Sensitivity Analysis Definition / Meaning

Sensitivity Analysis is a quantitative risk assessment technique used extensively in petroleum project management, economics, and regulatory decision-making. It systematically examines how the uncertainty in the output of a model or system can be apportioned to different sources of uncertainty in its inputs. In simpler terms, it answers the question: “If one key assumption changes, how much does the project’s value or outcome change?” This process helps identify which variables have the most influence on a project’s economic viability, such as Net Present Value (NPV), Internal Rate of Return (IRR), or breakeven price.

Core Methodology in Petroleum Economics

In practice, sensitivity analysis is performed by holding all variables constant at their base-case levels except for one specific input, which is then varied over a realistic range (e.g., ±10%, ±20%, or using P10, P50, P90 estimates). The resulting change in the chosen key performance indicator (KPI) is recorded. A typical sensitivity analysis for an upstream oil and gas project might include the following variables:

Usage Example: During a project gate review, the economics team presented a sensitivity analysis showing that a 10% drop in oil price would reduce the project’s Net Present Value by $22MM, triggering a requirement for downside risk mitigation measures such as hedging or securing lower-cost supply contracts.

Types of Sensitivity Analysis

Two primary types are used in the petroleum industry:

• One-at-a-Time (OAT) Sensitivity: The most common method in early-stage project evaluation. It changes one input variable while keeping all others at their base-case values. It is simple to understand and communicate to stakeholders but does not capture interactions between variables.
• Global Sensitivity Analysis (GSA): More advanced technique (e.g., using Monte Carlo simulation or Sobol indices) that varies all input parameters simultaneously across their probability distributions. This reveals the relative importance of each variable and accounts for synergies or dependencies, such as how oil prices and operating costs might move together.

Regulatory and Decision-Making Context

Regulatory agencies, such as the Bureau of Ocean Energy Management (BOEM) or the UK’s Oil and Gas Authority (OGA), often require sensitivity analysis as part of a Field Development Plan (FDP) or an Environmental Impact Statement (EIS). It demonstrates that the operator has considered a range of economic outcomes and has a robust plan for periods of low commodity prices or higher costs. Additionally, banks and project financiers use sensitivity analysis to gauge loan repayment risk under different scenarios.

Best Practices for Oil & Gas Professionals

• Always define the base case clearly using the most likely or P50 estimates from reservoir, technical, and commercial teams.
• Use a tornado diagram to visually present results. The width of each bar shows the magnitude of impact on the KPI (e.g., NPV) when that variable is changed from low to high. This is a standard deliverable in Joint Venture (JV) partner meetings.
• Include at least three key economics drivers: commodity price, production rate, and total project costs (CAPEX + OPEX). For global projects, also consider exchange rates and fiscal terms (e.g., royalty rates or tax holidays).
• Never rely solely on sensitivity analysis for risk management. It is a screening tool. For full probabilistic risk assessment, combine sensitivity analysis with Monte Carlo simulation.

Limitations

Standard sensitivity analysis (OAT) assumes input variables are independent, which is rarely true in petroleum economics. For example, a drop in oil prices often correlates with lower drilling costs and service company rates. More importantly, it does not provide a probability distribution of outcomes; it only shows the magnitude of impact if a change occurs. These limitations are why experienced project managers pair sensitivity analysis with scenario analysis (e.g., optimistic, pessimistic, most likely) and decision tree analysis.