Stop Measuring Safety Activity and Start Measuring Safety Effectiveness

Most organisations are not short of HSE data. If anything, the opposite is true.

Modern organisations can measure almost everything. Inspections completed, observations submitted, toolbox talks delivered, training hours recorded, audits conducted, corrective actions closed, leadership visits undertaken, permits issued, risk assessments completed, and incidents reported. Increasingly sophisticated dashboards can present this information instantly, complete with trends, traffic lights, rankings, and performance scores.

The problem is not whether we can measure HSE activity. The problem is whether all that activity is telling us what we actually need to know.

An organisation can complete 10,000 safety observations without becoming significantly safer. It can deliver thousands of hours of training without improving competence. It can close every audit action while allowing the underlying weaknesses that created those actions to remain. It can conduct hundreds of leadership safety visits that produce little more than another completed KPI.

All of those activities may have value. But completing an activity and achieving an outcome are not the same thing.

That distinction should fundamentally change the way we think about HSE performance measurement.

The Comfort of Counting Things

There is a reason organisations gravitate towards activity metrics: they are relatively easy to measure.

An inspection either happened or it did not. A toolbox talk can be counted. Training attendance can be recorded. An action can be classified as open or closed. A leadership visit can be entered into a system and displayed on a dashboard.

These numbers create an impression of control.

Senior management can see whether the organisation is completing the activities prescribed by the HSE management system. Business units can be compared. Targets can be established. Underperforming departments can be challenged.

There is nothing inherently wrong with this.

The problem begins when activity becomes a substitute for effectiveness.

Imagine an organisation establishes a target requiring every supervisor to conduct four safety observations each month. Initially, the programme may encourage supervisors to spend more time in the field and have useful conversations with employees.

But once the number becomes a performance target, something else can happen.

The organisation begins managing the number.

Supervisors know they need four observations, so four observations are submitted. Managers monitor completion percentages. Reminders are sent near the end of the month. Dashboards turn green when the target is achieved.

Eventually, the conversation shifts from “What are we learning from these observations?” to “Have you completed your four observations?”

The metric has quietly become the objective.

Activity Is Not the Same as Effectiveness

Consider something as familiar as HSE training.

An organisation may report that 98 per cent of employees completed mandatory training. From a compliance perspective, that can be important information.

But what does it tell us about risk?

Did employees understand the material? Can they apply it? Can they recognise the relevant hazards? Can they demonstrate the required skills under real operating conditions? Has their behaviour changed? Has the training addressed the actual causes of poor performance?

A completion rate cannot answer those questions.

The same principle applies across HSE.

A risk assessment completed is activity. A risk assessment that identifies the important hazards and results in effective controls is effectiveness.

An inspection completed is activity. An inspection that identifies deteriorating conditions before they contribute to an incident is effectiveness.

A corrective action closed is activity. A corrective action that prevents recurrence is effectiveness.

A toolbox talk delivered is activity. A conversation that causes the team to recognise a changing condition and modify the work is effectiveness.

A leadership safety visit completed is activity. A leadership interaction that identifies an organisational barrier and results in meaningful improvement is effectiveness.

Once we make this distinction, many traditional HSE dashboards start to look very different.

The Problem With Green Dashboards

One of the most dangerous situations in HSE is not a dashboard full of red indicators.

It is a dashboard full of green indicators in an organisation carrying significant uncontrolled risk.

Green creates confidence. It tells leaders that systems are functioning, actions are being completed, and expectations are being met. That confidence is useful when it is justified.

When it is not, the dashboard can become part of the problem.

An organisation might report excellent training compliance while critical competencies remain weak. Audit actions may show 95 per cent closure, while several important actions have been repeatedly extended. Inspection targets may be exceeded because large numbers of low-value observations are being recorded. Injury rates may remain low despite increasing exposure to serious hazards.

Every indicator can be technically correct while the overall picture is misleading.

This is particularly important in high-hazard industries where the absence of frequent incidents does not necessarily demonstrate effective control of catastrophic risk.

A business can operate for long periods without a major event, even when important barriers are degraded. Low-frequency, high-consequence events do not provide enough failure data for organisations to rely on outcomes alone.

The question, therefore, becomes: what evidence do we have that the controls preventing serious events are actually working?

That is a much more valuable question than asking whether the dashboard is green.

Lagging Indicators Still Matter

There has been a tendency in recent years to present leading indicators as modern and lagging indicators as outdated. The distinction is not that simple.

Lagging indicators matter.

Fatalities, injuries, occupational illnesses, environmental releases, process safety events, vehicle incidents and other unwanted outcomes contain important information.

Organisations need to understand what has happened, identify trends and learn from adverse events.

The limitation is that these indicators tell us about events that have already occurred.

They are also influenced by statistical variability. A business unit can experience a year without a serious injury because it managed risk exceptionally well. Another can achieve the same result while relying heavily on good fortune.

The number alone cannot tell us which organisation is stronger.

This becomes even more problematic when organisations use injury rates as the dominant measure of safety performance. A declining recordable injury rate may be encouraging, but it does not automatically demonstrate that fatal and major accident risks are under control.

The hazards that cause frequent minor injuries are not always the same hazards capable of causing catastrophic events.

Organisations therefore need a balanced picture: outcomes, activities, system health and evidence of control effectiveness.

Leading Indicators Need to Lead Somewhere

The response to the limitations of injury statistics has been an increased focus on leading indicators.

This is generally positive, but simply measuring something before an incident occurs does not automatically make it useful.

Many so-called leading indicators are actually measures of activity: number of inspections, percentage of training completed, number of observations submitted or percentage of audits performed.

These indicators tell us that preventive processes are taking place.

They do not necessarily tell us whether those processes are effective.

A genuinely useful leading indicator should provide information about the condition of the system and ideally signal whether risk is increasing or decreasing.

For example, instead of measuring only how many inspections were completed, an organisation might examine the percentage of critical-control verifications that identified a control as unavailable, degraded, or ineffective.

Instead of measuring only training completion, it might assess demonstrated competence for safety-critical roles.

Instead of measuring the number of corrective actions closed, it might evaluate the recurrence of similar findings.

Instead of counting management safety visits, it might examine the systemic issues identified through those visits and whether they were resolved.

The difference is subtle but important.

We move from measuring whether the organisation performed the safety activity to measuring what the activity tells us about the health of the system.

Measure What Prevents the Event

For organisations managing serious and fatal risks, one of the strongest opportunities is to connect performance measurement directly to critical controls.

Start with the event you do not want.

A dropped object. A loss of containment. An uncontrolled energy release. A vehicle collision. A lifting failure. A confined-space fatality. A major fire or explosion.

Then identify the controls that must function to prevent that event or mitigate its consequences.

Now measure those controls.

Are they present? Are they suitable? Are they being used? Are they functioning as intended? Are people competent to apply them? Are there recurring conditions that degrade them?

This produces a fundamentally different conversation.

Instead of telling senior management that 147 inspections were completed this month, we can tell them that 96 per cent of sampled critical controls were verified as effective, three recurring weaknesses were identified in energy isolation and one business unit is experiencing increasing degradation of vehicle journey-management controls.

One set of information describes activity.

The other supports decisions about risk.

That is what HSE performance information should ultimately do.

Beware of Targets That Change Behaviour

Every metric creates the potential to influence behaviour.

Sometimes that influence is exactly what we want.

If leaders know contractor safety performance will be examined during procurement decisions, contractor management may improve. If supervisors know critical controls will be verified in the field, greater attention may be given to those controls.

But poorly designed targets can create unintended consequences.

If managers are rewarded for having zero overdue actions, difficult actions may be closed prematurely.

If supervisors must submit a minimum number of observations, low-value observations may multiply.

If business units are heavily judged on injury rates, people may become reluctant to report injuries.

If departments are ranked according to the number of hazards reported, trivial hazards may suddenly become abundant.

This does not necessarily mean people are deliberately manipulating the system. Human beings naturally respond to the measures by which their performance is judged.

HSE professionals therefore need to ask an uncomfortable question whenever a KPI is introduced: If people wanted to make this number look good without actually reducing risk, how could they do it?

If the answer is easy, the metric deserves closer examination.

The Executive Dashboard Should Tell a Risk Story

Senior leaders do not need every piece of HSE data the organisation collects.

They need enough information to understand the risk story.

Where is exposure increasing? Which critical controls are weakening? Where are recurring problems appearing? Which business units are improving? Where are significant actions overdue? What organisational conditions are changing? What requires executive intervention?

An effective executive HSE dashboard should therefore help leaders make decisions rather than simply demonstrate that the HSE department is busy.

That may mean showing fewer indicators.

A dashboard containing 40 KPIs can create the appearance of sophistication while making it difficult to identify what actually matters. A smaller set of carefully selected indicators, supported by analysis and narrative, can be considerably more powerful.

The purpose of the dashboard is not to display everything we know.

It is to focus leadership attention on what they need to know.

This requires judgement. It also means the dashboard may need to change as the organisation's risk profile changes.

The indicators relevant during routine operations may not be the same indicators needed during a major shutdown, rapid expansion, organisational restructuring or the mobilisation of a large project.

Performance measurement should follow risk, not the other way around.

Ask Better Questions

Improving HSE measurement does not necessarily require another software platform or a major analytics programme.

It can begin by asking better questions about the measures already being used.

For every significant HSE KPI, ask: What are we trying to learn? What decision should this information support? Does the measure tell us about activity, effectiveness or outcome? Can the number look good while risk is deteriorating? Could the target unintentionally encourage the wrong behaviour? What other information is needed to interpret it correctly?

Some existing indicators will survive that test.

Others will not.

The goal is not to eliminate activity measures. Organisations still need to know whether required processes are happening. If inspections are not being performed or training is not being completed, that matters.

The mistake is assuming completion proves effectiveness.

It does not.

From Counting Safety to Understanding Risk

The future of HSE performance measurement should not be about collecting more data.

Most organisations already have more data than they effectively use.

The opportunity is to become much more selective about what we measure, why we measure it and what decisions the information supports.

Technology will make this increasingly important. Digital HSE systems, connected equipment, sensors, mobile inspections, artificial intelligence and advanced analytics will dramatically increase the volume of safety information available to organisations.

More data will not automatically produce better decisions.

Without a clear understanding of what matters, organisations risk creating increasingly sophisticated ways of measuring the wrong things.

The strongest HSE measurement systems will combine several perspectives. They will understand adverse outcomes, monitor preventive activity, examine organisational conditions and, critically, provide evidence that important controls are functioning.

They will also recognise that numbers require context.

A KPI should begin a conversation, not end one.

If an indicator changes, leaders should want to understand why. If it remains green for long periods, they should occasionally challenge whether it is still sensitive enough to tell them anything useful. If performance appears exceptionally good, they should understand what is driving that success rather than simply celebrating the number.

Ultimately, the purpose of HSE measurement is not to prove that safety activities are taking place.

It is to help the organisation understand whether risk is being effectively managed.

That is a much higher standard.

And it requires us to stop being impressed simply because everything on the dashboard is green.

References and Further Reading

International Association of Oil & Gas Producers (IOGP). Process Safety – Recommended Practice on Key Performance Indicators. Industry guidance addressing the use of leading and lagging indicators to understand process safety performance and the health of barriers and management systems.

Center for Chemical Process Safety (CCPS). Process Safety Metrics: Guide for Selecting Leading and Lagging Metrics. Guidance on developing balanced process safety indicators and using performance information to strengthen risk management.

UK Health and Safety Executive (HSE). Developing Process Safety Indicators: A Step-by-Step Guide for Chemical and Major Hazard Industries (HSG254). Guidance on developing leading and lagging indicators linked to the control of major hazards.

US Occupational Safety and Health Administration (OSHA). Recommended Practices for Safety and Health Programs. Guidance covering programme evaluation, leading indicators and the use of performance information to improve occupational safety and health management.

International Organization for Standardization (ISO). ISO 45001:2018 – Occupational Health and Safety Management Systems. Requirements relating to monitoring, measurement, analysis, performance evaluation and continual improvement.