Human Performance
in Practice.

Before asking why someone did something, understand the work they were trying to do.

When an unwanted event occurs, attention naturally moves toward the people involved.

Someone operated the equipment.

Someone made the decision.

Someone missed the signal.

Someone did not follow the procedure.

Someone was there.

Their actions matter.

But beginning and ending with those actions gives us a very narrow view of performance.

A person does not arrive at work and behave independently of everything around them.

They are working toward goals.

Using tools and equipment.

Interpreting information.

Following plans.

Coordinating with other people.

Responding to changing conditions.

Managing interruptions.

Balancing competing requirements.

Drawing on experience.

Making decisions with the information available at the time.

Their behaviour occurs inside work.

Human Performance starts by widening the field of view.

Instead of beginning with:

Why did they do that?

begin with:

What was happening?

What were they trying to achieve?

What did the work normally look like?

What was different?

What information was available?

What conditions shaped the decision?

What made the action reasonable at the time?

This does not make the person’s actions irrelevant.

It makes them understandable.

And understanding gives us more options for improvement than judgement alone.

A wider view produces different actions

Imagine a technician isolates the wrong item of equipment.

A narrow explanation might be:

The technician failed to identify the correct isolation point.

That may be factually accurate.

But it leaves important questions unanswered.

Were the isolation points clearly labelled?

Did two assets use similar identifiers?

Was the drawing current?

Could the isolation point be seen from the work location?

Was the technician interrupted?

Was the task different from normal?

How was isolation normally confirmed?

Had similar confusion occurred before?

Were people relying on local knowledge to distinguish between equipment?

Once those conditions become visible, the improvement options change.

The response may still include coaching or competency development.

But it might also involve equipment identification, drawings, interface design, planning, verification, access or the way the work itself is organised.

The objective is not to find a more complicated explanation.

It is to find a more useful one.

Start with the work. Then understand the action within it.

Work is never just the procedure.

Every organisation has a picture of how work should happen.

This is Work as Imagined.

It matters.

Without some shared expectation, complex work would quickly become unmanageable.

But work does not occur inside a document.

It occurs in real conditions.

Those conditions contain variation.

A job takes longer than planned.

Equipment behaves differently.

A contractor arrives late.

A patient deteriorates.

Weather changes.

Production demand increases.

Access is restricted.

A replacement component is unavailable.

Another crew is working nearby.

A worker knows from experience that one part of the procedure does not fit the current situation.

People respond.

That is part of Work as Done.

The difference is not automatically a failure

It is tempting to treat any difference between Work as Imagined and Work as Done as non-compliance.

Sometimes it is.

Sometimes the required way of working is clear, workable and important, and someone chooses not to follow it.

But that is not the only explanation.

The work may have changed.

The procedure may contain assumptions that no longer hold.

The equipment may have been modified.

The sequence may not work under certain operating conditions.

The written process may omit knowledge that experienced workers use every day.

The organisation may be asking for two things that cannot both be achieved as expected.

The difference between intended and actual work is therefore something to understand before it is something to correct.

The aim is not to eliminate the gap

There will always be some difference between the way work is described and the way it unfolds.

The objective is not perfect alignment.

It is to understand where the difference matters.

Ask:

Where is variation necessary?

Where has local practice become better than the written method?

Where is adaptation hiding poor design?

Where are people compensating for system weaknesses?

Where has the gap created additional risk?

Where does the intended method need to change?

Where does actual practice need to change?

Both Work as Imagined and Work as Done contain valuable information.

Good organisations learn from the relationship between them.

Behaviour does not occur in a vacuum.

We often talk about performance as though it belongs entirely to the individual.

Good performer.

Poor performer.

Experienced worker.

Careless worker.

Competent worker.

Complacent worker.

But performance is produced through an interaction between the person and the system around them.

Consider the same experienced worker doing the same task.

It is the same person.

But it is not the same work.

Conditions do not determine behaviour

This point needs care.

Time pressure does not make every person take the same shortcut.

Poor equipment does not guarantee an error.

Fatigue does not mean failure is inevitable.

People have agency.

They make choices.

They bring different experience, capability and judgement to the work.

But those choices occur within conditions that make some actions easier, harder, more attractive, less obvious or more likely.

That is why changing conditions can change performance even when the people remain the same.

The practical implication is important.

When something goes wrong, do not ask only:

What was wrong with the person?

Ask:

What was shaping the work?

And when something goes well:

What conditions helped people succeed?

Human Performance is interested in both.

People adapt because work requires it.

No procedure can anticipate every combination of conditions that people will encounter.

So people adjust.

They change sequence.

Find another tool.

Coordinate differently.

Slow down.

Speed up.

Ask for help.

Use experience.

Create a workaround.

Stop.

Recover.

Prioritise.

Most of these adaptations never appear in an incident report.

They appear as successful work.

That matters.

Because organisations often depend on people’s ability to adapt far more than they realise.

Adaptation is not automatically good

Human Performance can become unhelpful when adaptation is romanticised.

Not every workaround is clever.

Not every deviation should be preserved.

Not every local practice is safe.

Adaptation can maintain performance.

It can also reduce margin.

A workaround used occasionally can become normal.

A temporary arrangement can become permanent.

A shortcut that has worked repeatedly can become increasingly difficult to recognise as unusual.

Past success can reinforce confidence in a practice without demonstrating that the practice is robust.

The aim is not to remove human flexibility.

It is to understand where the system is depending on it.

The outcome changes what becomes obvious.

After an event, we know what happened.

We know which signal mattered.

Which decision was important.

Which condition became critical.

Which control failed.

Which action should have been different.

The people involved did not have that view.

They were moving forward through the work.

Not backwards from the outcome.

Local rationality

People generally do things that make sense to them in the situation they are in.

That does not mean every decision is good.

It means there is usually a reason.

A driver enters an area because previous access has been permitted.

A nurse uses a workaround because the normal equipment is unavailable.

A maintenance worker begins a task because the information available indicates the isolation is complete.

A supervisor continues work because the condition has occurred many times before without consequence.

Once the outcome is known, the weaknesses can appear obvious.

Before it, they may not have been.

The task for learning is to reconstruct that earlier view.

What did they know?

What did they expect?

What had happened before?

What cues were available?

What competing goals existed?

What would another person with similar experience reasonably have done?

This produces a more useful explanation than:

They should have known better.

Error tells us something happened. It does not yet tell us why.

Some explanations stop just as the useful investigation should begin.

Human error.

Poor judgement.

Complacency.

Failure to follow procedure.

Lack of attention.

Communication failure.

These descriptions may identify something real.

But they often rename the behaviour rather than explain it.

Keep going

If someone did not follow a procedure, understand why.

Was the procedure known?

Was it available?

Was it understandable?

Did it fit the task?

Was the required equipment available?

Did experienced workers normally perform the work differently?

Was there a competing requirement?

Was the deviation deliberate?

What benefit did the alternative provide?

What risk did the person believe they were taking?

This is not an exercise in finding a system explanation for every action.

Sometimes the answer may be straightforward.

A person understood the requirement, had reasonable conditions to meet it, understood the potential consequence and deliberately chose otherwise.

That matters.

But we should arrive at that conclusion through understanding, not assumption.

Do not stop at the label.

The same adaptations that help work succeed can also create vulnerability.

Consider an experienced team that has learned how to complete a difficult task efficiently.

They know where delays normally occur.

They know which parts of the process can overlap.

They know which information matters most.

They know when to wait and when to continue.

This expertise helps the work succeed.

Then one condition changes.

A different piece of equipment.

A new contractor.

An unusual configuration.

A delayed handover.

A less experienced team member.

An adaptation that worked hundreds of times now produces a different result.

Learn from both sides

If we only study failure, we can easily conclude that the behaviour associated with the outcome was itself abnormal.

Normal work may tell a different story.

The same behaviour may occur frequently.

The same workaround may be widespread.

The same judgement may normally produce success.

That does not make it safe.

It makes it important.

Success and failure are not always produced by completely different systems.

Sometimes the difference is smaller than we think.

If you want different behaviour, look at what the system is making easy, hard, likely or necessary.

Organisations often try to change behaviour through communication.

Reminders.

Campaigns.

Training.

Rules.

Posters.

Conversations.

These can help.

But behaviour is also shaped by the design of the work.

If the safe method requires more time than the schedule allows, that matters.

If the required tool is difficult to access, that matters.

If the procedure requires information people do not have, that matters.

If stopping work creates significant operational consequences, that matters.

If one metric rewards speed while another expects caution, that matters.

People notice what the system rewards, tolerates, makes difficult and repeatedly requires.

Instead of asking only:

How do we get people to comply?

also ask:

What makes the desired action easy to perform?

What makes it difficult?

What makes another action attractive or locally reasonable?

What conditions repeatedly produce the behaviour we are trying to change?

Can we redesign those conditions?

Behaviour change becomes more sustainable when the work supports the behaviour being asked for.

Most work ends well. That does not mean nothing interesting happened.

A task was completed.

Production continued.

The patient received care.

The train departed.

The equipment returned to service.

The project moved forward.

No incident occurred.

So the organisation moves on.

But inside that successful work may have been:

a control that was difficult to use

a last-minute equipment substitution

an experienced worker recognising a weak signal

a team recovering from a planning error

a workaround everyone knows about but nobody has documented

a conflict between two procedures

a supervisor creating additional time

a worker catching a mistake before consequence.

Normal work contains information about how the system actually functions.

Learn from what helps

Learning from normal work is not simply looking for hidden problems.

It is also understanding success.

What made the work easier today?

What helped the team coordinate?

Which information was useful?

What did an experienced person notice?

What allowed the team to recover?

Which part of the system worked particularly well?

What should we avoid accidentally removing when we redesign the work?

Organisations can become very good at identifying what went wrong and remain surprisingly weak at understanding what routinely goes right.

Both matter.

The quality of learning depends on the quality of the questions.

People quickly recognise the purpose behind a question.

A question can sound open while still carrying judgement.

Why didn’t you follow the procedure?

may really mean:

Explain why you were wrong.

When people expect judgement, they protect themselves.

Answers become shorter.

Complexity disappears.

Uncertainty is removed.

The story becomes safer to tell.

If we want to understand work, our questions need to create room for the work to be described.

Questions that open the work

Then listen.

Do not use the question simply as a bridge to the answer you already believe.

Curiosity is not a script.

Work is difficult to understand from a dashboard.

Data can tell us a great deal.

Where events occur.

Which actions are overdue.

What people reported.

How often an activity was completed.

Which controls were verified.

Where performance differs.

But data is a representation of work.

It is not the work itself.

Some things become visible only when we spend time where work happens.

Field interactions can do more than check

There is an important place for field assurance.

Are controls present?

Are requirements being met?

Is the work within agreed limits?

But every field interaction does not need to be only an inspection.

Leaders can also learn.

Ask:

What is going well today?

What is making this harder than it should be?

What has changed since the job was planned?

Where do you have to adapt?

What feels fragile?

What are you relying on?

What would make this easier to do well?

Then look at the work together.

The purpose is not to catch someone doing something wrong.

It is to understand what the system looks like from inside the work.

Bring the people who know the work together to understand the work.

Some work is too complex to understand through a single interview.

Different people see different parts.

Operators.

Maintainers.

Supervisors.

Planners.

Engineers.

Contractors.

Support functions.

Each may understand something the others do not.

A Learning Team creates a structured way to bring those perspectives together.

What a Learning Team is not

It is not a disciplinary meeting.

It is not an investigation interview conducted in a group.

It is not a root-cause workshop with different language.

It is not a forum for validating an explanation already decided.

And it is not simply a conversation.

The purpose is to develop a richer understanding of how the work functions.

Good Learning Teams stay close to the work

The value is not the workshop itself.

It is what the organisation understands and improves because the workshop happened.

The response to failure shapes what the organisation will learn next.

People pay attention to what happens after bad news.

If reporting difficulty leads to blame, people learn to be careful about what they report.

If admitting uncertainty damages reputation, certainty increases.

If raising a workaround creates trouble for the person who describes it, workarounds become harder to see.

The organisational response therefore affects future visibility.

But learning does not require the removal of accountability.

A mature response can hold both.

Accountability needs context

Different situations warrant different responses.

An unintended mistake is not the same as a deliberate act.

A reasonable adaptation under difficult conditions is not the same as knowingly exposing others to serious harm.

A person working within an accepted local practice is not in the same position as someone deliberately bypassing a clear and workable requirement for personal convenience.

Intent matters.

Capability matters.

Foreseeability matters.

The quality of the system matters.

The conditions matter.

The potential for harm matters.

Human Performance should improve the quality of accountability.

Not remove it.

A useful response asks both:

What should we learn about the system?

and:

What is a fair and appropriate response to the behaviour?

The strongest improvements change the conditions around performance.

Many corrective-action systems repeatedly produce the same responses.

Retrain.

Remind.

Communicate.

Update the procedure.

Add a warning.

Tell people to be more careful.

Sometimes those are exactly the right actions.

Often they are not enough.

If the problem was created by equipment design, another briefing will not redesign the equipment.

If two goals conflict, retraining does not remove the conflict.

If information arrives too late, a reminder does not make it arrive earlier.

If the task routinely requires a workaround, rewriting the procedure without changing the work may simply create a newer procedure that people still cannot use.

Make good work easier

The aim is not to design people out of work.

It is to design work that makes good performance more achievable.

What leaders pay attention to shapes what the organisation can see.

Field leadership often has an assurance purpose.

Check the permit.

Review the controls.

Confirm the procedure.

Inspect the equipment.

Verify the standard.

These activities matter.

But if every interaction is a check, leaders see only part of the work.

Sometimes the more useful role is to learn.

Know which conversation you are having

Problems arise when assurance is disguised as learning.

A leader asks an apparently curious question but is really testing compliance.

Workers recognise this quickly.

Be clear.

Sometimes:

I need to confirm that this control is in place.

Other times:

I want to understand how this work actually happens. Show me what makes it difficult.

Both conversations are legitimate.

They simply serve different purposes.

Strong operational leadership can do both.

A good conversation is not the end point.

An organisation can run Learning Teams, conduct field interactions, teach Human Performance principles and collect hundreds of operational insights without materially changing the work.

Learning activity is not the same as improvement.

Eventually, something needs to change.

Close the learning loop

People who contribute operational knowledge should be able to see what happened because they spoke.

Not every suggestion will be adopted.

Not every issue can be solved immediately.

But silence after consultation teaches people something too.

It teaches them that sharing information changes little.

That is how learning becomes organisational capability rather than a series of good conversations.