Green Doesn't Mean Safe: What is your safety dashboard actually telling you?

I've sat through more safety performance reviews than I care to remember. Some are genuinely useful, while others involve working through endless slides of statistics, completed actions, training figures and colourful dashboards. Inevitably, there's a lot of green, which is reassuring, isn't it? But I've often wondered what we're actually reassuring ourselves about. If an organisation has completed 98% of its safety inspections, delivered all its mandatory training and closed 95% of its actions, that's good news, but does it mean the organisation is any safer? Possibly, possibly not. It tells us that we've done what we said we were going to do, but what it doesn't necessarily tell us is whether any of it has made a difference. I think that's where we sometimes get ourselves into trouble, confusing the delivery of safety management activity with evidence that risks are being effectively controlled.

If it ain't broke...

There's a tendency in organisations to assume that if things are going well, there's no particular reason to change anything. We haven't had a serious incident, the numbers look good and our performance is green, so we must be doing something right. Maybe we are. There are plenty of organisations that manage their risks extremely well, and I certainly wouldn't suggest that every green dashboard is concealing some terrible problem. But how do we know? I've heard people use the absence of serious incidents as evidence that existing arrangements are working, and I've seen organisations celebrate completing programmes of activity without really establishing whether they've achieved the improvements they were supposed to deliver. I think there's a danger that we become more interested in demonstrating that we're managing safety than understanding how effectively we're managing risk. Once that positive performance narrative becomes established, it can be surprisingly difficult to challenge, particularly when nobody has experienced anything serious enough to question it.

“There's a danger that we become more interested in demonstrating that we're managing safety than understanding how effectively we're managing risk.”

The green light culture

Nobody necessarily sets out to mislead anyone, but organisations are made up of people, and people respond to expectations. If positive reports are welcomed and negative reports attract difficult conversations, it's hardly surprising that people become quite good at presenting things positively. There's also the question of career progression. People want to be seen as successful, they want their departments to perform well, and they generally want their bosses to have confidence in them. That doesn't mean they're dishonest, but it can influence what gets emphasised, what gets challenged and what gets left out. Over time, a rather comfortable cycle can develop. No serious incidents means confidence in the existing arrangements, confidence encourages reassuring reports, reassuring reports reduce the appetite for challenge, and without challenge, weaknesses can remain unnoticed. Until something happens.

Of course, there's another dimension to this. A green dashboard can provide a certain amount of perceived protection for those receiving it. If everything was reported as satisfactory, how were they supposed to know there was a problem? I'm not suggesting that this necessarily provides any legal defence, but psychologically, I think it's a very comfortable position to be in. It's easier to point towards an established reporting process than acknowledge that perhaps we didn't ask enough questions. In some respects, the dashboard becomes evidence that governance has taken place, rather than evidence that risks are being effectively managed.

Safety professionals aren't innocent either

It's easy to blame senior leaders for not asking the right questions, but I think we need to look at ourselves as safety professionals too. I've certainly made mistakes in how I've presented concerns. Earlier in my career, I sometimes painted a more negative picture than was necessary because I wanted to secure attention, resources or support for an improvement. The problem is that if everything is presented as urgent, eventually nothing is. You can become the boy who cried wolf, and when something genuinely significant needs attention, people may already have become sceptical. I've learned that establishing credibility and building relationships is just as important as understanding the technical issues. When I've been most successful in influencing senior leaders, it's generally been because I've explained my professional judgement, how I've arrived at it, what I think needs to change, what resources are required and how we'll know whether the changes have worked.

Sometimes leaders have challenged my evidence, and that's not necessarily a bad thing. They should challenge it. The important distinction is whether they're challenging it to understand the problem or to avoid having to deal with it. Executives aren't expected to be experts in every aspect of safety, but they should create an environment where expertise is respected, difficult conversations are possible and people feel able to challenge established thinking. Equally, safety professionals need to recognise that technical knowledge alone doesn't automatically make their advice credible or their recommendations proportionate. Effective scrutiny requires trust, professional competence and a willingness to listen on both sides.

Are we asking the wrong questions?

I've recently spent some time revisiting research on safety performance measurement, particularly Andrew Hopkins' work on process safety indicators, research examining the relationship between leading and lagging indicators on a major construction project, and the guidance provided by ISO 45004. What struck me wasn't simply the difficulty of finding the right measures, but how easily we can convince ourselves that a measure tells us more than it actually does. We often describe inspections, training and completed actions as leading indicators, but leading indicators of what? They might demonstrate that a safety management process is operating, but that doesn't automatically establish whether the controls intended to prevent serious harm are effective. There's an important distinction between measuring the activity of a safety function and measuring how well an organisation manages its risks.

Hopkins' work particularly reinforces this distinction. An organisation might have encouraging personal safety statistics while significant weaknesses remain in the systems and controls intended to prevent a catastrophic event. That's not an argument for abandoning conventional measures, but it does suggest that we need to understand precisely what each indicator can and cannot tell us. It also raises an uncomfortable question about incentives. When performance measures become targets, particularly where reputation or remuneration is involved, are we encouraging people to manage the risk or manage the indicator? I don't think the answer is as straightforward as removing every safety-related target or incentive. In organisations where safety is already struggling to compete for attention, that could create a different problem. But we should at least recognise the potential for unintended consequences.

The construction research also prompted me to question how confidently we describe indicators as predictive. Just because something happens before an incident doesn't necessarily mean it reliably predicts one, and the relationship between organisational activity and eventual outcomes is rarely straightforward. Equally, I think we sometimes underestimate the value of looking backwards. If similar incidents continue to occur after investigations have been completed and actions closed, surely that tells us something about the effectiveness of our management arrangements? Perhaps the more useful distinction isn't leading versus lagging, but whether the information helps us identify weaknesses, understand their significance and do something about them.

ISO 45004 offers a useful approach to performance evaluation, but there's another issue that I think receives insufficient attention: organisational capability. Developing a meaningful framework requires an understanding of the organisation's risks, the limitations of available data and the competence to interpret the findings. It also requires somebody to translate that analysis into information executives can use without stripping away everything that makes it meaningful. I suspect many organisations are better equipped to produce dashboards than to critically evaluate what those dashboards are telling them. And if the people preparing the information don't fully understand its limitations, how can we reasonably expect the board to understand them?

“I suspect many organisations are better equipped to produce dashboards than to critically evaluate what those dashboards are telling them.”

Red isn't necessarily any better

Here's another observation. I've seen organisations where a completely red dashboard produces much the same response as a completely green one. Nothing changes. With green, everyone assumes things are working, while with red, poor performance can become so familiar that people simply accept it. Everyone loves being a firefighter when something goes wrong. It's visible, urgent and often attracts resources, but the less glamorous work of anticipating problems, questioning assumptions and making improvements before an incident occurs doesn't always receive the same attention. So perhaps the issue isn't really the colour of the dashboard at all, but what we do with the information. If the same concerns appear month after month without meaningful intervention, what exactly is the reporting process achieving?

Making green mean something

I don't think we need to throw away dashboards or develop another complicated set of performance measures, and I certainly don't think every executive needs to become a safety specialist. What I'd like to see is a little more curiosity. When the dashboard is green, ask what evidence supports that assessment. When it's red, ask what's being done differently. When an activity is reported as complete, ask whether it achieved what was intended. And when someone tells you that everything is going well because there haven't been any serious incidents, perhaps ask how they know that isn't simply good fortune. There are organisations that genuinely manage their risks well, and you can often recognise them by the quality of the professional scrutiny and constructive challenge that takes place, even when performance is good.

The objective isn't to make everything red or to find fault for the sake of it. It's to have confidence that the information we're relying on means something. We need to be interested not just in what the dashboard says, but in how those conclusions were reached, what they tell us about the organisation and whether they're influencing the decisions being made. After all, a performance framework that doesn't inform decisions or lead to improvement is little more than an elaborate reporting exercise.

So next time someone presents you with a dashboard, don't just look at the colours. Ask a few questions. You might learn considerably more from the conversation than the dashboard itself.

References and further reading

• ISO 45004:2024 — Occupational health and safety management: Guidelines on performance evaluation.

• Hopkins, A. (2007) — Thinking About Process Safety Indicators.

• Leading or Lagging? Temporal Analysis of Safety Indicators on a Large Infrastructure Construction Project.