What is Sensitivity Analysis?

Fahad Usmani, PMP September 9, 2024

Sensitivity analysis can help you determine how changes in one variable will affect a system’s output. It involves changing one input while
keeping others constant to observe the impact on the results.

This analysis lets you identify the most influential variables, providing insights into the model’s behavior and the robustness of its predictions.

In project management, sensitivity analysis can help assess how changes in project variables (e.g., costs, timelines, or resources) will affect the
project objectives. You will alter one variable while keeping others constant to determine which factors will most significantly affect the
project or its outcome.

Using this method, you can identify risks and uncertainties and develop their response plans.

For example, sensitivity analysis can show how delays in specific tasks can influence the schedule baseline or how fluctuations in any resource
cost could affect the cost baseline.

By understanding these relationships, you can prioritize your attention on high-impact areas and make more informed decisions. Sensitivity
analysis enhances project control by providing insights into variables that require careful monitoring.

How to Use Sensitivity Analysis

You have two approaches to using sensitivity analysis. These approaches are:

Direct Approach
In a direct approach, you change one or more input variables to observe the corresponding changes in the output. This method is
straightforward, as it directly manipulates the inputs to see how sensitive the system is to those changes.

Common techniques used in the direct approach include one-at-a-time (OAT) analysis, where each variable is changed individually while
others are held constant. The direct approach provides clear, easily interpretable results, making it useful for identifying the most influential
factors in a model.

Indirect Approach
The indirect approach uses mathematical or statistical techniques to analyze sensitivity without explicitly altering the input variables. These
methods include regression analysis, variance-based sensitivity analysis, or factorial design.

The indirect approach involves more complex calculations and is used when the model is too complicated for straightforward input
manipulation. It provides insight into the relationships between inputs and outputs, thus identifying the main effects and interactions
between variables.

This method is useful in models with many variables or where direct manipulation is impractical.

Step-by-Step Guide to Conducting Sensitivity Analysis

1. Define the Objective: Clearly outline the goals of the sensitivity analysis. Determine which outputs or results you want to assess in
response to changes in input variables.

2. Identify the Key Variables: List all input variables that may affect your model’s outcome. Prioritize them based on their relevance and
potential impact.

3. Develop the Model: Create or use an existing model that relates input variables to the output. Ensure that the model accurately reflects
the relationships between variables.

4. Select the Analysis Method: Choose a sensitivity analysis method (e.g., one-at-a-time [OAT], scenario analysis, or Monte Carlo
Simulation) based on your requirements and the available data.
 5. Set the Input Range: Define the range or scenarios for each input variable. For example, you can change an input by ±10% or use
different predefined scenarios for a more comprehensive analysis.

6. Conduct the Analysis: Apply the chosen method to vary the input variables. Adjust one variable at a time for direct methods, while
keeping others constant. For indirect methods or simulations, use statistical techniques to assess the impact of variations across multiple
variables.

7. Analyze the Results: Review the output data to determine how input changes affect the results. Use graphical tools (e.g., Tornado or
Spider Diagrams) to visualize the model’s sensitivity.

8. Interpret the Findings: Identify which variables have the most significant impact on the output. Assess the robustness of your model
and understand which factors are critical to achieving the desired outcomes.

9. Develop Recommendations: Based on the analysis, form recommendations to mitigate risks, optimize performance, and/or make
informed decisions. Consider contingency plans for variables with high sensitivity.

Sensitivity Analysis Example

A construction company is planning a building project. The project manager is concerned about the volatility of steel prices, a major
component of the construction costs. To perform a sensitivity analysis, the project team changes the cost of steel within a reasonable range
(e.g., ±20%) while keeping other variables constant. They will then observe how this change affects the total project cost.

Sensitivity analysis shows that a 10% increase in steel prices will lead to a 5% increase in the total project cost, thus significantly impacting the
project’s budget. This insight will allow the project manager to develop contingency plans (e.g., securing fixed-price contracts for steel or
adjusting the project’s scope to stay within budget).

Sensitivity Analysis Techniques

The two popular sensitivity techniques are:

Tornado Diagram
A Tornado Diagram is a graphical representation used in sensitivity analysis to display the relative importance of different variables on a
model’s outcome. It arranges the variables in descending order of their impact on the output, with the most influential variable at the top and
the least influential at the bottom.

The diagram resembles a tornado, with bars extending to the left and right, indicating the range of possible outcomes as each variable varies.
This visual tool helps you quickly identify which factors most affect the results, and then you can prioritize areas for further analysis or
intervention.

Monte Carlo Simulation

Monte Carlo Simulation is a powerful technique used in sensitivity analysis to assess the impact of uncertainty in input variables on a model’s
output. It involves running the model thousands or millions of times, using randomly selected values for the input variables based on their
probability distributions. This process generates a range of possible outcomes, which can then be analyzed to determine the likelihood of
different scenarios and to understand the overall risk and uncertainty in the model.

Monte Carlo Simulation is widely used in finance, engineering, and project management to provide a more comprehensive view of potential
outcomes and to support robust decision-making under uncertainty.

Sensitivity Analysis Pros and Cons

Pros
Identifies Key Drivers: Sensitivity analysis can help identify which variables will significantly impact the model’s output, thus allowing
decision-makers to focus on critical factors.

Enhances Decision-Making: Sensitivity analysis enhances decision-making by understanding how input changes affect outcomes.

Improves Risk Management: Sensitivity analysis can help assess risks by revealing a model or project’s vulnerabilities to changes in key
inputs.

Improves Model Robustness: Sensitivity analysis can highlight areas where a model may need refinement, thus leading to more reliable,
accurate predictions.
 Increases Versatility: Sensitivity analysis can be applied across various fields (e.g., finance, engineering, project management, and
environmental science).

Highlights Areas for Improvement: Sensitivity analysis aids decision-makers in determining where they can make modifications.

Cons
More Time-Consuming: Conducting sensitivity analysis—especially with complex models or numerous variables—can be time-intensive
and resource-heavy.

Requires Assumptions: The results depend heavily on the assumptions made during the analysis, which may not always accurately
reflect real-world conditions.

Its Scope Is Limited: If only a few variables are analyzed, sensitivity analysis might miss other important factors or interactions between
variables.

Leads to Oversimplification: Simplistic approaches (e.g., one-at-a-time analysis) might overlook interactions between variables, thus
leading to incomplete insights.

Leads to Complexity with Multiple Variables: When dealing with many variables, sensitivity analysis can become complicated, and
interpreting results can become challenging.

Creates Variable Dependency: Since variables often depend on each other, it is impossible to analyze them separately. For instance,
changes to the selling price will result in changes to the sales volume.

Sensitivity Analysis Vs Scenario Analysis

Scenario analysis and sensitivity analysis are both methods used to assess the impact of variable changes, but they differ in their approach
and scope. Sensitivity analysis examines how variations in individual input variables can affect a model’s output by altering one variable at a
time while keeping others constant, thus allowing you to identify the most influential variables.

Scenario analysis evaluates the effects of multiple variables that change simultaneously based on predefined scenarios. It involves creating
and analyzing different “what-if” scenarios (e.g., best-case, worst-case, and most likely scenarios) to understand how combinations of changes
will impact the outcome.

While sensitivity analysis focuses on the impact of single variables, scenario analysis provides a broader view by exploring future states and
their combined effects. This makes scenario analysis more comprehensive for understanding complex, interrelated uncertainties.

Conclusion
Sensitivity analysis is a useful tool for understanding the impact of variable changes on the outcomes of models and projects. Identifying the
most influential factors enables decision-makers to focus on critical areas, manage risks, and optimize performance.

Sensitivity analysis enhances the reliability of predictions and supports informed, strategic decision-making. As uncertainties and complexities
grow, anticipating and responding to changes through sensitivity analysis remains essential for success.

Further Readings:

What is a Tornado Diagram?

What is Monte Carlo Analysis?
What is Scenario Analysis?
What is What-if Scenario Analysis?
What is 5 Why Analysis?

This topic is important from a PMP exam point of view.

Fahad Usmani, PMP

I am Mohammad Fahad Usmani, B.E. PMP, PMI-RMP. I have been blogging on project management topics since
2011. To date, thousands of professionals have passed the PMP exam using my resources.