Sensitivity Analysis

Decision Analysis

•

Medium

Sensitivity analysis tests how much a conclusion depends on its assumptions. It helps you see whether your decision is robust or whether one uncertain input is carrying too much of the result.

Reasoning type

Model robustness analysis

Certainty level

Conditional on assumptions

Cognitive load

Medium

Formality

Medium to High

Core Idea

Definition

Sensitivity analysis varies key assumptions or parameters to see how strongly the final conclusion changes.

In Plain English

If a small change in one assumption flips the answer, that assumption matters more than your neat final number.

Framework Structure

Components

Decision Model

Key Assumptions

Parameter Variations

Conclusion Stability

Flow

Build baseline view -> Identify assumptions -> Vary them one by one or together -> Observe how the conclusion changes

How to Apply

1.State the baseline assumptions behind the decision clearly
2.Identify the few variables most likely to affect the result
3.Vary those assumptions across plausible ranges
4.Notice which inputs materially change the conclusion and which barely matter
5.Use the result to focus research, add buffers, or change the decision

When to Use

•Financial models and forecasts
•Product, hiring, or strategic decisions with uncertain inputs
•Any plan that depends heavily on assumptions
•Stress-testing models before acting on them
•Finding which variable deserves the most attention or validation

When NOT to Use

•When the model is so poor that refining inputs does not rescue it
•When the ranges used are arbitrary and ungrounded
•When the decision is too simple to justify the extra modeling
•When people are using the exercise to appear rigorous without improving judgment

Example

Problem

A team is deciding whether a new product line is financially attractive.

Application

1.Build a baseline model using assumptions for demand, price, churn, and support cost
2.Vary each key assumption across plausible ranges
3.Discover that profitability depends heavily on retention but only modestly on launch timing
4.Shift effort toward validating retention assumptions before committing

Conclusion

The team learns which variables are decision-critical and where false confidence would be most dangerous.

Takeaway

Sensitivity analysis tells you whether your answer is sturdy or only true under one narrow set of assumptions.

Common Mistakes

•Varying trivial inputs instead of the ones that actually drive the outcome
•Using unrealistic parameter ranges
•Testing variables independently when they are strongly linked
•Treating the baseline as truth and the sensitivity run as optional
•Ignoring what the analysis reveals about fragility

How to Practice

top three inputs

For each meaningful model, identify the three assumptions most likely to move the result.

range not point

Replace single-number assumptions with plausible low, base, and high ranges.

fragility flag

If a small assumption change flips the decision, mark the model as fragile and investigate before acting.

Related Cognitive Biases

anchoring

People often become attached to one baseline assumption and fail to explore how much it matters.

overconfidence

A clean model can feel more robust than it really is unless assumptions are varied.

planning fallacy

Teams often understate uncertainty in timelines and inputs until forced to test ranges.

Related Frameworks

Scenario Analysis

Monte Carlo Simulation

Value of Information

Decision Tree Analysis

Related Skills

option evaluation

risk identification

confidence estimation

prioritizing factors

Variants & Extensions

One-way sensitivity analysis

Multi-variable stress testing

Break-even analysis

Tornado chart reasoning

Typical Failure Modes

•Bad ranges
•Ignored variable dependencies
•Cosmetic rigor