Tag: Decision Making

A leadership team asks for a strategy. What it really wants is a guarantee.

The executives want to know which campaign will produce the forecasted revenue, which technology will create the promised efficiency, which market will grow, and which reorganization will solve the problem without creating another one. The strategist can provide research, models, experience, and a recommendation. What the room wants is certainty.

So everyone participates in a familiar performance. Assumptions become projections. Projections become targets. Targets become commitments. The presentation grows more precise as the underlying situation remains stubbornly unpredictable.

If the plan works, its authors were prescient. If it fails, they can point to the data, the methodology, the consultant, or the industry standard. Every decision was defensible. No one quite decided.

This is not strategy. It is responsibility laundering.

In his episode of The Unfolding Thought Podcast, strategist Steve Kozel describes strategy as a series of nested choices. At any altitude, he argues, strategy requires multiple feasible options, an understanding of the situation, a desired effect, and a choice among the alternatives.

That definition carries a consequence organizations often try to avoid: if the alternatives are genuinely feasible and the future is genuinely uncertain, the choice can be wrong.

A strategy that cannot fail is not a strategy. It is either an operating procedure, a foregone conclusion, or a story told to make uncertainty feel controllable.

Strategy begins where prediction ends

There are decisions for which certainty is a reasonable expectation. Payroll should run. An invoice should calculate correctly. A proven manufacturing process should stay within tolerance. When cause and effect are stable and repeatable, leaders should demand reliability.

Strategy is different because it concerns choices whose value depends on reactions that have not happened yet. Competitors respond. Customers reinterpret what they want. Employees change the plan while implementing it. Technology alters the economics. The choice itself changes the environment in which its success will be judged.

Frank Knight drew a useful distinction in Risk, Uncertainty, and Profit. Risk describes situations in which probabilities can be estimated. Uncertainty describes situations in which the probabilities themselves are not reliably knowable.

Most organizations are comfortable managing risk. They create ranges, reserves, scenarios, and controls. They struggle when uncertainty cannot be converted into a percentage. The absence of a reliable probability feels like the absence of management.

But uncertainty is not a temporary defect in the strategy process. It is the condition that makes strategy necessary. If the correct action and its outcome were already known, the organization would not need a strategist. It would need an operator.

Certainty has a job inside the organization

The demand for certainty is not simply a reasoning error. It serves an organizational purpose.

A confident forecast lets a project receive funding. A familiar methodology reassures procurement. A famous consultancy gives an executive cover. A dashboard suggests control. A best practice allows everyone to say the decision met the accepted standard.

These things reduce personal exposure even when they do not reduce uncertainty. That difference matters.

Kozel calls the deeper pattern “Fear OS,” an unspoken operating system in which people are expected to maintain control, avoid failure, produce positive metrics, and explain the world as though it were more rational than it is. Under those conditions, the safest decision is often the one whose failure will be easiest to defend.

This helps explain why a company can be filled with intelligent people and still repeat choices no one strongly believes in. Each participant is responding rationally to a local incentive. The analyst avoids an unsupported recommendation. The manager protects the quarterly number. The executive selects the respectable vendor. The board receives a forecast with enough detail to feel governed.

The organization has not eliminated uncertainty. It has distributed responsibility so thoroughly that the eventual result will seem to belong to no one.

Best practice cannot create strategic advantage

Best practices are valuable when the goal is dependable execution of a known activity. I want the people operating an airplane, administering medication, or securing financial data to use procedures that have survived serious scrutiny.

The problem begins when leaders ask best practice to answer a strategic question.

Michael Porter’s explanation of strategy centers on a unique position, trade-offs, and fit among activities. Strategy requires choosing what not to do. An organization that copies the accepted practices of its category may become more competent. It does not become meaningfully different.

This is the contradiction Kozel surfaces in the episode. Organizations want a proven solution and an advantage. They want evidence without experiments. They want innovation without the waste, ambiguity, and failed attempts through which new knowledge is produced.

If every competitor follows the same research, hires the same experts, adopts the same technology, and optimizes against the same benchmark, best practice becomes a convergence mechanism. It improves the category while compressing the differences within it.

An experiment is not an excuse

None of this means leaders should become casual about evidence or celebrate failure as a cultural virtue.

“We are experimenting” can become its own form of responsibility laundering. Teams can use the language of learning to excuse poor preparation, unclear goals, oversized bets, and projects that continue long after the original hypothesis has collapsed.

A real experiment is more demanding than a confident plan because it must say what is not known and how the organization intends to learn it.

Amy Edmondson describes an intelligent failure as an undesired result in new territory. The attempt should pursue a meaningful goal, rest on an informed hypothesis, and be no larger than necessary to produce the needed knowledge.

Those constraints turn experimentation from a slogan into a management discipline.

• What specific assumption are we testing?
• What evidence would increase or reduce our confidence?
• What is the smallest credible action that can produce that evidence?
• What downside are we exposing, and who has agreed to it?
• When will we stop, expand, or change direction?
• Where will the learning alter a decision, budget, process, or belief?

The purpose of an experiment is not to make failure harmless. It is to buy information at a price the organization can afford.

Learning must reach the assumptions

Organizations often claim to learn while protecting the belief that produced the disappointing result.

A campaign misses its target, so the team changes the creative. A product struggles, so marketing receives a larger lead goal. A transformation stalls, so managers schedule more training. The response corrects activity without reconsidering the governing assumption.

Chris Argyris distinguished this kind of correction from deeper learning. Single-loop learning asks how to perform the existing approach more effectively. Double-loop learning asks whether the underlying goals, rules, and assumptions still deserve to govern the work.

In his work on defensive reasoning, Argyris argued that capable professionals often respond to difficulty by avoiding examination of their own contribution to it. Expertise can make the defense more articulate without making it less defensive.

This is why adding more data does not necessarily make an organization more empirical. Data can test a claim. It can also decorate a conclusion the organization is unwilling to reconsider.

Double-loop learning creates a more uncomfortable review:

• Did the execution fail, or was the strategic premise wrong?
• Are we improving the offer, or only improving the way we promote it?
• Does the metric represent value, or is it merely easy to make rise?
• Are we asking a team to solve a problem whose cause sits above its authority?
• What would we stop believing if we took this result seriously?

If every review ends with a better way to execute the same idea, the organization is not learning strategically. It is becoming more efficient at defending the past.

Exploitation wins the quarterly argument

Even leaders who understand this problem face a structural disadvantage.

James March’s classic paper on exploration and exploitation in organizational learning describes the tension between refining what an organization already knows and searching for new possibilities. Exploitation produces clearer, nearer, and more predictable returns. Exploration is uncertain, delayed, and frequently disappointing.

That asymmetry gives the known approach an advantage in every budgeting cycle. The established product has revenue. The existing channel has benchmarks. The familiar process has owners. The new option has a hypothesis.

March warned that adaptive processes can become effective in the short run and self-destructive in the long run because organizations refine exploitation faster than exploration. The company becomes increasingly competent at a world that is becoming less relevant.

This is another reason certainty can be dangerous. It does not merely misdescribe the future. It systematically moves resources toward whatever already has a history.

What honest strategy sounds like

An honest strategy does not replace confidence with vagueness. It makes confidence proportional to evidence.

It names the choice. It explains why this option is preferable to other feasible options. It distinguishes what is known from what is assumed. It defines the risk the organization is accepting. It makes clear who owns the decision. It identifies the earliest evidence that would justify changing course.

It also separates commitments from predictions.

A leader can commit to allocating resources, protecting an experiment, reviewing evidence, and changing direction when the premise no longer holds. A leader cannot honestly commit that customers, competitors, employees, regulators, and technology will behave according to the forecast.

This kind of strategy can sound less impressive in the room. It replaces false precision with boundaries, hypotheses, and decision rules. It admits that expertise improves a choice without making the future obedient.

That is not weaker leadership. It is leadership willing to remain responsible after the illusion of control has been removed.

The choice comes before the certainty

Organizations do not need more comfort with failure in the abstract. They need a more precise relationship with uncertainty.

Preventable mistakes should be prevented. Known processes should be reliable. Large irreversible choices deserve more scrutiny than small reversible ones. Evidence should shape judgment wherever evidence exists.

But strategy cannot wait until the uncertainty has disappeared. By then, the choice may be obvious, the advantage may be gone, or the organization may have spent years perfecting the wrong thing.

The discipline is to decide without pretending to know, act without becoming reckless, and learn without protecting the assumptions that made the action possible.

Strategy is not a promise that the future will obey. It is a commitment to choose, learn, and remain accountable before certainty arrives.

Listen and read further

• Steve Kozel on The Unfolding Thought Podcast
• Steve Kozel on LinkedIn
• Frank Knight, Risk, Uncertainty, and Profit
• Michael Porter’s explanation of strategy
• Chris Argyris, “Teaching Smart People How to Learn”
• James March, “Exploration and Exploitation in Organizational Learning”
• Amy Edmondson on intelligent failure