Marginal gains is the idea that small, incremental improvements across a process can, over time, lead to meaningful and sustained performance benefits.
In a previous paper, Dominic explored how this concept can be applied to SCTA, alongside the importance of balancing efficiency and thoroughness. These ideas resonate with many organisations because they reflect the reality of working environments: SCTA must be both practical and effective, and improvement rarely comes from large, disruptive change. Instead, it is often achieved through continuous refinement, better engagement, and small adjustments to how analysis is carried out.
However, there is another side to this conversation that is less often discussed.
Building on that work, this blog takes a different perspective. It explores what happens when those small changes move in the wrong direction. Instead of marginal gains, we consider marginal losses: the subtle, often unnoticed ways in which the quality of analysis can erode over time. Alongside this, we introduce the idea of being an intelligent customer of SCTA: not just accepting that an analysis has been completed, but actively questioning how it has been carried out and whether it truly delivers meaningful insight.
Completion vs Understanding
Most SCTAs do not fail in obvious or dramatic ways. They are rarely missing entirely, rarely carried out by the wrong people, and rarely appear incomplete when reviewed. In fact, many analyses present well. The structure is clear, the steps are documented, and the outputs appear to cover the expected ground.
And yet, there are times when something does not feel quite right.
This feeling can be difficult to articulate. It may arise during a review of the analysis, or from those who were involved in the process itself. Everything appears to have been done, and yet there could be a sense that something important has not been fully captured. It might be a workshop where the right individuals were present, but the discussion remained close to the written procedure and never quite reached the realities of how the task is actually performed. It might be a walkthrough that was shortened or simplified due to time pressures, limiting the opportunity to observe the task in its full context.
These are not failures in the conventional sense; they are far more subtle than that. What they point to is a gap, not between doing and not doing, but between completing an analysis and truly understanding a task.
Why Balance Could Be Harder Than It Sounds
At its core, SCTA is about achieving a balance between two competing demands. On one hand, the analysis must be thorough enough to identify where human performance could influence safety, uncovering vulnerabilities, assumptions, and potential failure pathways. On the other hand, it must be efficient enough to be carried out within the constraints of real operational environments.
Those constraints are significant and unavoidable. Time is often limited, and the availability of relevant, experienced personnel can be difficult to secure. Operational priorities may shift unexpectedly, with urgent issues taking precedence over planned analysis work. In some cases, there may also be organisational pressures to deliver outputs quickly, particularly where SCTA is being conducted to meet regulatory or internal requirements.
Within this context, trade-offs are inevitable. Organisations must make decisions about how much time to allocate, who to involve, and how deeply to explore different aspects of a task. In principle, this is entirely reasonable. SCTA is not intended to be an academic exercise, but a practical tool that supports safe and reliable operations.
However, while the need for balance is widely acknowledged, the assumption that this balance is actively maintained is worth examining more closely.
Where The Balance Starts To Shift
In practice, the balance between efficiency and thoroughness is rarely static. It does not sit neatly in a “just right” position, carefully maintained over time. Instead, it is influenced by a range of subtle factors, many of which relate to mindset and intent rather than formal process.
One of the most important of these factors is how SCTA is perceived within the organisation. Over time, it can begin to shift from being viewed as an opportunity to explore and understand how work is really carried out, to being seen primarily as an exercise that needs to be completed. This shift is rarely explicit. It does not involve a conscious decision to reduce quality or depth. Rather, it emerges gradually through experience, pressures, and competing demands.
When SCTA is approached as something to complete, the nature of engagement begins to change. Discussions may become more focused on confirming existing knowledge rather than probing uncertainty. Procedures may be accepted as accurate representations of work, rather than starting points for deeper exploration. Participants may still contribute, but the space for challenge, reflection, and shared learning becomes more limited.
The process, in other words, is still followed, but the intention behind it subtly shifts, and this has a direct impact on the outcome.
How These Small Compromises Add Up
The concept of marginal gains is widely understood and widely applied. It recognises that small, incremental improvements, often individually insignificant, can accumulate over time to produce substantial benefits. This way of thinking has proven valuable in many fields, including human factors and safety analysis.
However, there is an inverse dynamic that is far less visible, but equally important.
Small compromises can also accumulate.
These compromises are rarely dramatic or concerning when viewed in isolation. They are often entirely reasonable responses to practical constraints. Relying on existing procedures instead of developing a detailed task analysis may seem efficient and proportionate. Limiting the number of participants in a workshop may be necessary due to availability. Reducing the scope of a walkthrough may be justified by time pressures or access limitations. Focusing on key steps while spending less time on supporting detail may appear to be a sensible prioritisation of effort.
Each of these decisions is understandable. Each can be justified in context.
But collectively, they begin to shape the nature of the analysis. The depth of discussion is slightly reduced. The level of challenge is slightly lower. The opportunity to surface tacit knowledge is slightly constrained. No single decision fundamentally alters the outcome, but together they redefine it.
This is the essence of marginal losses, not a failure of the process, but a gradual reduction in its ability to generate meaningful insight.
When the Detail Doesn’t Fully Reflect Reality
Even when all the expected steps of SCTA are followed, the resulting analysis may not fully capture the complexity of real-world task performance.
This is because many of the factors that influence how work is carried out are subtle, context-dependent, and not formally documented. Informal workarounds, local adaptations, variations between individuals, and knowledge developed through experience often sit outside of written procedures. These elements are rarely captured unless there is sufficient time, engagement, and openness within the analysis process to surface them.
If the analysis remains too close to the procedure, or if discussions are not sufficiently exploratory, these aspects can remain hidden. The result is an analysis that appears complete and coherent, but does not fully reflect work-as-done.
This gap is critical. It is often within these subtle variations and informal practices that vulnerability exists. When they are not identified, understood, or addressed, the analysis may provide reassurance without fully capturing the underlying risk.
The Role of the Intelligent Customer
This is where the concept of the intelligent customer capability becomes particularly important.
An intelligent customer does not simply accept that an SCTA has been completed, or that it meets expected standards in terms of format and documentation. Instead, they take a more active role in understanding and questioning the process by which the analysis was produced.
They consider how the analysis was conducted, not just what it contains. They ask whether it was used as a tool for genuine exploration, or whether it primarily served to confirm existing assumptions. They reflect on who was involved, and whether the perspectives included were sufficient to capture the realities of the task. They examine where compromises may have been made, and whether those compromises were appropriate given the risks involved.
Perhaps most importantly, they recognise that the quality of an analysis is defined by the depth of understanding it represents.
Being an intelligent customer is not about demanding more analysis for its own sake, or pushing for unnecessary detail. It is about ensuring that the analysis remains meaningful, proportionate, and grounded in reality, even within the constraints that organisations face.
Holding the Line Between Practical and Meaningful
SCTA will always involve trade-offs. It must remain practical and achievable, and it must operate within the constraints of time, resource, and organisational priorities.
However, there is an important distinction between consciously managing these trade-offs and gradually allowing the depth of analysis to reduce without recognising it. The former involves deliberate decisions, informed by an understanding of risk and context. The latter is a more passive process, where small adjustments accumulate over time and gradually shift the nature of the analysis.
Maintaining this distinction requires awareness and, in many cases, a shift in mindset. It involves moving away from a focus on completion and towards a focus on understanding. It means asking not just whether the analysis has been done, but whether it has genuinely achieved its purpose.
This is not always easy. It requires time, engagement, and a willingness to challenge assumptions. However, it is essential if SCTA is to remain an effective tool for managing risk.
A Final Reflection
Marginal gains remain a powerful way to improve SCTA, and the ideas explored in Dominic’s paper continue to provide valuable guidance for enhancing performance.
However, recognising marginal losses is just as important.
The challenge with marginal losses is that they are difficult to detect. There is no single point at which quality clearly declines, and no obvious failure that signals a problem. Instead, there is a gradual shift, shaped by a series of small, reasonable decisions that, over time, influence the depth and value of the analysis.
The risk is not that the analysis is wrong.
It is that it is almost right.
And in safety-critical environments, that difference matters. Because “just enough” has a tendency, over time, to become not quite enough, often in ways that only become visible when it is too late.
You can download our full paper on “Enhancing the Performance of Safety Critical Task Analysis: The Goldilocks Problem and Marginal gains” by clicking here.
If you’d like to read Dominic’s original blog on the topic, you can find it here.
We have a webinar and audit tool for checking the quality of your SCTAs. It’s free and we share lots of hints, tips and expert advice. Check it out here.
