A thoroughly enjoyable and thought-provoking evening at yesterday’s SPE London Distinguished Lecture, “Beyond the Hype: Transforming Subsurface & Production with AI Innovation.”
Shripad Biniwale of SLB, recently relocated to London after several years in Abu Dhabi, brought a particularly valuable perspective: practical experience of applying AI to real subsurface and production challenges.
Across an excellent hour-long lecture, Shripad explored applications ranging from waterflood management and history matching to fast-track field development planning and the emerging potential of multi-agent engineering.
What particularly interested me was not AI in isolation, but the engineering architecture around it.
One of Shripad’s slides captured the principle particularly well:
“The strongest solutions combine data, physics, optimisation and engineering judgement, rather than forcing every problem into one AI model.”
His field development example demonstrated exactly that. Full-physics simulation preserved fidelity. Machine learning proxies accelerated the exploration of uncertainty. Risk analytics helped rank development options. Engineering judgement ultimately determined which opportunities were robust.
That distinction matters.
AI does not remove the need for engineering judgement. It potentially allows engineering judgement to operate across a much richer body of evidence, much faster.
It also prompted me to think about another dimension of that evidential environment: comparative subsurface benchmarking.
AI can become increasingly sophisticated at understanding the reservoir in front of us: its data, physics, uncertainty and possible development responses.
Comparative benchmarking introduces a complementary question:
What can experience elsewhere, under genuinely comparable subsurface and development conditions, tell us about what may be achievable here?
As subsurface workflows become increasingly AI-enabled, and potentially increasingly agentic, establishing the right engineering context, appropriate comparators and defensible reference points may become more important, rather than less.
There is a fascinating intersection here between AI, physics, comparative evidence and engineering judgement, and one that I think deserves further exploration across the profession.
Many thanks to Shripad for an excellent lecture and for the thoroughly enjoyable discussions during the evening, and to SPE London Section for bringing together such a good group of people.
A particular thank you also to Farid Hadiaman, Wells Perfomance Advisor at bp and SPE London Section Treasurer, for the invitation and for encouraging me to come along. I’m very glad I did.
Recovery factor benchmarking helps identify where performance diverges. Analogue benchmarking helps establish what remains achievable and what it may take to get there. This article explores why accumulated analogue intelligence remains one of the most powerful tools available for identifying and extending recovery potential across mature hydrocarbon systems.
Across mature oil and gas portfolios, recovery factor benchmarking is often one of the first tools used to identify where additional value may exist.
It helps answer an important question:
Which assets appear to exhibit the greatest remaining recovery opportunity?
However, recovery factor benchmarking alone rarely provides the answer operators ultimately need.
It may indicate where performance diverges.
It does not necessarily establish what remains achievable, nor what it may take to realise it.
This is where analogue benchmarking becomes essential.
Recovery Factor Does Not Explain Itself
A recovery factor number, by itself, has limited meaning.
Suppose a field has achieved a recovery factor of 35%.
Is that exceptional?
Average?
Or materially below what remains achievable?
The answer depends entirely on context.
Recovery outcomes are shaped by geology, reservoir architecture, drive mechanism, development maturity, pressure support, intervention history, recovery strategy, operating philosophy, infrastructure constraints, and the decisions that have shaped the asset over decades.
The same recovery factor can therefore represent very different outcomes in different reservoirs.
Meaningful recovery factor benchmarking requires meaningful analogues.
Identifying The Gap Is Not The Same As Explaining The Gap
Recovery factor benchmarking and analogue benchmarking serve different purposes.
Recovery factor benchmarking helps identify where performance diverges.
Analogue benchmarking helps establish why that divergence exists, what remains achievable, and what may be required to realise it.
The objective shifts from:
What recovery factor has been achieved?
to:
What remains achievable?
and ultimately:
What would it take to get there?
This is where many benchmarking exercises stop too early.
The most valuable insights rarely emerge from identifying the gap itself.
They emerge from understanding what explains the gap.
The Real Asset Is The Independent Reference Point.
Many organisations possess data.
Many organisations possess technical expertise.
The challenge is rarely a lack of information.
The challenge is maintaining valid comparison across large populations of hydrocarbon systems that have evolved under different geological conditions, recovery mechanisms, development philosophies, infrastructure architectures, operating environments, and maturity states.
The value of analogue benchmarking does not lie in just identifying fields that appear similar.
Its value lies in maintaining valid comparison across systems that differ in important ways while remaining sufficiently comparable to establish what remains achievable.
This requires accumulated analogue intelligence.
Not assembled.
Accumulated.
Developed through long-term observation of hydrocarbon systems across operators, basins, recovery mechanisms, maturity states, and development pathways.
The real asset is the accumulated analogue intelligence behind the benchmark.
Why This Matters
The greatest long-term value losses in sophisticated hydrocarbon systems rarely arise through obvious operational failure.
More often, they emerge when accepted interpretations gradually become embedded within recovery expectations, development plans, infrastructure pathways, investment decisions, and optimisation frameworks.
Over time, these interpretations influence what is considered realistic.
They influence what is pursued.
And ultimately, they influence what is achieved.
The most significant opportunities often emerge where accepted interpretations no longer reflect what remains achievable.
Identifying and Extending What Remains Achievable
Recovery factor benchmarking helps identify where performance diverges.
Analogue benchmarking helps establish what remains achievable and what it may take to get there.
Together they provide one of the most effective mechanisms available for identifying and extending what remains achievable across mature hydrocarbon systems.
Because the objective is not simply to understand performance.
The objective is to extend it.
Benchmarking is the discipline of establishing what world-class looks like under comparable conditions – and understanding how far performance remains from that potential.
It provides an independent reference point for what remains achievable.
As systems evolve across field life, development phases, operating conditions, and changing business priorities, benchmarking helps maintain visibility of what world-class performance continues to look like – and what it may take to achieve it.
Recovery factor comparisons also have implications for the wider development strategy. FDP Benchmarking provides a broader comparative examination of development assumptions, concept choices and alternatives, while distinguishing what external evidence supports from what requires further operator-specific technical appraisal.
AI is rapidly becoming embedded across reservoir management, production systems, and facilities operations.
From subsurface modelling and well placement to process optimisation and predictive maintenance, AI is improving how decisions are made.
Across Middle East NOCs, this is already translating into:
The direction is clear. And the progress is real.
But as AI becomes more capable, a different question begins to emerge.
Not how quickly decisions can be made –
but how those decisions are anchored.
AI learns from data:
But it does not define what “best” looks like.
It does not determine:
Two systems may be optimised using AI – and arrive at very different outcomes:
But neither necessarily reflects best-in-class performance.
This is where a deeper distinction emerges.
Between:
Across reservoirs, facilities, and integrated systems, performance is shaped by context:
AI can optimise within that context.
But it does not establish whether that context itself reflects what is achievable.
This creates a subtle but important gap.
One that is not immediately visible – particularly as systems become more data-rich and more responsive.
Without a consistent basis for comparison:
Over time, this limits how far performance is extended.
Not because capability is lacking –
but because the reference point is not clear.
The challenge, then, is not only to improve performance.
It is to define it.
This is where benchmarking operates.
Not as an alternative to AI – but as a foundational layer.
Providing an independent, consistent, validated basis for understanding performance across systems, over time, and as conditions evolve.
Best-in-class performance does not immediately reveal itself.
It exists – but it is not obvious.
Benchmarking is the discipline of establishing what world-class looks like under comparable conditions – and understanding how far performance remains from that potential.
It provides an independent reference point for what remains achievable.
As systems evolve across field life, development phases, operating conditions, and changing business priorities, benchmarking helps maintain visibility of what world-class performance continues to look like – and what it may take to achieve it.
© Fulcrium 2026. All rights reserved. Proprietary analysis and methods.
Reservoir engineers routinely draw on experience from other fields when evaluating development decisions. Yet observing that another carbonate development has achieved a different recovery or production outcome does not establish what caused that difference, still less whether the result can be replicated. A useful comparison must distinguish the observed outcome from the explanation offered for it.
Consider a population of carbonate developments selected as relevant to a particular performance question. Most occupy a recognisable range; a few perform materially differently. Differences in reservoir architecture, fracture connectivity, pore structure, wettability, fluid properties or aquifer behaviour may explain part of the variation. Development maturity, well configuration, pressure support, injection strategy and operating history may explain more. Production-system constraints can also influence what the reservoir is able to deliver.
If a material difference remains after the principal explanations have been examined, the enquiry changes. The question is no longer simply which operator performs better, but what have we failed to explain?
Comparative evidence can be valuable precisely because it tests whether the explanations currently used to understand reservoir performance are sufficient. An unexplained difference is not, by itself, an improvement opportunity or evidence of superior practice. It is a reason to examine the interpretation more closely.
There is no single population of comparable carbonate reservoirs suitable for every engineering enquiry. A comparison relevant to well productivity may be inappropriate for sweep efficiency. Developments useful for examining pressure depletion may not be the right comparators for waterflood response, infill performance, recovery efficiency or late-life redevelopment.
This distinction matters particularly in carbonates. Reservoirs with apparently similar geological descriptions can behave differently, while reservoirs that are not close geological analogues may nevertheless provide useful evidence about a particular production or development response. Comparability must be judged in relation to the behaviour being investigated, rather than assumed from a broad reservoir classification.
The appropriate system boundary follows the same logic. A waterflood comparison may initially concern fracture connectivity, matrix properties, wettability, aquifer contribution, producer–injector configuration and injection strategy. But a difference in produced-water handling capacity can affect which high-water-cut wells remain on production. Injection capacity or processing limitations may similarly constrain the development choices available to an operator.
In such circumstances, a comparison confined to the reservoir would give an incomplete account of the observed outcome. For some questions the reservoir remains the appropriate boundary; for others, the relevant unit is the reservoir-development system – including wells, development decisions, facilities and operating constraints. The boundary should extend only as far as the performance question requires.
Suppose one carbonate development appears to achieve higher recovery than a relevant comparison population, assessed on a consistent recovery basis and with differences in development maturity taken into account. It may be tempting to attribute that outcome to better reservoir management. Yet higher recovery does not, on its own, establish that interpretation. Reservoir characteristics, development maturity, well density, completion architecture, pressure support, surveillance, intervention history and production constraints may all influence the result.
The observation that an operator used a particular practice and subsequently achieved a favourable outcome establishes an association. It does not establish that the practice caused the outcome.
The engineering task is to assess whether the available evidence supports the proposed explanation. A difference initially associated with operating practice may prove consistent with variations in fracture behaviour, injection configuration or production-system capacity. In other cases, those factors may account for only part of what has been observed.
What remains is a residual difference: a difference not adequately explained given the available evidence, selected comparators and explanations examined so far. It must be interpreted cautiously. It may reflect incomplete data, measurement uncertainty, imperfect comparability, interactions not adequately represented in the interpretation, or relevant conditions that have not been recognised.
An unexplained residual is not evidence, by itself, of superior operator practice or previously unrecognised reservoir behaviour. Its value is that it defines the next technical question.
Sometimes that question concerns a development choice or a difference in surveillance. Sometimes it concerns the interpretation of reservoir behaviour. If apparently relevant developments repeatedly produce outcomes inconsistent with the prevailing explanation, it may be necessary to reconsider the explanation itself – including the assumptions on which the comparison was constructed.
What if the unexplained difference is telling us something about the reservoir rather than the operator?
Cross-operator comparisons are generally based on observational evidence, not controlled experiments. Even where a particular practice is associated with a favourable outcome, other reservoir and development differences may have contributed to that outcome.
Attribution therefore requires disciplined language. The available evidence may support an explanation, make it more plausible or challenge competing explanations without establishing a single cause with certainty. An observed advantage should not be assigned to operator behaviour simply because it remains after the most obvious differences have been considered.
An asset team may understand its own reservoir in considerable detail while having limited visibility of how other developments have responded under different conditions and decisions. Wider comparative evidence can expose observations that are difficult to recognise within one field history. Its contribution is not merely a ranking of operators, but a more demanding test of what the observed differences permit us to infer.
Even where the explanation for a favourable outcome is credible, a separate question remains: would the same practice be appropriate elsewhere?
The effectiveness of an intervention may depend on fracture behaviour, mobility ratio, pressure conditions, well configuration, development maturity, surveillance capability or facility capacity. A practice associated with improved performance in one development may be ineffective – or unsuitable – in another.
Attribution asks why an apparent advantage occurred. Transferability asks whether the conditions associated with that advantage exist elsewhere.
The distinction matters because identifying a credible explanation for another operator’s performance is not the same as establishing a development option for one’s own reservoir. The latter requires judgement about the local reservoir and production system, the proposed intervention and the consequences that can reasonably be expected.
The same discipline applies when comparing the use of AI, advanced surveillance, digital workflows or more sophisticated reservoir models. The fact that another operator adopted a technology establishes little about its engineering consequence.
A more useful enquiry is whether the technology changed what the operator understood or decided. Did it reveal previously unrecognised connectivity, alter infill placement or injection allocation, overturn an existing interpretation, or reduce uncertainty sufficiently to change a development decision? Did earlier recognition of reservoir behaviour lead to a different intervention?
Even where decisions changed, a further distinction remains between the decision and the subsequent reservoir outcome. Have materially different decisions associated with these technologies produced materially different results, and what evidence supports that interpretation? Comparative experience may help examine that question, but technology adoption alone cannot answer it.
Across Fulcrium’s comparative reservoir work, apparent performance differences have sometimes become less straightforward as they are examined more closely. Outcomes initially associated with reservoir-management practice may also reflect reservoir characteristics, development configuration or production-system constraints. At times, a material difference remains without a sufficiently supported explanation.
The residual is not the conclusion. It is the point at which a different technical enquiry begins.
These observations draw on Fulcrium’s comparative reservoir work involving approximately 15 international and national oil companies, considered in the context of wider full-value-chain benchmarking experience across approximately 90 oil and gas operators. These are descriptions of the scope of experience, not a statistical sample from which general reservoir-performance relationships are claimed.
A useful comparison may change how a reservoir team interprets a performance difference, which explanation it tests next, or which development option it regards as credible. That consequence does not necessarily depend on identifying the highest-performing operator or a practice to copy. Sometimes the more consequential result is discovering that the existing interpretation does not adequately account for the evidence.
The question then becomes whether another operator is doing something we have not understood – or whether the comparison is revealing something about reservoir behaviour that we have not yet understood.
And sometimes the most valuable thing we can learn from another operator is not what to replicate.
It is what to question in our own reservoir.
Andrew Gould discusses why establishing what world-class looks like under comparable conditions is fundamental to long-term competitiveness across integrated oil & gas systems.
Andrew reflects on performance, leadership, and the realities of operating at scale across global oil & gas systems.
The conversation explores how performance is shaped across assets, services, and the wider value chain – and why these interactions are often not fully visible within individual organisations.
Many of the perspectives shared remain highly relevant today, particularly in how performance is understood and evolves across systems and over time.
The discussion spans multiple dimensions of benchmarking – from upstream performance and cost, through to value chain, organisational, and service company dynamics.
When you step into a supergiant field such as Ghawar, Burgan, or Rumaila, it becomes immediately clear that no two reservoirs are alike – even when production levels appear similar on paper.
Differences in structure, drive mechanisms, fluid systems, development history, and operating strategy all shape how performance is expressed over time.
In these environments, performance is not defined by a single metric.
It is the result of how multiple factors interact:
pressure support, sweep efficiency, well placement, completion design, and the timing and sequencing of interventions.
Across large-scale reservoir systems, this complexity is well understood.
Significant capability has been built – in modelling, surveillance, data integration, and increasingly in the application of AI – to better characterise behaviour and support more informed decisions.
As a result, performance is continuously being improved.
But as capability increases, the nature of the question begins to shift.
It is no longer only about how to improve performance within a reservoir.
It is also about how that performance is anchored against what is fully achievable under comparable conditions.
Two reservoirs may exhibit similar production profiles, while operating at different levels of recovery efficiency.
Two development strategies may deliver comparable short-term outcomes, while diverging materially over the life of the field.
And two assets may both be improving – while progressing toward different performance horizons.
These differences are not always immediately visible.
Not because they are hidden – but because they are expressed through the interaction of many variables over time.
Best-in-class performance does not immediately reveal itself.
It exists – but it is not obvious.
As a result, the interpretation of performance becomes increasingly important.
Not just within a reservoir, but across reservoirs.
This is where benchmarking plays a distinct role.
Not as comparison alone, but as a discipline that maintains valid, comparable interpretation of performance across fields, reservoirs, over time, and as conditions evolve – helping ensure that performance is anchored against what is fully achievable, and that insight can be applied without distortion.
Best-in-class performance is not obvious – but it exists, can be identified, and applied within and across oil & gas systems.
In a discussion with Fulcrium, Sir Jim Ratcliffe – Founder and Chairman of INEOS – reflects on the role benchmarking plays in building world-class industrial performance.
The discussion spans petrochemicals, upstream and downstream oil & gas, operational performance, cost, safety, commercial outcomes, and the management of complex industrial systems.
Although the industries differ, a common principle emerges.
Strong performance does not necessarily mean best-in-class.
Across sophisticated industrial systems, organisations continuously improve.
Yet continuous improvement alone does not establish whether performance remains close to what is achievable.
The more important question is not:
“Are we improving?”
It is:
“What remains achievable under comparable conditions?”
That question cannot be answered through current performance alone.
Not simply as comparison, but as the discipline through which organisations establish what remains achievable.
Across petrochemicals, upstream and downstream oil & gas, and integrated industrial systems, this perspective becomes increasingly important as systems mature and operating environments evolve.
Establishing what remains achievable depends upon comparative context, longitudinal observation, system-wide visibility, and understanding how performance has evolved across comparable industrial systems over extended periods of time.
Few organisations possess and maintain visibility across all of these dimensions simultaneously, over time, and as conditions evolve.
Without that perspective, even sophisticated organisations may continue to optimise increasingly constrained interpretations of future potential.
Performance differences are rarely explained by a single decision.
They emerge through the interaction of technical disciplines, operating philosophies, investment decisions, development pathways, infrastructure configuration, and the way those decisions propagate throughout complex systems over time.
As a result, the gap between current performance and what remains achievable is often:
This is where benchmarking moves beyond measurement.
It establishes what remains achievable under comparable conditions.
It provides the comparative perspective through which opportunities that were previously difficult to recognise become visible.
Ultimately, benchmarking is not simply about understanding performance today.
It is about maintaining visibility of what remains achievable as industrial systems evolve.
Fulcrium helps hydrocarbon systems establish what remains achievable under comparable conditions – revealing opportunities for long-duration value extension.
In LNG systems, performance is not constrained by individual components, but by how tightly integrated systems operate under continuously shifting conditions.
On the surface, trains within the same complex often look similar – comparable output, comparable efficiency.
But over time, that similarity starts to drift. Not dramatically. Not all at once. Just enough to matter.
Small differences begin to show up:
Individually, none of these raise concern. Collectively, they start to shape how performance evolves.
What makes this more complex is that LNG trains don’t operate in isolation.
So what you’re seeing isn’t just train performance – it’s system behaviour. And this is where things get interesting.
Two trains can sit side by side, deliver similar results, yet behave very differently when conditions change. Over time, that difference becomes more than noise. It becomes structural.
Later trains often reflect improvements – design refinements, operational learning, adjustments based on earlier experience.
But those improvements are shaped by what has already been seen. And just as importantly, what hasn’t.
What works in one context doesn’t always carry across cleanly to another.
This is where benchmarking starts to shift in meaning. It’s no longer just about comparison. It’s about understanding what actually holds up.
The reality is, large-scale LNG operations already have data. They have visibility. They have highly capable teams.
Benchmarking itself isn’t new. It happens internally, across joint ventures, and with external providers.
And yet, outcomes vary. Not because of a lack of effort – but because maintaining comparability as systems evolve is far from straightforward.
What matters is not just what is measured, but how performance is interpreted in context.
When you look across LNG systems in this way, patterns begin to emerge.
These differences are rarely visible from within a single system.
What distinguishes more advanced approaches is access to unbiased data, and the ability to assess performance across systems, and over time, and in a way that remains consistent.
At that point, benchmarking becomes something different.
Not a comparison exercise –
but a discipline through which performance can be understood, tested, and extended…
…beyond the limits of any individual system or organisation.
Best-in-class performance does not immediately reveal itself.
It exists – but it is not obvious.
Benchmarking is the discipline of establishing what world-class looks like under comparable conditions – and understanding how far performance remains from that potential.
It provides an independent reference point for what remains achievable.
As systems evolve across field life, development phases, operating conditions, and changing business priorities, benchmarking helps maintain visibility of what world-class performance continues to look like – and what it may take to achieve it.
An FDP may be technically rigorous, internally assured and robust across the development assumptions and uncertainty ranges examined. That does not necessarily establish that the assumptions, ranges and alternatives themselves remain adequate.
This distinction becomes consequential when a mature field approaches a major Field Development Plan (FDP) revision, redevelopment concept selection or substantial infrastructure commitment.
Operators routinely examine alternative development concepts, optimise production strategies and test the sensitivity of development economics to uncertain reservoir and operating conditions. The more difficult question is whether the appraisal has adequately represented the development possibilities and evidence relevant to the decision.
Does continued optimisation of the existing FDP offer the most defensible development choice, or does the available evidence warrant strategic reassessment?
The question does not presume a deficiency in the operator’s technical capability or assurance processes. Sophisticated operators already possess substantial development experience, field-specific knowledge, external analogue information and established investment decision procedures.
Rather, it concerns the limits of what an appraisal can establish from the assumptions and alternatives it contains.
An appraisal may be internally consistent and its preferred alternative robust across the uncertainty ranges examined. Yet additional field-specific information, relevant external experience or changes in development conditions may still raise questions about the completeness of those ranges, the plausibility of particular assumptions or the development alternatives considered.
The technical task is to distinguish an opportunity for further optimisation from evidence that may justify reconsidering the development strategy itself.
An established FDP is not necessarily an exhausted development strategy.
Opportunities may remain through improved well placement, completion design, injection management, reservoir surveillance, production allocation and facilities performance.
At Upper Zakum, ADNOC and its partners, including ExxonMobil and INPEX/JODCO, adopted a development architecture combining artificial islands and extended-reach drilling. In 2022, ADNOC reported that further extended-reach drilling could access previously undeveloped resources, potentially increasing production capacity by 15,000 barrels of oil per day without additional infrastructure [1].
This illustrates how an established development architecture can continue to offer opportunities as technology and technical understanding advance.
Other fields may encounter constraints that incremental improvement cannot adequately address. Ageing infrastructure, changing distributions of remaining hydrocarbons, well configurations or development sequencing may warrant materially different approaches.
In gas developments, the relevant constraints may extend across reservoir deliverability, wells, gathering, compression, processing and export infrastructure.
The Schiehallion–Loyal Quad 204 redevelopment illustrates the breadth of alternatives that may be examined during redevelopment planning. A published technical account describes an appraisal involving 16 customised depletion plans, combining four development options with four offtake-rate cases. The development options included refurbishment of the existing FPSO, a new FPSO with subsea development, a tension-leg platform and a semisubmersible concept [2].
These examples illustrate different forms of development choice. They do not establish that either operator overlooked an alternative or pursued an inappropriate strategy.
The distinction between optimisation and strategic reassessment concerns the development choices being examined, not simply the physical scale of an intervention.
Optimisation seeks additional value within the strategy being pursued. Strategic reassessment reconsiders material assumptions, objectives or choices underlying that strategy, including recovery approach, depletion sequencing, development configuration, infrastructure commitments or investment timing. A major FDP revision may involve both.
The economic task is to compare feasible development choices on a consistent basis, considering remaining-field value, uncertainty, risk and the operator’s applicable strategic and investment criteria.
Higher recovery, increased production capacity and extended field life may contribute to an attractive development outcome, but none necessarily maximises economic value.
Remaining-field net present value (NPV) is an important measure, although production-capacity commitments, resource stewardship, domestic supply obligations, infrastructure dependencies, emissions requirements and portfolio considerations may introduce additional objectives or constraints.
A technically attractive development alternative is therefore not necessarily the preferred investment choice.
Development alternatives may also differ in the flexibility they preserve. Staged investment, deferred commitments and adaptable infrastructure can allow subsequent decisions to respond to improved reservoir understanding or changing economic conditions. Their value must be assessed alongside production performance, capital requirements and the consequences of irreversible commitments [12].
Strategic reassessment is not justified merely because another operator has achieved higher recovery, greater productivity or adopted a different development approach.
The relevant question is whether new understanding, changed circumstances or comparative evidence raises a material question about the development choices currently being pursued.
Three conditions are particularly important.
A material development question. There must be a credible reason to reconsider an existing assumption, reassess the relative merits of alternatives or investigate a development possibility not adequately represented in the current appraisal. This may arise from improved reservoir understanding, new technologies, changing infrastructure constraints, revised economics or relevant experience elsewhere.
Technical relevance. The evidence must be sufficiently applicable to the subject field to justify examination. Geological characteristics, fluid properties, reservoir architecture, recovery mechanisms, wells, facilities and implementation requirements must be considered. Success elsewhere does not establish local feasibility.
Potential decision consequence. The question must have sufficient prospective significance, relative to the costs, risks and uncertainties of further examination, to warrant appraisal. Its consequence may concern the selected development concept, investment scale, sequencing, implementation conditions or the evidence required before commitment.
Strategic reassessment does not necessarily require additional data acquisition. Existing evidence may be sufficient to justify reconsidering the alternatives or assumptions defining the decision. Where additional information is proposed, its value depends on whether it could improve the decision sufficiently to justify the associated cost, delay and risk [11].
These conditions do not require proof that a different strategy will outperform the existing FDP. They require a defensible basis for believing that further examination could materially affect the decision.
The practical distinction is whether the question can be adequately addressed by optimisation within the existing strategy, or whether the development choices and assumptions underlying that strategy themselves require reconsideration.
A change in injection strategy, well configuration or facilities capacity may constitute optimisation within the existing FDP. However, where evidence raises a material question about the recovery approach, allocation of development capital, depletion sequencing or infrastructure strategy, broader reassessment may be warranted.
This distinction is not a rigid threshold. Several individually modest changes may collectively justify a different development strategy, while a substantial capital investment may remain consistent with the original approach.
Strategic reassessment need not involve reopening the entire FDP. Its scope should follow the assumptions and choices affected by the evidence.
Where no material question is established, continued optimisation may remain the appropriate course.
Major operators already undertake uncertainty analysis, sensitivity testing and comparative evaluation of development alternatives.
These established practices can demonstrate whether a preferred strategy remains robust across the assumptions and uncertainty ranges examined. They do not, solely through internal consistency or sensitivity analysis, establish that those assumptions, ranges and alternatives adequately represent all material development possibilities.
The adequacy of those boundaries may already have been examined through field-specific evidence, external analogues and independent assurance. Further comparative examination is justified only where a material question remains insufficiently resolved.
Three distinct elements must be separated.
Comparative evidence supplies observations about relevant fields, development practices, interventions and outcomes.
Comparative analysis evaluates the relevance of those observations, examines possible explanations for differences and establishes what can legitimately be inferred for the subject development.
Independent judgement provides a professionally accountable interpretation of the evidence and analysis, including their limitations and potential implications for the decision.
Independence does not compensate for inadequate evidence or analytical capability. Nor does external information necessarily improve an appraisal that already contains sufficient relevant evidence.
Where a consequential evidential question remains, comparative analysis may affect development appraisal through three mechanisms.
FDP appraisals depend on assumptions about reservoir behaviour, recovery, well productivity, production decline, operating performance, development costs and uncertainty.
Appropriately selected external analogues may support, qualify or challenge performance expectations underlying the appraisal.
For example, a comparison may reveal differences in well productivity or decline behaviour that remain unexplained after accounting for relevant reservoir, completion and operating characteristics.
The observation does not establish that the subject field could reproduce another operator’s outcome. Nor does it establish that the original development assumption is incorrect.
It identifies a question requiring further examination.
The explanation may lie in reservoir characteristics, well design, operating conditions, measurement conventions or factors not adequately represented in the comparison. In some cases, the evidence may challenge not merely an assumed parameter value, but the adequacy of the explanation or model used to derive it.
The significance of a comparative observation therefore depends on whether the difference can be explained, whether any unresolved difference affects a material development assumption and whether that assumption matters to the decision.
A difference that is explained by relevant field characteristics may require no change to the appraisal. An unresolved difference may justify further investigation without necessarily affecting the development choice.
Only where the evidence supports a consequential revision to an assumption, alternative or decision condition does the comparative finding become material to the FDP decision.
The usefulness of external evidence depends on the question. Development practices, productivity outcomes and cost relationships may be amenable to comparison, while reservoir connectivity and field-specific subsurface behaviour may depend predominantly on proprietary information and detailed modelling.
External evidence cannot automatically provide statistically reliable probability distributions or substitute for field-specific reservoir models.
An operator may already have evaluated several technically credible development alternatives.
Their relative attractiveness depends on assumptions about production profiles, recovery, capital requirements, operating costs, implementation risks and uncertainty.
Consider an appraisal comparing continued infill drilling within existing infrastructure with a more substantial redevelopment involving different well configurations and facilities investment.
The preferred alternative may depend materially on expected well productivity, decline behaviour, recovery and infrastructure performance.
Relevant external evidence might indicate that a productivity assumption warrants reconsideration under comparable reservoir and completion conditions.
The operator can then examine the consequences through its field-specific technical and economic appraisal.
The significance of the comparative finding depends not simply on the size of the observed difference, but on the sensitivity of the development decision to the assumption being examined.
A substantial discrepancy may have limited decision significance if the preferred strategy remains robust across plausible values. Conversely, a relatively modest discrepancy may be consequential where it changes the relative attractiveness of development alternatives.
Even where comparative evidence does not change the preferred concept, it may influence investment timing, development sequencing, contingency provisions, implementation safeguards or the further evidence required before commitment.
A comparative finding is therefore most useful when its implications can be traced from the observation, through a defensible interpretation and any consequential development assumption, to a potential decision consequence.
This is an illustration of the analytical mechanism, not a reported project outcome.
External experience may identify a credible development configuration or approach not adequately represented in the current appraisal.
Changes in technology, reservoir understanding, infrastructure constraints or economics may also justify reconsidering an option previously rejected.
This does not establish that the operator overlooked an alternative or made an incorrect earlier decision. Such conclusions require examination of the original appraisal, its assumptions and the circumstances prevailing when the decision was made.
The technical question is whether the evidence identifies a development possibility that remains feasible and potentially consequential under current conditions.
A previously rejected alternative may warrant renewed examination because the circumstances affecting its feasibility or economics have changed. Equally, an approach successful elsewhere may remain unsuitable for the subject field.
Challenging assumptions, changing comparisons and changing alternatives describe three ways comparative evidence may affect an FDP decision.
The outcome may be confirmation of the existing strategy, qualification of its assumptions, reconsideration of alternatives, modification of investment conditions or identification of further evidence required.
Consequential development decisions do not always lend themselves to established benchmarking methodologies or predefined performance indicators.
Conventional benchmarking can establish relative performance against selected measures and comparators. A major FDP revision or unusual development concept may instead require a comparative assessment constructed around the specific decision.
The analytical question determines the evidence required, the appropriate comparators, the system boundary and the limits of defensible inference.
Comparative evidence must be evaluated against the characteristics relevant to the question being investigated.
Recovery factor benchmarking may identify differences in observed recovery performance. Analogue benchmarking can then examine reservoir characteristics, recovery mechanisms, development approaches and operating conditions that may explain those differences.
Longitudinal comparisons may examine development decisions, interventions and reservoir-management practices alongside performance over time.
Such evidence can reveal patterns, inform hypotheses and test the plausibility of proposed explanations. Historical associations do not themselves establish causation, particularly where reservoir characteristics, earlier development decisions and simultaneous interventions influence performance.
A performance gap is therefore a starting point for investigation, not a measure of recoverable opportunity.
The question is what the evidence permits one to infer about the subject development, including the uncertainty and limitations of that inference.
The appropriate analytical boundary depends on the development question.
Reservoir performance cannot always be examined independently of wells, facilities and infrastructure.
In a mature waterflood approaching its produced-water handling capacity, changes to injection strategy may affect sweep, water production, facilities utilisation, infill well productivity and the economics of additional drilling or processing capacity.
Removing an apparent constraint does not necessarily increase economically recoverable production. Other constraints may remain binding, production may shift between periods, or the additional output may not justify the expenditure required.
Similar interactions arise in gas developments, where reservoir deliverability, wells, compression, gathering, processing and export infrastructure jointly influence production capability and economics.
The analytical boundary should therefore expand where necessary to explain the observed difference or assess a feasible development alternative, rather than being fixed by a conventional disciplinary classification.
The methodology, comparators and analytical boundaries depend on the decision and the available evidence.
Field-specific information may include reservoir models, production forecasts, development alternatives, capital estimates, operating costs, fiscal assumptions and other technical and commercial inputs.
Where the evidence permits, external comparisons can be examined against these inputs, including through comparative economic evaluation using consistent assumptions.
A defensible comparative opinion must distinguish three matters:
What the evidence establishes. The observed outcomes, relevant comparisons and conclusions supported by sufficiently applicable evidence.
What remains uncertain. Differences between comparators, limitations of the available information, unresolved explanations and assumptions requiring further field-specific examination.
What follows for the decision. Whether the evidence supports existing assumptions, warrants their reconsideration, changes the assessment of alternatives or identifies matters requiring further appraisal.
The purpose is not to replace the operator’s development appraisal with an external benchmark. It is to establish what comparative evidence can legitimately contribute to the assumptions and choices underlying that appraisal.
The operator retains responsibility for developing the FDP and making the development and investment decision.
An independent comparative institution is responsible for the independence, evidential basis, methodology, limitations and defensibility of its own opinion. That responsibility does not constitute comprehensive technical verification or certification of the FDP, reservoir model or reserves estimates.
Fulcrium’s approach to decision-specific comparative judgement draws on accumulated cross-operator experience and benchmarking disciplines spanning reservoirs, wells, drilling, facilities, capital projects, operations and costs.
Its Relationship Governance Standard (FRGS-001) addresses independence, trust and confidentiality, while its Opinion Governance Standard (FOGS-001) addresses evidential sufficiency, scope integrity, methodological integrity, transparency and professional responsibility.
Where the evidence cannot responsibly support the intended conclusion, the appropriate outcome may be a qualified opinion, additional evidence, revised scope or no opinion.
The balance between continued optimisation and strategic reassessment cannot be determined by a general rule.
For some fields, the greatest remaining economic value lies in improving the existing strategy. For others, changes in reservoir understanding, technology, infrastructure economics or remaining resource distribution may justify reconsidering material elements of it.
The distinction becomes particularly important before a major FDP revision, redevelopment concept selection, material recovery-strategy change or substantial infrastructure commitment.
Four questions follow from the preceding analysis:
Are the material performance and economic assumptions underlying the preferred development concept consistent with sufficiently relevant evidence?
Have credible development alternatives been examined using assumptions and uncertainty ranges that remain defensible?
How sensitive is the development decision to plausible changes in the assumptions that matter most?
Could relevant comparative evidence change the assessment of alternatives, the conditions under which investment proceeds or the need for further appraisal?
These questions do not presume that the existing FDP is deficient or that redevelopment is preferable to continued optimisation.
They concern the evidential boundaries within which the development decision is being made.
An appraisal may demonstrate that a preferred development strategy is robust across the assumptions and alternatives examined. That is a meaningful result, but its significance depends on whether those assumptions and alternatives adequately represent the material possibilities relevant to the decision.
Those boundaries may already have been examined and found defensible through the operator’s own technical work and assurance processes. Where a material question remains unresolved, relevant comparative evidence may help support existing assumptions, explain observed differences, challenge an interpretation or identify development possibilities warranting further appraisal.
The contribution of independent comparative judgement is conditional on the relevance and sufficiency of the evidence, the quality of its interpretation and its potential significance to the decision.
A defensible FDP decision requires both a robust appraisal of the development choices examined and an adequate evidential basis for the boundaries within which those choices were evaluated.
Independent FDP Benchmarking provides an external comparative basis for examining development assumptions, evaluating relevant analogues and identifying where optimisation may need to give way to a more fundamental reassessment of the development strategy.
Independent FDP Benchmarking
For operators examining the assumptions, alternatives and development choices underlying a Field Development Plan, Fulcrium provides independent comparative evidence and judgement.
Further information: FDP Benchmarking – Field Development Plans.
1. ADNOC (20 October 2022). ADNOC Drilling Delivers New World Record for the Longest Well. https://www.adnoc.ae/en/news-and-media/press-releases/2022/adnoc-drilling-delivers-new-world-record-for-the-longest-well
2. Duncan, E., and Burchell, M. (2015). BP Employs Lessons Learned for North Sea EOR Projects. Offshore, 6 May 2015. https://www.offshore-mag.com/production/article/16758510/bp-employs-lessons-learned-for-north-sea-eor-projects
3. Schiozer, D. J., et al. (2019). Model-based decision analysis applied to petroleum field development and management. Oil & Gas Science and Technology, 74, 46. https://doi.org/10.2516/ogst/2019019
4. Mirzaei-Paiaman, A., Santos, S. M. G., and Schiozer, D. J. (2021). A review on closed-loop field development and management. Journal of Petroleum Science and Engineering, 201, 108457. https://doi.org/10.1016/j.petrol.2021.108457
5. Sun, S., and Pollitt, D. A. (2022). An Empirical Analog Benchmarking Workflow to Improve Hydrocarbon Recovery. SPE Reservoir Evaluation & Engineering, 25(2), 349–366. https://doi.org/10.2118/208612-PA
6. Fulcrium. Reservoir Benchmarking. https://www.fulcrium.com/reservoir-benchmarking/
7. Fulcrium. Oil & Gas Benchmarking. https://www.fulcrium.com/oil-gas-benchmarking/
8. Fulcrium. About Fulcrium. https://www.fulcrium.com/about-fulcrium/
9. Fulcrium. Fulcrium Relationship Governance Standard (FRGS-001), Issue 1.1. https://www.fulcrium.com/fulcrium-relationship-governance-standard/
10. Fulcrium. Fulcrium Opinion Governance Standard (FOGS-001), Issue 1.0. https://www.fulcrium.com/fulcrium-opinion-governance-standard/
11. Bratvold, R. B., Bickel, J. E., and Lohne, H. P. (2009). Value of Information in the Oil and Gas Industry: Past, Present, and Future. SPE Reservoir Evaluation & Engineering, 12(4), 630–638. https://doi.org/10.2118/110378-PA
12. Smith, J. E., and McCardle, K. F. (1999). Options in the Real World: Lessons Learned in Evaluating Oil and Gas Investments. Operations Research, 47(1), 1–15. https://doi.org/10.1287/opre.47.1.1
Fulcrium provides independent benchmarking and comparative judgement to the oil and gas industry.
Its work draws on accumulated cross-operator experience and complementary benchmarking disciplines spanning reservoirs, wells, facilities, operations, capital projects and commercial performance.
Engagements are constructed around the specific technical, operational or investment questions requiring examination, combining relevant comparative evidence with the operator’s own information where appropriate.
Fulcrium’s independent benchmarking opinions are governed by institutional standards addressing professional relationships, independence, evidential sufficiency, methodological integrity and professional responsibility.
© Fulcrium 2026. All rights reserved. Proprietary analysis and methods.
Two LNG trains may have similar nameplate capacities, comparable liquefaction technologies and broadly similar operating histories, yet deliver materially different annual LNG production. The difference is observable; its significance is not. Higher output may reflect better equipment reliability or operating decisions, but it may equally reflect feed-gas availability, ambient conditions, plant configuration, maintenance requirements or restrictions elsewhere in the production system.
The question is not simply which train produces more. It is how much of the observed difference can be explained, what remains uncertain and whether any part could realistically be reduced at the asset being assessed.
An unexplained difference is not an improvement target. It is a technical question requiring further investigation.
There is no single population of comparable LNG assets suitable for every performance question. A comparison of refrigeration efficiency requires attention to process configuration, compressor arrangements, cooling systems, feed composition and ambient conditions. A comparison of annual LNG production must also consider feed supply, planned maintenance, equipment failures, operating restrictions, storage and loading availability.
Two assets may therefore be suitable comparators for one engineering question and unsuitable for another. Similar nameplate capacity alone does not establish comparability.
The production measure also matters. Nameplate capacity, current rated capacity, maximum demonstrated throughput, LNG produced and LNG loaded are not interchangeable. A train may demonstrate a high production rate for several days without being capable of sustaining it over a year. Conversely, a train may have additional production capability that cannot be used because feed gas is unavailable.
Comparability must be established for the specific question and the particular measure being investigated.
Suppose one of two nominally similar LNG trains produces materially more LNG over twelve months. The difference may reflect refrigeration duty, compressor condition, maintenance strategy or operational decisions. It may also reflect upstream feed restrictions, ambient conditions, planned shutdowns or limitations on storage and loading. Until these factors have been examined, the production difference says little about the relative effectiveness of the operators.
Even after the principal known constraints have been considered, a remaining difference does not establish inferior performance. It may arise from incomplete records, measurement uncertainty, unsuitable comparators, unrecognised interactions or a variable omitted from the explanation. Equally, the underlying model, comparator population or framing of the question may be inadequate. The residual identifies what the analysis has not yet explained; it does not quantify what can be recovered.
Three quantities should be kept distinct:
A compressor modification may increase physical capability without increasing annual LNG production if feed supply remains limiting. A reliability improvement may reduce equipment downtime without increasing LNG deliveries if storage or loading restrictions prevent the additional production from reaching the market. The engineering question is not whether a component could perform better in isolation, but what would change in the output of the connected system.
Production losses are not always equal to the duration of an equipment interruption multiplied by a reference production rate.
Consider a hypothetical LNG train capable of producing 500 tonnes per hour. A compressor-related outage lasts 100 hours. A simple calculation suggests a loss of 50,000 tonnes. But during 35 of those hours, no feed gas is available. Repairing the compressor would not restore production during that interval.
For the counterfactual, assume the compressor is repaired at the start of the 100-hour interval, but restarting the train would still require ten hours after the 35-hour feed interruption. A downstream restriction would then have limited production to 300 tonnes per hour for twenty hours; only during the remaining 35 hours could production have reached 500 tonnes per hour. For comparison, assume that without the compressor repair the train would have produced nothing throughout the same 100-hour period.
Under these deliberately simplified assumptions, the repaired train produces nothing during the 35-hour feed interruption or the subsequent ten-hour restart, and its incremental plant production relative to the unrepaired case would be:
(20 hours × 300 tonnes/hour) + (35 hours × 500 tonnes/hour) = 23,500 tonnes.
That is substantially less than the apparent 50,000-tonne loss. Nor is it necessarily 23,500 additional tonnes of LNG sold or delivered: storage movements, boil-off, loading restrictions and the assessment period may affect the commercial outcome.
The example is illustrative, not an empirical finding. It demonstrates why the cause of an equipment interruption should not automatically be assigned an equivalent quantity of recoverable production. Removing one constraint may leave another in place or expose a bottleneck that was previously concealed.
A defensible estimate therefore requires reconstruction of what the facility actually produced during the relevant intervals, what it could have produced, and what would plausibly have happened if the specific restriction had been removed. Aggregated downtime categories alone cannot establish that counterfactual.
An LNG train operates within a system extending from feed-gas receipt through treatment, liquefaction, utilities, storage and loading. The appropriate boundary of analysis depends on the proposed inference.
For a narrowly defined compressor-efficiency question, refrigeration equipment, drivers, operating conditions and supporting utilities may provide the relevant boundary. For a question about additional annual LNG production, that boundary may be too narrow. More efficient compression will not necessarily increase output if another process stage is limiting; improved liquefaction availability may not increase deliveries where feed supply is persistently constrained.
The boundary should not be expanded indiscriminately. It should expand where necessary to determine whether a subsystem performance difference has a consequence for whole-asset production. A technically meaningful equipment improvement and a recoverable production opportunity are not the same proposition.
An operator may understand its own equipment and operating history in considerable detail while having limited evidence of how similar equipment and configurations perform elsewhere. Cross-operator comparisons can challenge assumptions that have become accepted within a single asset. They may show, for example, that a presumed operating limitation is not universal or that a different level of compressor reliability has been demonstrated under relevant conditions.
The comparator population nevertheless determines what can legitimately be inferred. A group selected for exceptional performance does not establish typical performance. A group containing persistently feed-constrained assets may understate demonstrated technical capability. Typical peer performance, upper-range demonstrated performance and capability under specified engineering conditions answer different questions.
Nor does the observation that another operator achieved a better result establish what caused it. A practice associated with superior performance may depend on equipment condition, feed characteristics, process configuration or infrastructure that differs materially from the asset being assessed. The observation can support an engineering enquiry, but it cannot by itself establish causation or transferability.
The useful question is therefore not only what another operator achieved, but under what conditions it was achieved and whether those conditions permit a defensible inference about this asset.
Consider a train experiencing repeated refrigeration-compressor trips. Relevant peer evidence may indicate that other trains experience fewer interruptions. That difference merits investigation, but it does not quantify additional LNG that the affected train could produce.
The assessment must examine failure mechanisms, frequency and duration, associated operating conditions and the restrictions coinciding with each trip. It must then identify a plausible intervention and test its consequences. Would the intervention prevent failures, shorten recovery time or alter the operating envelope? What other constraints would limit production once compressor interruptions were reduced? Would additional plant production become additional LNG available for delivery?
An improvement demonstrated elsewhere can strengthen the engineering case that an intervention is feasible. It cannot replace examination of the local failure mechanism, operating conditions and interacting constraints. A performance difference becomes a credible opportunity only when there is a supportable explanation of how a specific change could reduce it.
Even a carefully constructed comparison will retain uncertainty. Production measurements may require reconciliation; failure histories may be incomplete; models may not represent all operating conditions; and comparator populations may differ in ways that cannot be fully adjusted. Future feed availability, maintenance requirements and intervention effectiveness are also uncertain.
These limitations should be reflected in the conclusion rather than hidden within a single improvement estimate. If an apparent opportunity disappears under a reasonable alternative comparator population, the evidence may be insufficient to support it. If the benefit requires simultaneous improvements in feed supply, reliability and loading capacity, those dependencies must be explicit. An opportunity that remains material under several credible scenarios provides a stronger basis for further engineering work.
Technical achievability is also distinct from commercial justification. Incremental LNG must be considered alongside investment cost, implementation downtime, operating risk, integrity requirements and the economic consequences of the change. Benchmarking may establish that an intervention is technically plausible without establishing that it should be undertaken.
LNG performance reflects the interaction of engineering design, equipment condition, operating decisions, feed supply, reliability and the wider production system. Some differences between assets are inevitable; others may reveal opportunities that cannot be recognised from one facility’s operating history alone.
Independent comparative evidence is useful because it tests whether explanations accepted within an individual asset remain convincing when examined against relevant experience elsewhere. It may show that an apparent gap is largely explained by conditions outside the operator’s control, that an assumed limitation is less fundamental than believed, or that a material difference remains unexplained. The last outcome is not a failed comparison: it may identify the most important question for the next stage of technical investigation.
The objective is not to convert every observed difference into an improvement target. It is to establish which differences warrant intervention, which require further evidence and which should not be interpreted as recoverable performance at all.
The most consequential question may therefore not be why another LNG asset produces more.
It may be what would actually have to change for our own asset to produce more – and whether the evidence supports the conclusion that it could.
'
