Why Wait for Something to Go Wrong? Learning From the Work That Goes Right

Safety professionals spend a great deal of time studying failure. An incident occurs and an investigation begins. Evidence is collected, interviews are conducted, causes are identified, corrective actions are developed and lessons are shared. Near misses receive similar treatment, particularly when the potential consequences could have been serious.

This makes sense. When something goes wrong, organisations should understand why. But there is an obvious limitation to this approach: most work does not result in an incident.

Every day, millions of tasks are completed successfully across construction sites, offices, factories, processing facilities, drilling operations, warehouses and transportation networks. Conditions change, equipment behaves unexpectedly, information is incomplete, people become unavailable, schedules move, contractors interact, priorities compete and weather changes. Yet most of the time, people adapt and the work succeeds.

We rarely investigate that success with anything approaching the same curiosity we apply to failure. Perhaps we should, because if organisations only learn when something goes wrong, they are ignoring the largest source of operational information available to them: normal work.

Success Is Not Simply the Absence of Failure

When a job finishes without an incident, we tend to conclude that everything worked as intended. That may be true, but it is worth asking what actually happened. Did the work proceed exactly according to plan? Were all the expected resources available? Did the procedure perfectly describe the task? Did conditions remain unchanged? Did nobody encounter an unexpected problem?

In many workplaces, the answer to at least one of those questions will be no. People adjusted. A supervisor changed the sequence, a technician noticed something unusual and investigated, a worker requested additional equipment, two contractors coordinated an interface that had not been anticipated, or an experienced employee recognised that conditions were different and stopped the task before continuing.

The work succeeded partly because people responded intelligently to the conditions they encountered. IOGP's Learning from Normal Work guidance was developed specifically to help organisations extract actionable lessons from everyday operations rather than relying only on incidents as the trigger for learning.

This represents an important shift in thinking. Success is not simply the absence of something going wrong. Success is often the result of people actively making things go right.

The Procedure Describes Expected Work

Procedures are necessary. They establish expectations, define important controls, capture organisational knowledge and create consistency. But no procedure can describe every condition a worker may encounter.

The procedure describes the work as anticipated. The employee experiences the work as it actually unfolds. Sometimes those two versions align almost perfectly; sometimes they do not. A valve is harder to access than expected, another contractor is working nearby, a tool is unavailable, a component has been replaced with a slightly different model, weather deteriorates or the job simply takes longer than planned.

People have to respond to these conditions. This difference between prescribed work and actual work is not automatically evidence of poor compliance; it is often evidence that real operations contain variability. The UK Health and Safety Executive describes human factors in terms of the interaction between the job, the individual and the organisation, including factors such as task design, workload, equipment, procedures and the working environment.

The important question is therefore not whether adaptation occurs. It almost certainly does. The more useful question is whether those adaptations strengthen the organisation's ability to manage risk or quietly weaken it.

People Are Constantly Solving Problems

Think about an experienced operator, technician or supervisor. Part of their value is precisely that they do not need every possible situation described in advance. They recognise patterns, anticipate problems, know when something does not feel right and understand which aspects of the job can tolerate adjustment and which boundaries must never be crossed.

This expertise can become almost invisible because the job gets done and nothing happens. Nobody writes an incident report because a supervisor noticed that two planned activities would interfere and separated them. There may be no formal record when a technician recognises unusual vibration and requests an inspection, when a lifting team changes the plan because wind conditions are deteriorating, or when a worker challenges an isolation because the equipment configuration does not match the drawing.

Yet the organisation has just witnessed successful risk management. We should be interested in what the person noticed, why they interpreted it as important, what experience informed their decision, what allowed them to intervene and whether another person in the same circumstances would have recognised the same warning signs. Those questions can turn everyday performance into organisational learning.

We Often Don't Know How Much Adaptation Is Occurring

Management systems frequently assume that work is being performed substantially as designed. That assumption may be wrong. Employees may have developed small adaptations throughout a process, each of which appears insignificant on its own but collectively means that the real operating system differs considerably from the documented one.

A permit step may routinely be completed earlier because the responsible person becomes unavailable later. Employees may carry a particular tool because the specified one does not work effectively. Supervisors may maintain their own spreadsheet because the corporate system is too slow. Workers may use an informal communication method because radio coverage is unreliable.

None of these adaptations necessarily creates harm. Some may actually make operations safer and more efficient. Others may gradually erode important controls. Unless management understands normal work, it cannot reliably distinguish between the two.

This is one reason organisations need to spend time observing work without approaching every interaction as an inspection. If employees believe management is always looking for violations, they will naturally show management the version of work that conforms most closely to the procedure. The organisation may then obtain excellent assurance about work as imagined while learning very little about work as actually performed.

Go and See the Job Without Looking for Faults

Traditional workplace inspections create a particular mindset. We look for missing PPE, poor housekeeping, incorrect barricades, documentation gaps and other deviations. That has value, but learning from normal work requires a different approach.

Observe a task with the deliberate intention of understanding it. Ask what makes the job easy or difficult, where people need to make decisions, where the activity differs from the plan, what information employees rely upon and what happens when something unexpected occurs. A particularly useful question is simply: “What makes this job difficult?”

That question avoids the generic “Any safety concerns?” and invites the employee to talk about the work itself. Perhaps the equipment is awkward, information arrives late, staffing is tight at particular times, another department frequently changes the plan, or the task itself is straightforward but obtaining the necessary permit is unnecessarily difficult.

These answers can reveal conditions that formal risk assessments overlook. They can also identify opportunities for improvement that have nothing to do with an incident, which is precisely the point. Safety improvement should not require somebody to get hurt first.

Ask What Surprises People

Another useful question is: “What tends to surprise you when doing this job?” Surprises matter because they identify where expectations and reality diverge.

Equipment may occasionally behave differently from what operators expect. Work may regularly take longer than planned. Underground conditions may be less predictable than drawings suggest. Contractor personnel may change more frequently than anticipated. Experienced workers often understand these variations extremely well even when the formal management system does not.

Understanding recurring surprises allows organisations to improve planning before those surprises combine with less favourable circumstances. It also provides an opportunity to capture knowledge that may otherwise remain inside the heads of a small number of experienced employees.

Every organisation has people who make difficult work look easy. They rarely attract attention because they do not create problems, but that is precisely why they deserve attention. How do they prepare? What do they notice? When do they slow down? What causes them to stop? How do they recognise when a situation is becoming unstable?

Some of this expertise will never fit neatly into a procedure. It can, however, be shared through mentoring, coaching, job observation and structured conversations. This becomes particularly important where experienced workforces are approaching retirement or turnover is increasing. Organisational knowledge does not reside entirely in documents; a great deal of it walks through the gate every evening.

Learning From Success Does Not Mean Everything Was Safe

There is an important qualification. Learning from normal work does not mean assuming that because a task succeeded, the way it was performed was safe.

A shortcut can succeed. A degraded control may not fail. A dangerous condition may produce no consequence. Successful outcome does not prove good process.

That is why learning requires analysis. If employees adapted the work, was the adaptation beneficial? Did it preserve critical controls or introduce another risk? Was it necessary because the designed process was inadequate? Should the adaptation become part of the formal method, or should the system be redesigned so that the adaptation is no longer required?

The goal is not to endorse everything people do. It is to understand it.

This distinction becomes particularly important in high-risk activities where some operational boundaries should remain firm. People need flexibility to respond to changing circumstances, but they also need absolute clarity about controls that cannot be compromised. Good systems combine disciplined management of significant risk with enough flexibility for competent people to respond intelligently when reality differs from the plan.

Normal Work Can Reveal Organisational Strain

Workload problems are not always obvious in incident statistics because people compensate. They work faster, delay lower-priority tasks, multitask, stay late or rely more heavily on memory. For a while, everything continues functioning.

HSE guidance notes that people have limited capacity for processing information, remembering items, making decisions and performing tasks, and that workload can drift over time as additional activities are gradually introduced. Observing normal work can therefore reveal pressure long before failure occurs.

How many activities is a supervisor juggling? How frequently is an operator interrupted? What gets postponed when workload increases? What happens when the operation moves away from normal conditions? A staffing level that works perfectly during steady operations may be completely inadequate during an emergency, shutdown or major upset.

This raises an important question that traditional performance measures often miss:

How much effort is required to make normal performance look normal?

A team may continue meeting every target while becoming progressively more fatigued.

People compensate until eventually they cannot. HSE guidance links fatigue with working arrangements and workload and notes its effects on attention, reaction time, information processing and decision-making.

A system that succeeds only because employees continually make extraordinary effort is not necessarily a resilient system. It may simply be borrowing resilience from its workforce.

Near Misses Are Valuable, but We Can Learn Earlier

Near-miss reporting is rightly encouraged. A near miss allows an organisation to learn without experiencing the full consequence, but it still means that something undesirable has already happened. A control failed, a hazard was released or someone was unexpectedly exposed.

Learning from normal work asks whether we can learn even earlier.

Suppose a team identifies a conflict during planning and changes the work before the activity begins. Nothing happens. There is no incident and no near miss. But there is still something worth understanding. Why did that team recognise the problem? Would another team have recognised it? Should the planning process be changed? Could the same conflict exist elsewhere?

The same principle applies to good-catch programmes. Counting how many reports were submitted tells us relatively little. The real value lies in asking what those reports are teaching us. If several workers independently identify similar problems, there may be a systemic weakness. If experienced employees repeatedly compensate for the same equipment issue, perhaps the equipment requires redesign. If one team consistently identifies problems earlier than others, perhaps its planning practices deserve closer study.

The report is not the learning. What the organisation does with the information determines whether learning occurs.

Study the Teams That Consistently Perform Well

Safety management understandably directs significant attention toward poor performers, but organisations should also study teams that consistently manage difficult work effectively.

Perhaps their pre-job discussions are better. Supervisors may spend more time in the field. Employees may be more comfortable challenging one another. Planning may begin earlier, or the team may have developed better methods of sharing information across shifts.

These practices may be transferable.

This approach is normal elsewhere in business. Companies study high-performing sales teams, manufacturers study their most productive facilities and sporting organisations study elite performers. Safety should be equally interested in understanding excellence.

The objective is not to create a simplistic list of “best behaviours.” It is to understand the conditions that support reliable performance and determine whether those conditions can be strengthened elsewhere.

Technology Makes Normal-Work Learning More Important

Automation, artificial intelligence and digital systems are rapidly changing how work is performed. It is tempting to assume that more automation means less human variability, but technology usually changes the role people play rather than removing it.

Employees may perform fewer routine actions while becoming even more important when conditions are unusual, systems fail or automation encounters something it was not designed to handle. Energy Institute guidance on implementing new technologies specifically highlights the need to consider human and organisational factors when introducing technologies including automation and AI.

This makes normal-work learning increasingly valuable. How are people actually using the new system? Which features do they ignore? What workarounds have appeared?

Where does automation help and where does it create additional workload? What happens when the technology is unavailable?

Implementation is not complete simply because the software went live. The organisation needs to understand how the technology becomes part of real work.

Create Time for Reflection

Learning does not happen automatically. People can perform excellent work every day without the organisation capturing any of the knowledge involved.

There needs to be deliberate reflection, but this does not require another complicated corporate programme. After a significant task, a short conversation can explore a few practical questions: What went well? What was different from what we expected? What did we have to adapt? What made the job difficult? What would we do differently next time? What should somebody performing this job for the first time know?

The important part is not creating another form. It is creating a mechanism through which useful insights can travel beyond the people who already know them.

Leadership response is critical. If every difference between the procedure and reality triggers criticism, employees quickly learn to report that work occurs exactly as designed. Management receives reassuring information but becomes progressively disconnected from reality.

The objective should be to create enough trust that people can describe the messy parts of work without assuming that every adaptation will automatically be treated as misconduct. That does not lower standards. It improves visibility.

Safety Is Something People Create Every Day

It is easy to think of safety as the absence of something: no injuries, no incidents, no releases and no failures. But safety is also something people actively create. They plan, communicate, notice changing conditions, anticipate problems, adapt to circumstances, intervene when something does not look right, stop work when necessary and solve problems as they emerge.

Most of these actions will never appear in an incident database precisely because they worked.

That does not make them unimportant. It makes them worth understanding.

So continue investigating incidents. Continue learning from near misses and analysing failures. But do not wait for something to go wrong before becoming curious about how your organisation works.

Go and observe a difficult job that consistently succeeds. Talk to the people who make it happen. Ask what they notice, what they adapt, what makes the work difficult and what helps them succeed. Look for the difference between a system that performs well because it is well designed and one that appears to perform well because experienced people are constantly compensating for its weaknesses.

You may discover vulnerabilities that have never produced an incident. You may find practices worth spreading across the organisation. You may also discover that some of the most valuable safety knowledge in the business has been sitting in front of you every day.

Nothing had to go wrong for it to be worth learning.

References and Further Reading

International Association of Oil & Gas Producers (IOGP). Learning from Normal Work. Guidance for operational and corporate leaders on extracting actionable learning from everyday operations in addition to learning from incidents.

UK Health and Safety Executive (HSE). Introduction to Human Factors. Guidance addressing interactions between the job, individual and organisation and the influence of task design, workload, equipment, procedures and working conditions on human performance.

UK Health and Safety Executive (HSE). Workload – Human Factors. Guidance addressing human information-processing limitations, workload drift, staffing and the need to consider workload during normal and abnormal operating conditions.

UK Health and Safety Executive (HSE). Fatigue – Human Factors. Guidance addressing the effects of working time, shift design, workload and fatigue on human performance.

Energy Institute. Guidance on Human and Organisational Factors Aspects of Implementing New Technologies. Guidance addressing human and organisational considerations when introducing technologies including automation and artificial intelligence.