Digital Transformation
Why one forecast is not enough for Oil & Gas finance
Why scenario planning matters more than a single forecast in volatile markets
One forecast shows the outcome of one set of assumptions. Oil and gas finance teams usually need Oracle EPM scenario planning because price, production, cost and timing can move materially. Base, upside, downside and stress scenarios help management see the range of possible outcomes and identify which assumptions have the greatest effect on cash flow and returns.
A forecast is not a fact. It is the financial consequence of a set of assumptions.
That distinction matters in oil and gas because several of the most important assumptions are uncertain by design. Commodity prices cannot be known with precision years in advance. Production curves change. Costs move. Projects are rescheduled.
A single forecast can therefore answer only one question: what happens if this exact combination of assumptions is broadly correct? Scenario planning asks a better question: what happens across a credible range of outcomes?
What are the four useful scenario types?
- Base case – the organisation’s current central planning view
- Upside – a defined set of more favourable assumptions, not simply “everything improves”
- Downside – a defined adverse scenario used to understand resilience
- Stress case – a more severe scenario intended to test liquidity, funding, covenants, project economics or another material decision threshold
The labels matter less than the governance. Each scenario should have a clear purpose and documented assumptions.
Why should scenarios be driver-led, not duplicated workbooks?
A scenario becomes difficult to manage when finance duplicates the entire forecast and changes cells manually. Instead, scenario design should focus on the drivers that genuinely change the outcome: price, volume, cost, CAPEX, timing, FX, tax and other business-specific assumptions.
At Wood Mackenzie, GKEPM and Oracle discussed changing price and volume assumptions by plus or minus 10% as a simple example of sensitivity analysis, not a recommendation that every business should use 10%, but a useful illustration of the principle: change a controlled input and observe the effect downstream.
How do you avoid ‘scenario theatre’?
Running five scenarios is not automatically better than running two. If the scenarios do not correspond to real decisions, they create work without insight. Finance should be clear about what management is trying to learn:
- Can the business absorb a lower-price environment?
- What happens if a major project moves?
- Which asset drives downside cash-flow exposure?
- How much headroom exists under a stress case?
- Which assumptions explain most of the variance between cases?
The model should serve those questions.
Featured content
What is enterprise performance management
The best way to kickstart a digital transformation
Share this post
Why is scenario capability an architecture issue?
If every new scenario requires copying a workbook, changing formulas and rebuilding reports, the limitation is not the creativity of the FP&A team — it is the structure of the model. A governed planning environment should separate scenario assumptions from core calculation logic, preserve versions and make comparisons straightforward.
Oracle EPM includes scenario-modelling and predictive capabilities, and GKEPM’s NEO Planning case explicitly includes versioning and what-if analysis as part of the implementation.
Four scenario types and what they are for
| Scenario | Purpose | Example driver changes | Decision supported |
|---|---|---|---|
| Base | Central planning view | Approved current assumptions | Operating plan |
| Upside | Test favourable case | Higher price/volume, improved timing | Investment / capacity |
| Downside | Test resilience | Lower price/volume, higher cost | Cost and cash response |
| Stress | Test severe exposure | Multiple adverse drivers | Liquidity / funding / risk |
In conclusion:
The instinct to chase a single “right” forecast is understandable, it’s simpler to present and easier to explain. But in a market where price, production and cost can each move independently, one number tells management what happens under one combination of assumptions and nothing about what happens if that combination is wrong. Scenario planning isn’t about generating more numbers; it’s about making the range of plausible outcomes visible before the market forces the issue.
The organisations that do this well aren’t the ones running the most scenarios, they’re the ones where producing a new scenario takes minutes, not a manual rebuild, because the architecture separates governed assumptions from core calculation logic. If a downside case still means copying a workbook and changing cells by hand, that’s usually the real constraint worth addressing before adding more forecasting complexity on top.
Frequently Asked Questions
How many scenarios should an oil and gas company model in Oracle EPM? There’s no fixed number -most mature finance teams run a base, upside, downside and occasionally a stress case, but the right number depends on which real decisions the scenarios need to inform. Adding scenarios that don’t map to a decision creates work without insight.
What’s the difference between a base case and a downside case? A base case reflects the organisation’s approved, central planning assumptions. A downside case applies a defined, documented set of adverse changes — such as lower price or production, to test resilience and cash response, rather than simply making every number worse.
Does Oracle EPM support scenario and what-if modelling out of the box? Yes. Oracle EPM includes scenario-modelling and predictive capabilities, and implementations like GKEPM’s NEO Planning case explicitly built in versioning and what-if analysis as part of the core solution.
What’s the fastest way to tell if our scenario process needs an architecture review rather than more analyst time? Time how long it takes to produce a genuinely new scenario. If it means duplicating a workbook, rewriting formulas and rebuilding reports rather than changing a governed set of driver assumptions, the bottleneck is structural, no amount of additional analyst effort fixes that on its own.
The goal is not just better forecasting
It’s eliminating the common EPM failure points that lead to fragmented models, manual scenario building, and low confidence in the numbers.
Why most finance teams aren’t AI ready
Simplifying data transfers
