Sensitivity analysis in project management is the quantitative technique for identifying which uncertain input variables have the greatest influence on project outcomes — enabling project managers to focus risk management attention and contingency planning on the variables that matter most. While Monte Carlo simulation shows the range of possible project outcomes, sensitivity analysis answers the more specific question: which individual uncertain variables are most responsible for that outcome variability? This distinction makes sensitivity analysis particularly valuable for risk prioritisation — it points directly to where additional information gathering, risk mitigation investment, or contingency reserve allocation will have the highest leverage on reducing project outcome uncertainty.
What Is Sensitivity Analysis?
Sensitivity analysis systematically examines how much a project’s output (total cost, completion date, NPV, or other key metric) changes when each uncertain input variable is varied, while holding all other inputs constant at their baseline values. By varying each input independently and measuring the resulting output change, sensitivity analysis ranks inputs by their leverage — the degree to which uncertainty in that input translates into uncertainty in the output.
The result is a sensitivity ranking: input Variable A causes a 25% change in project cost when varied ±20%; Variable B causes a 12% change; Variable C causes a 3% change. Variable A is therefore 2x more sensitive than Variable B and 8x more sensitive than Variable C — making it the highest-priority target for risk mitigation and estimating refinement.
The Tornado Diagram: Sensitivity Analysis Visualised
The tornado diagram is the standard visualisation for sensitivity analysis results. It displays each uncertain input variable as a horizontal bar, with bar width representing the total output variation caused by that variable’s uncertainty range. Variables are ordered from the widest bar at the top to the narrowest at the bottom — producing the characteristic inverted pyramid shape that gives the chart its name. The tornado diagram immediately answers the question “which variables matter most?” in a format that is immediately interpretable by both technical and non-technical audiences.
Reading a tornado diagram: the widest bar at the top represents the variable with the highest sensitivity — the one where uncertainty has the most impact on the project outcome. If this variable were perfectly known, it would reduce total outcome uncertainty more than improving knowledge of any other variable. This makes the tornado diagram both an analysis output and a decision support tool: it shows where to invest in better estimates, more detailed scope definition, or additional risk mitigation effort to achieve the greatest reduction in outcome uncertainty.
One-Way vs Two-Way Sensitivity Analysis
One-way sensitivity analysis varies one input at a time while holding all others constant — the simplest and most common form. It answers: how sensitive is the output to each individual input? Two-way sensitivity analysis varies two inputs simultaneously, producing a matrix of output values that reveals how the combination of two variables interacts to affect the outcome. Two-way analysis is useful for identifying variable pairs with significant interaction effects — where the combined uncertainty of both variables together is greater or lesser than the sum of their individual sensitivities would suggest.
Sensitivity Analysis in Practice: A Cost Model Example
A project manager is building a cost model for a software development project with five major uncertain variables: developer day rate, team size, project duration, rework rate, and infrastructure cost. Baseline assumptions are established for each variable, and each is varied ±20% independently. The results show: developer day rate variation causes a 28% change in total cost (widest bar — top of tornado); team size causes 22%; project duration causes 18%; rework rate causes 9%; infrastructure cost causes 3% (narrowest bar — bottom of tornado). The sensitivity analysis tells the PM that developer day rate and team size are by far the most important variables to estimate precisely and manage closely — together they account for 50% of cost variability. Infrastructure cost is relatively unimportant to refine further — its uncertainty has negligible impact on total cost outcomes.
“Sensitivity analysis tells you where to look. It is the intelligence report that focuses your risk management resources where they will have the highest return.” — David Hillson, Risk Doctor Partnership
Sensitivity Analysis vs Monte Carlo Simulation
Sensitivity analysis and Monte Carlo simulation are complementary tools that project managers should use together rather than choosing between them. Monte Carlo simulation takes all uncertain variables simultaneously and produces a probability distribution of total project outcomes — answering “what range of outcomes is possible and how likely is each?” Sensitivity analysis takes each variable independently and ranks their influence — answering “which variables are most responsible for the outcome uncertainty the Monte Carlo simulation reveals?” The typical workflow is: run Monte Carlo to understand the overall outcome distribution and risk level, then run sensitivity analysis to identify which variables are driving that distribution, then focus risk management efforts on those high-sensitivity variables.
Sensitivity Analysis Applications
| Application | Input Variables | Output Metric | Decision Supported |
|---|---|---|---|
| Cost estimation | Day rates, team size, duration | Total project cost | Contingency allocation |
| Schedule risk | Task durations, dependencies | Completion date | Schedule risk mitigation priority |
| Business case | Benefits realisation rate, costs | NPV / ROI | Investment decision robustness |
| Vendor selection | Delivery risk, price, capability | Total cost of ownership | Which criteria to weight most |
Key Takeaways
- Sensitivity analysis identifies which uncertain input variables have the greatest influence on project outcomes — enabling focused risk management on the variables that matter most.
- The tornado diagram is the standard sensitivity analysis visualisation — widest bar at top = highest sensitivity = highest-priority target for risk mitigation and estimation refinement.
- One-way analysis varies one input at a time; two-way analysis varies two simultaneously — use one-way for initial prioritisation and two-way when interaction effects between key variables are suspected.
- Sensitivity analysis answers “which variables matter most?”; Monte Carlo simulation answers “what range of outcomes is possible?” — use both together for comprehensive quantitative risk analysis.
- Apply sensitivity analysis to cost models, schedule risk, business case NPV, and vendor selection scoring to identify where estimating refinement and risk mitigation will produce the highest return.
- The practical output of sensitivity analysis is a clear prioritisation of where to invest limited risk management resources — it converts a vague sense of uncertainty into a specific, ranked action list.
