Good Safety is Good Business – But How Do We Know?

Griff Bromfield-Jenkins

Oct 8

11 min read

Updated: Oct 14

“Good safety is good business” is one of those phrases that has become so familiar within the safety profession that we rarely stop to question what it actually means. I have used it myself, and on the face of it the argument seems obvious. Fewer injuries should mean less disruption and absence, lower costs and potentially better productivity and reputation. All of those things should be good for business.

But the more I have thought about it, the less convinced I am that this tells us very much about either safety or business performance. I initially approached this question by wondering whether there was a demonstrable relationship between safety and financial performance. Could we show that organisations with better safety performance make more money? There is research that attempts to establish those relationships, but I increasingly think that starting there risks reducing business performance to something much narrower than it actually is.

Businesses routinely make decisions that make their short-term financial position worse in order to improve their long-term prospects. They invest in assets, acquire businesses, enter new markets, restructure, develop people and build capability. A loss in a particular year does not necessarily mean that a business has performed badly, just as a healthy profit does not necessarily mean that it is well positioned for the future. If business performance is broader than financial outcomes, perhaps the more interesting question is not whether good safety causes better financial performance, but how safety should be understood as part of organisational performance in the first place.

"If safety is genuinely part of business performance, why do we continue to measure and manage it largely as a separate system?"

Looking at performance more broadly

This isn't a particularly new question in management thinking. Kaplan and Norton's original Balanced Scorecard paper was published in 1992 and argued that financial measures alone were insufficient to understand organisational performance. Reading it more than thirty years later, parts of it inevitably show their age. The particular perspectives and measures reflect the period in which it was written, and what organisations now consider material to performance has developed considerably. I wouldn't take the original scorecard and suggest that it represents how a business should measure itself today.

But dismissing it because of its age would also miss the point. I found the thinking behind the framework much more interesting than the individual measures within it. It brought together information from different parts of an organisation because no single function could adequately describe how the business was performing. Finance, operations, customers and organisational capability contributed different information to a wider understanding of whether the organisation was delivering its strategy. The measures weren't necessarily permanent either; they could change as priorities and circumstances changed. More importantly, performance information wasn't simply there to demonstrate whether the strategy was being delivered. It should help management learn, challenge assumptions and, where necessary, adjust the strategy itself.

None of that sounds particularly radical today. What interested me was what happened when I compared it with how we often measure safety.

We have got very good at measuring safety

I have sat through more safety performance reviews than I care to remember. In large organisations, the amount of information available can be extraordinary. I have seen performance packs running to 50 or 60 slides covering injuries, incidents, near misses, audits, inspections, training, action closure, observations, safety conversations and various combinations of leading and lagging indicators. I have also contributed to those systems myself, so this isn't a criticism I can make entirely from the sidelines.

None of those measures is inherently wrong. In the right circumstances, many of them are useful. What I increasingly question is whether collecting more of them necessarily gives us a better understanding of performance. The limitations of relying on injury statistics aren't a new discovery either. Pawłowska's research into leading and lagging indicators was published in 2015, based on organisational data collected considerably earlier, and shows that the profession has been wrestling with this issue for a long time. More recently, ISO 45004 provides a considerably broader approach to performance evaluation, recognising different organisational contexts and levels, quantitative and qualitative information, and the need to select indicators that reflect what an organisation is actually trying to achieve.

What struck me when reading the older research was how familiar many of the measures still looked. Our ability to collect, manipulate and visualise safety data has changed enormously over the last twenty years; I am less convinced that the underlying measures have evolved at the same rate. We can now produce exceptionally sophisticated dashboards containing essentially the same information.

I also wonder whether we have created some confusion between leading and lagging indicators and being proactive or reactive. There is an entirely understandable desire for organisations to demonstrate that they manage safety proactively. The difficulty is that this can push us towards things that are relatively easy to count before an unwanted event occurs: inspections completed, conversations held, training delivered, observations submitted or actions closed. Because they happen before the accident, they readily acquire the label of a leading indicator and, by implication, something positive.

I am not sure the logic always holds. A thousand safety conversations certainly tells me that a thousand safety conversations happened. Without understanding their quality, purpose, what they revealed or what changed as a result, I am less certain what they tell me about future safety performance. Conversely, information from something that has already happened may reveal an important change in the system and allow an organisation to intervene before something more serious occurs. Perhaps the objective shouldn't be to find more leading indicators simply to demonstrate that we are proactive. It should be to identify the information that tells us something meaningful about performance and helps us make better decisions.

"A thousand safety conversations certainly tells me that a thousand safety conversations happened. I am less certain what it tells me about future safety performance."

What are we actually trying to understand?

This is where I think the comparison with wider business performance becomes useful. We wouldn't expect every part of a business to demonstrate its performance using exactly the same measures, yet with safety there can be a tendency to roll out a common set of indicators and expect them to mean something everywhere.

Take a relatively low-risk office environment. Counting minor injuries may tell me very little about whether that workplace is being managed well. I might be more interested in DSE assessments, what workplace inspections are telling me, wellbeing outcomes or patterns in absence. Move into a manufacturing environment and the things I need to understand change. Machinery condition, maintenance, defects, competence and process deviations might suddenly be far more important. Move up again to executive level and I probably don't need either set of measures in that detail.

This sounds obvious when written down. I'm not convinced it is always reflected in how organisations report safety. Measures can survive because they are easy to collect consistently and aggregate, rather than because they remain the best way of understanding performance. One of the things I found interesting in Kaplan and Norton's thinking was the expectation that measures themselves would evolve as priorities changed or particular issues were addressed. Safety can be remarkably reluctant to let go of a KPI once it has found its way onto a dashboard. Again, there are exceptions. HS2 is probably one of the better examples I have encountered of being prepared to change the emphasis of its measures as the organisation, its maturity and its priorities changed.

So I don't think the challenge is to find the definitive set of safety KPIs. The measures should depend on what the organisation is trying to achieve, where the risk sits, who needs the information and what they can actually do with it.

The same organisation can need very different measures at different levels and at different times.

"The right measure, in the right place, at the right time."

And some of that information won't belong to safety

The more I thought about this, the more I realised that some of the information I would want if I were trying to understand an organisation's safety performance wouldn't come from the safety team at all. If maintenance backlog is increasing, I want to know. If employee turnover is rising, vacancies are remaining open or overtime is increasing, that might matter too. The same applies to capital investment, asset condition, defects, rework, production pressure or contractor performance. Depending on the organisation, any combination of those things could tell me something about its ability to continue operating safely.

But I don't think the answer is to turn all of those into safety KPIs or move ownership of them into the safety function. That would simply give us an even bigger safety dashboard and, arguably, make the problem worse. This is where I found myself coming back to the Balanced Scorecard. Its value wasn't that every measure belonged to the same function; quite the opposite. It brought together information from different parts of the business because collectively they helped management understand whether the organisation was working as intended.

Perhaps we need to think about safety in the same way. Instead of asking what business information should be brought into the safety performance framework, perhaps we should ask where safety needs to appear within the organisation's existing performance framework. Safety doesn't need a fifth box. It potentially sits across all of them. Investment decisions can have safety implications. Asset performance can have safety implications. Workforce capability, quality, customer outcomes and operational performance can all intersect with safety. At shareholder and stakeholder level, safety can contribute to a broader understanding of governance, resilience, organisational conduct and long-term value.

That also means the safety function doesn't have to own all the information required to understand whether an organisation is performing safely. In many cases, it probably shouldn't.

Are we actually talking about the same kind of safety performance?

The OECD guidance on Safety Performance Indicators made me reconsider another part of my argument. Its focus is heavily on process safety and major-hazard industries and, initially, I found myself questioning why its discussion of enterprise value seemed largely concerned with preventing costly incidents, protecting reputation and making risk management more efficient. It felt like another attempt to demonstrate that safety was 'good for business' without really integrating the two. On reflection, I think that criticism was partly unfair.

The guidance is concerned with process safety. At that level, detailed technical indicators are exactly what we should expect. If I am operating a chemical plant, I need to know whether the process is within its operating parameters, whether barriers are functioning, whether equipment is performing as intended and whether the controls designed to prevent a major event remain effective. That isn't the same problem as understanding organisational performance, and we shouldn't expect the same framework to solve both.

It also illustrates why I think we need to be clearer about what we mean when we talk about 'safety performance'. At one level there is system or process safety performance. This can be highly technical and, in many environments, overlaps significantly with engineering, operations and quality. I would be comfortable designing the framework through which a chemical engineering organisation manages safety. I would not be comfortable designing the technical controls for its chemical processes. I am not a chemical engineer and I am not competent to do so.

At another level there is the safety management system. ISO 45001 provides the framework here: leadership, planning, competence, operational arrangements, performance evaluation, improvement and the other organisational processes through which OH&S is managed. ISO 45004 then provides more specific guidance on performance evaluation. Then there is what I would describe as organisational safety performance, where the question becomes broader: is this organisation capable of delivering what it is trying to achieve safely?

These things are clearly connected, but I think we mix them up quite readily, particularly when talking about 'integrating safety'. Good process-safety measurement doesn't automatically tell me that the organisation is performing well, just as a good management-system audit doesn't necessarily tell me that the business strategy is sustainable. Nor do I think the answer is simply to aggregate everything upwards. The technical detail needs to remain close to the people competent to understand and act upon it; what needs to travel through the organisation is the information that becomes relevant to decisions being made at the next level.

Which brings me back to strategy

This is probably the part that troubles me most from my own experience. ISO 45001 establishes planning, objectives, performance evaluation and management review as connected parts of the management system. Yet I have seen safety strategies developed with considerable effort, launched, communicated and then gradually disappear into the background while the established monthly reporting continues largely unchanged. Again, I have been involved in developing safety strategies, so I need to be as critical of my own practice here as I am of anybody else's.

If an organisation changes its safety strategy but continues to measure exactly the same things, discusses the same dashboard and makes the same decisions, it is reasonable to ask what has actually changed. The Balanced Scorecard thinking adds another dimension: measures aren't simply there to demonstrate whether the strategy is working. What management learns from performance should be capable of changing the strategy itself. That feedback loop is important, and I don't think we see enough of it in safety.

"If changing the safety strategy doesn't change what we measure, discuss and act upon, what has actually changed?"

Businesses constantly make choices about investment, resources, staffing, utilisation, maintenance, outsourcing and the pace at which work is delivered. Those choices aren't inherently unsafe, and I don't think it helps when the safety profession talks as though there is always an absolute, ideal level of resource or control that the business simply has to provide. In reality, organisations operate within changing parameters.

There will be technical limits that aren't negotiable. Physics doesn't change because a business is under financial pressure. But the organisational arrangements around those limits involve choices. How much preventative maintenance is undertaken? How much redundancy is retained? When is an ageing asset replaced? How much competence is held internally? What level of utilisation is acceptable? How much contingency is built into a programme? Those decisions involve safety, but they also involve finance, operations, engineering, people and strategy.

A reduction in maintenance expenditure might be perfectly manageable. Higher utilisation might be manageable. Increased overtime or a smaller workforce might be manageable. But when several of those things move together, the organisation may gradually consume capacity, resilience or safety margin without any single traditional safety KPI showing it particularly well. That is exactly why I don't think safety performance can be understood entirely from safety information.

And when those signals do emerge, the response isn't necessarily to change the measure, launch another safety initiative or add another action to the plan. It might be the business decision or strategy that needs to change. For me, that is a much stronger test of strategic integration. Safety isn't integrated simply because the safety strategy references the corporate strategy. It is integrated when business decisions influence what we need to understand about safety and what we learn about safety is capable of influencing business decisions in return.

So, is good safety good business?

I still think it probably is. But having started by wondering whether I could demonstrate a straightforward relationship between safety and financial performance, I now think that question is too simplistic. An organisation with good leadership, competent people, reliable assets, appropriate investment, sensible governance and strong operational discipline might reasonably be expected to perform well across several dimensions. Safety, quality, productivity, reliability, customer outcomes and ultimately financial performance may all be influenced by the same underlying organisational conditions.

That creates a causality problem too. If organisations with strong safety performance also perform well financially, it doesn't necessarily follow that safety caused the financial performance. Financial strength might enable greater investment in safety. Both might result from better management. The relationship may work in both directions and may change over time. Perhaps that is why the phrase 'good safety is good business' has always felt intuitively right while being surprisingly difficult to demonstrate cleanly.

Maybe we have been trying to prove the wrong thing. Rather than asking only whether injury rates are improving, whether our leading indicators are green or even whether safety produces a financial return, perhaps leaders should be asking whether the organisation remains capable of delivering its strategy safely, what is changing that might threaten that capability, and whether they are seeing those changes early enough to act.