction leads to a deeper question.
If different decision contexts require fundamentally different approaches, how should organizations design systems that can respond to each of them appropriately?
Answering that question requires moving beyond decision-making as an activity toward decision-making as a system.
References
Arthur, W. B. (1989). Competing technologies, increasing returns, and lock-in by historical events. The Economic Journal, 99(394).
Holland, J. H. (1992). Adaptation in Natural and Artificial Systems. MIT Press.
Klein, G. (1998). Sources of Power: How People Make Decisions. MIT Press.
Knight, F. H. (1921). Risk, Uncertainty and Profit. Houghton Mifflin.
Snowden, D. J., & Boone, M. (2007). A Leader’s Framework for Decision Making. Harvard Business Review.
Weick, K. E. (1995). Sensemaking in Organizations. Sage Publications.