Industry insights
Oil & Gas Sensitivity Analysis: Turning price and volume uncertainty into decisions
Sensitivity analysis changes one or more defined assumptions and measures the effect on an outcome. In oil and gas finance, common sensitivities include:
- commodity price
- production volume
- project timing
- OPEX
- CAPEX
The value is not in producing more numbers but rather in identifying which assumptions have the greatest effect on cash flow, profitability or returns.
When management asks, “What happens if price is 10% lower?”, finance should not need to rebuild the plan.
That question is a classic sensitivity test: hold the core model together, change a defined assumption, and observe the effect.
At the Wood Mackenzie fireside discussion, GK EPM and Oracle used a deliberately simple example: 10% lower or higher price and 10% lower or higher volume.
That example is useful because it shows how quickly uncertainty becomes multidimensional.
If price has three states, lower, base, higher – and volume has three states, finance already has nine combinations to examine before changing cost, timing or fiscal assumptions.
There is no magic number of spreadsheets at which finance must implement an enterprise planning platform. A single badly controlled workbook can be risky. A large number of well-designed analysis files can be perfectly manageable.
The useful indicators are behavioural.
Start with the decision, not the percentage
A sensitivity should answer a management question.
The question might be:
- Which driver has the largest effect on asset cash flow?
- At what price does a project stop meeting a return threshold?
- How much production underperformance can the plan absorb?
- What happens if CAPEX moves by a defined amount?
- Which assumptions create the greatest downside exposure?
The percentage change is only a modelling device.
Using 10% because it is easy to communicate can be useful for an illustrative test. For a real decision, the sensitivity range should be based on the organisation’s risk view, historical volatility, management assumptions or another defensible basis.
Featured content
What is enterprise performance management
The best way to kickstart a digital transformation
Share this post
Separate sensitivity from scenario planning
The two terms are often used interchangeably, but they answer slightly different questions.
A sensitivity usually isolates the effect of changing one driver, or a small set of drivers, while holding other assumptions constant.
A scenario normally represents a coherent future in which several related assumptions change together.
For example:
- Price -10% is a sensitivity.
- A downside case combining lower price, lower production, delayed project timing and higher cost is a scenario.
Both are useful. The planning model should support both without requiring duplicate calculations.
The financial impact should flow through the model
Sensitivity analysis becomes less useful if finance has to calculate the impact separately in several places.
A production change should feed revenue. Revenue and cost changes should affect P&L and cash. Project timing should affect CAPEX phasing. Depending on the model, tax, funding or other outputs may also change.
That is why sensitivity capability is partly an architecture question.
Our NEO Next Planning case includes versioning and what-if analysis as part of the Oracle EPM solution, alongside integrated quantity and financial forecasting.
Focus on the response, not only the output
The best sensitivity analysis does more than show that cash flow falls.
It should help management understand what action becomes relevant.
For example:
Sensitivity | Financial result | Management question |
Price -10% | Lower revenue/cash | Do we defer discretionary spend? |
Volume -10% | Lower revenue, potential unit-cost pressure | Which asset drives the gap? |
CAPEX +10% | Higher cash requirement | Does funding headroom remain adequate? |
First production +6 months | Delayed cash inflow | Does project timing change return or liquidity decisions? |
That moves scenario work from finance modelling into decision support.
Build a sensitivity library
Finance teams do not need to reinvent the analysis every time management asks a question.
A mature planning process can maintain a defined library of common tests, with clear owners and ranges.
Examples might include:
- commodity-price shifts;
- volume shortfalls;
- project delays;
- cost inflation;
- FX movements;
- fiscal changes.
The library should not become a rigid list. It is a starting point that allows finance to respond quickly while retaining governance.
Table: Sensitivity matrix
Volume / Price | Price -10% | Base price | Price +10% |
Volume -10% | Case 1 | Case 2 | Case 3 |
Base volume | Case 4 | Base | Case 5 |
Volume +10% | Case 6 | Case 7 | Case 8 |
In conclusion:
Sensitivity analysis is simple in principle: change a defined assumption and see what happens. The difficulty is practical. If every “what if price is 10% lower?” question means copying a model, re-keying numbers and rebuilding outputs, management gets its answer too late to influence the decision it was meant to inform.
The most useful sensitivity work starts with the management question rather than the percentage, keeps sensitivities and scenarios distinct, and lets a change in one driver flow through revenue, cost, cash and returns automatically. When that is true, the nine-case matrix in this article takes minutes rather than days, and finance spends its time interpreting results instead of producing them.
What leading finance teams do
- Start with the decision, not the percentage. They define what management needs to learn, such as the price at which a project misses its return threshold, and choose sensitivity ranges that reflect real volatility rather than a convenient round number.
- Let one change flow through the whole model. A change in price, volume, CAPEX or timing updates revenue, cost, cash and returns automatically, with no recalculating in several places.
- Keep a governed sensitivity library. Common tests such as price shifts, volume shortfalls, project delays and cost inflation have clear owners and ranges, so finance can answer “what if?” quickly without losing control.
Frequently asked questions:
What’s a typical sensitivity range to use for oil price in a financial model? There’s no universal figure. ±10% is a commonly used illustrative range for demonstrating the method, but a defensible range for real decisions should be based on the organisation’s historical price volatility, risk appetite and management view, not a round number chosen for convenience.
Can Oracle EPM automate sensitivity analysis across scenarios? Yes. When sensitivity capability is built into the planning architecture, as in GKEPM’s NEO Planning implementation, which includes versioning and what-if analysis, a change to a driver like price or volume can flow through revenue, cost and cash automatically rather than requiring finance to recalculate manually in each output.
How many variables should a sensitivity test at once? A sensitivity typically isolates one or two variables at a time, such as price and volume, while holding everything else constant. Testing many variables together moves into scenario planning territory, where a coherent combination of changes represents a plausible future rather than isolating a single driver’s effect.
What’s the difference between sensitivity analysis and what-if analysis? In practice the terms overlap. Both change an assumption and observe the effect. Sensitivity analysis is usually more systematic, testing defined ranges across key drivers to see which matter most, while what-if analysis is often a quicker, one-off question from management.
Thinking about your current EPM setup?
If your team can produce a base forecast but every sensitivity requires manual rework, the constraint may sit in the model rather than the people.
Why most finance teams aren’t AI ready
Simplifying data transfers
