Every E&P company tracked production losses in some form. Whether it is called deferral tracking, loss accounting, or forecast versus actual variance analysis, the intent is always the same. Someone needed to explain where the missing hydrocarbon molecules went.
Yet ask a simple question in many asset reviews: where exactly did we lose production last month? The room often goes quiet. Numbers exist somewhere in spreadsheets, dashboards, or enterprise systems, but turning them into a clear operational story has always been much harder than it should be.
In smaller operations, Excel often carried this responsibility for years. It was flexible, familiar, and fast to deploy. But as assets grew, wells multiplied, and production systems became more interconnected, spreadsheets eventually reached their limits. Version control broke down. Formulae became fragile. The person who built the workbook left the company, and nobody fully understood what sat behind the tabs.
At that point most organisations began migrating toward enterprise solutions designed to track production losses across multiple assets and facilities. Yet the transition often created a different set of problems. Engineers felt like they were feeding a system rather than solving problems. They spent time entering loss events, categorising causes, and filling mandatory fields without seeing clear operational benefits in return. Meanwhile, management still struggled to obtain reliable answers to that fundamental question. Where exactly did we lose production, and why?
This disconnect was common across the industry. And it usually pointed to the same root issue.
Production deferral tracking was treated as an administrative requirement rather than an engineering tool.
In reality, it should sit at the core of every production engineer’s responsibilities. When designed properly, the system becomes the operating memory of the asset. It records what failed, when it failed, how much it cost the asset, and what patterns emerged over time. It transforms operational anecdotes into measurable information.
In other words, it makes the data confess.
Why Production Engineers Should Care
If you have ever felt that the time spent entering production loss data could have been better used fixing the issue itself, it is worth stepping back and considering the broader purpose of the process.
The data serves several important objectives. It allows engineers to identify and quantify losses in a consistent way across assets. It helps teams prioritise where to act first based on impact rather than instinct. It improves production forecast assumptions by providing historical evidence of downtime and operational constraints. It strengthens surveillance by highlighting emerging patterns before they become major failures.
When used properly, deferral data feeds directly into better forecasting, stronger surveillance practices, and more informed operational decisions. It also increases accountability within teams because the cost of operational disruptions becomes visible and measurable.
That is why production deferral tracking sits at the centre of production engineering practice. It is not about accounting for lost molecules. It is about building a system that allows the asset to explain its own behaviour.
Planned Versus Unplanned Deferment
One of the most fundamental classification decisions in any deferral tracking system is the distinction between planned and unplanned events.
Planned deferment covers production losses that were known in advance and scheduled into the production plan. Turnarounds, planned maintenance shutdowns, and pipeline pigging campaigns all fell into this category. These events appeared in the forecast. The production impact was estimated before the event began, and actual performance was compared against the plan once it concluded.
Unplanned deferment was everything else. Equipment failures, process upsets, compressor trips, unexpected well performance changes, and third party interruptions that were not anticipated. These were the events that created the gap between forecast and actual production, and they were typically where the greatest operational value lay in understanding root causes and driving improvement.
This distinction matters because planned and unplanned losses require fundamentally different management responses. Planned shutdowns need to be integrated with production forecasts and maintenance schedules so that inconsistencies are surfaced early and resolved before the event begins. Unplanned events require rapid root cause analysis, pattern recognition across equipment and systems, and prioritisation frameworks that direct resources to where they will recover the most production.
Getting this classification right at the beginning gives the asset team a much clearer picture of where operational improvement opportunities actually exist.
System Design Before Data Capture
Building an effective deferral tracking system was never straightforward. One of the biggest challenges lay in deciding the appropriate level of detail. If the system was too granular, engineers struggled to maintain it and data quality quickly deteriorated. If it was too high level, the information became too vague to support operational decisions. Both extremes reduced the value of the system.
Automation and artificial intelligence were often proposed as solutions to this challenge. Automated classification sounded attractive because it promised to capture detailed information without overwhelming engineers.
But automation cannot replace understanding.
If engineers themselves were uncertain how to classify events manually, automated classification was never going to be trusted. More importantly, if everything became automated too early, engineers risked losing familiarity with the data they were supposed to interpret. Engineers needed to understand the data intimately before automation could meaningfully assist them.
One principle consistently improved system design. The people who analyse the data must be involved in designing how the data is captured. Reservoir engineers who performed pressure transient analysis participated in the design of drillstem tests. Production engineers who analysed production losses should likewise participate in designing the deferral tracking system itself.
How Should Loss Events Be Recorded?
One of the first design choices concerned how production loss was represented in the data model. Should loss be captured as discrete events, or as frequent status records for each component in the system, like on, off, and ready traffic lights on each well?
Most systems adopted the event-based approach. A loss event was created with attributes such as start date, end date, equipment involved, and estimated production impact. A multi day shutdown was captured as a single record with editable start and end times. This approach kept the dataset compact and made historical editing relatively straightforward.
The alternative was a status based model where each well or component recorded its operating state at regular intervals, often daily. In such systems, a single event generated many rows of data. While this structure proved useful for certain analytical workflows, historical corrections became more difficult because multiple records had to be updated from the start to the end of the event.
The choice between these two approaches carried long term implications for data quality, analytical capability, and system maintainability. Most organisations found the event based model easier to sustain over time.
Building the Equipment Hierarchy
Production systems are complex networks of facilities, wells, pipelines, and surface equipment. The hierarchy used in the deferral tracking system should ideally align with the physical structure of the asset and, where possible, with enterprise systems such as SAP.
This alignment matters because production losses are closely linked with equipment reliability and maintenance cost. Integrating the equipment hierarchy allows organisations to connect production deferrals with maintenance records, reliability analysis, and risk based maintenance strategies.
A typical hierarchy starts at the field and well level. From there it extends into major equipment groups such as completions, instruments, and Christmas tree assemblies. In more detailed implementations, the hierarchy continues down to individual equipment tags.
The choice depends on the level of diagnostic capability the organisation wants to achieve. A system built with only field level granularity will never support equipment reliability analysis without significant rework. Investing in the right hierarchy from the start avoids expensive redesign later.
Where It Broke Versus Why It Broke
Should the system record where the loss occurred, or attempt to capture the underlying root cause?
Root cause classification can be powerful but also dangerous. Without clear definitions, teams easily fell into endless debates about the true origin of a problem. Engineers knew that operational failures often involved multiple contributing factors. When root cause categories were poorly designed, everything eventually ended up labelled as “reservoir” or “facility constraints,” which added little practical insight.
For this reason, many systems focused first on where the loss manifested in the production system before attempting deeper causal analysis. A well designed reason tree with clear category definitions did more to improve data quality than an ambitious root cause framework that nobody used consistently.
How Much Did We Actually Lose?
Perhaps the most difficult aspect of production deferral tracking was the measurement of impact. Calculating how much production was lost during an event was rarely straightforward. Engineers had to decide how to treat ramp up and ramp down periods, how to establish baseline production rates, and how to account for measurement uncertainty.
Many organisations relied on measurement by difference, comparing expected production against measured output. This approach introduced its own challenges, especially when wellhead meters, allocation systems, and aggregated facility measurements were involved. Variation in measurement uncertainty across different flow meters, process upsets, and shrinkage factors all influenced the calculated loss. Small assumptions in the methodology could significantly change the reported numbers.
For that reason, transparency in the calculation methodology was essential. Documenting the assumptions behind impact estimates mattered just as much as recording the events themselves.
Make the Data Confess
After careful design and diligent data capture, the engineer’s work does not end with producing charts and reports. The real value comes from interrogating the data for insights.
Patterns in unplanned failures reveal reliability problems. Trends in planned deferment highlight opportunities for improved maintenance planning. Loss statistics inform operational strategies, whether the philosophy is run to fail or pre emptive maintenance supported by predictive analytics.
Organisations that treated deferral tracking as a genuine engineering discipline consistently uncovered hidden production. They identified recurring operational weaknesses and addressed them systematically.
The question that remains is not whether to track production losses. Every company does that already. The question is how to design the system well enough to capture the right data, then let the digital production engineers make it confess.