By Ryan Hodson, Founder, SafeCompli
How difficult is it to make a safety report look successful?
Choose a reassuring comparison. Lead with the improving figures. Give completed training and closed actions plenty of space. None of the numbers need to be false for the reader to come away reassured.
But would the people doing the work recognise that picture?
Imagine reporting a near hit. Two minutes later, a notification tells you it has been investigated and closed. Nobody has spoken to you. Nothing you can see has changed. You are still working with the same equipment, in the same conditions, maybe even still stood in the same location you were when you clicked send on the near miss report, wondering whether anybody understood what you were reporting! This is so demoralising, and it happens more than we would like to admit.
From the management dashboard, that might look efficient. From your end, it might feel rather different.
There will be reports that can reasonably be resolved quickly. Every near hit does not require a site visit and a lengthy investigation. But the response needs to make sense to the person receiving it. Otherwise, what have we taught them about bothering next time?
That gap between what our systems record and what people experience is where measuring success in safety becomes interesting.
What are we actually measuring?
Looking back over my career, I am struck by how straightforward some safety objectives used to be. We wanted a percentage of managers trained, a percentage of operatives trained, inspections completed and incidents reduced by ten per cent.
Report the figures, explain the actions, set next year’s targets.
Those measures still matter. I want people trained. I want inspections undertaken. I certainly want fewer people hurt. But I also want to know whether the training changed how somebody approached their work, whether the inspection found anything useful, and whether the reduction in incidents reflects better controls.
“We haven’t had an accident in five years” is something I have heard delivered with considerable pride.
That may be good news. In a small business with relatively few people and hours worked, it may also be entirely plausible. But it leaves questions unanswered. What gets reported? What counts as an accident? What have people noticed, challenged or stopped before somebody was injured?
A zero is a number. It needs some company before it tells us very much.
The same applies when presenting information to senior leaders. Context helps people understand the data. Selecting a presentation that makes everybody feel comfortable is a different exercise.
If the information is uncomfortable, I would rather see leaders own the situation and decide what it needs. Sometimes that is a straightforward fix. Sometimes it requires investment, additional resources or reconsidering an arrangement that has become difficult to defend.
We should be able to see what decision the information helped them make.
Go and ask somebody doing the job
A report may show that training is up to date and procedures have been issued. Meanwhile, somebody on site is struggling with equipment that should have been replaced, or a new system that makes doing the job properly harder than it needs to be.
You will not necessarily discover that by adding another column to the spreadsheet.
Psychological safety matters here. Can someone explain a difficulty, challenge an instruction or admit a mistake without expecting to be embarrassed or treated unfairly?
One observation I find encouraging is when a conversation barely changes after someone discovers that I work in health and safety. They carry on talking about the work much as they did before. There is no sudden switch to the answer they think the safety person wants. That does not prove everything is working.
It does give me more confidence that I am hearing something useful.
We all know we should involve people in risk assessments, safety meetings and discussions about changes. We can describe those talking points at considerable length, but that’s boring!
The question is whether that involvement has any effect.
HSE’s guidance on monitoring worker involvement includes checking whether people have been encouraged to participate and whether the issues and ideas they raise have led to improvements.
Ask somebody what changed because they spoke up. You will learn more than you would from checking whether their name appears on the attendance sheet.
For near-hit reporting, the feedback loop is particularly important. I have helped develop systems that provide a response to the original reporter. They need to know what happened next, including when the conclusion is that further action is unnecessary.
It need not be War and Peace. Explain what was considered, what will happen and, where appropriate, why nothing further is needed. If the work will take time, say so.
People can understand that replacing equipment or changing an operation takes longer than sending a notification. What is harder to accept is being told something is resolved when their experience suggests otherwise.
Trust takes time to earn. We should be careful about what our systems do with it.
Where does the risk actually sit?
A business can collect thousands of observations and near hits without becoming particularly good at understanding them.
What interests me is the information underneath the total. Are we seeing the same equipment, task, location or working arrangement repeatedly? Is one weakness appearing under different labels across several sites?
And is the important report getting lost among hundreds of routine, low-risk observations?
We need to consider what could credibly have happened, alongside what actually happened. An event without an injury may still expose a serious weakness in the controls.
Equally, ask leaders about their biggest risks and they may understandably describe the most dramatic possibilities. The data may show repeated manual-handling injuries, slips or occupational-health exposures causing substantial harm over time. Both perspectives deserve attention.
AI can help examine large volumes of information and identify possible connections. Used carefully, that can be valuable. Sensitive information needs protection, and the findings need checking. A pattern involving one person might reflect their workload, equipment or exposure rather than anything peculiar to that individual.
The point is to help us decide where to look and what to question. Producing more analysis is useful only if it improves the decisions that follow.
Then we need evidence of what those decisions achieved – the key is verification of the claim.
After a significant finding, one practical approach is to put relevant questions into routine inspections and audits. Does this issue apply where we are today? If so, are the controls present, understood and working?
That checks the original response and whether the learning reached other places where it was needed. Return later, too. Does it still work when the workload changes or the person who introduced it has moved on?
“Nothing to report this period” can contribute to that picture, provided we understand what somebody checked before saying it.
Success should also prompt curiosity. When difficult work goes well, what helped? Can we repeat those arrangements, or did an experienced person quietly compensate for a weakness?
What would convince you?
If a board asked me how we know we are succeeding in safety, given the type of character I am, I would probably turn around and say that you’re asking the wrong person, and I would wait in silence to see their response (because, in all truthfulness, they probably are asking the wrong person). I would want operational managers to describe improvements their teams had made and how they knew they worked.
By the time a concern reaches the head of health and safety, it may already have escalated considerably. It is more telling when useful responses happen throughout the business.
I would also suggest asking someone who joined three months ago how this workplace compares with their previous experience. Ask a longer-serving employee what has improved and what still makes safe work difficult.
Listen across roles, sites and shifts. Compare those accounts with incidents, exposure monitoring, inspections and verified controls. HSE’s performance guidance supports considering preventive information alongside incident data and checking effectiveness.
Certificates, awards, reduced lost time and lower claim costs all contribute. They do not answer everything.
Before presenting the next reassuring report, take it out onto the shop floor.
Would the people doing the work recognise the business it describes?
About the Author: Ryan Hodson
Ryan Hodson is the Founder of SafeCompli, a consultancy helping businesses take a practical, common-sense approach to health, safety, quality and compliance. A business guided by one simple belief: nobody should be priced out of working safely.
SafeCompli: Your QHSE & Compliance Department, Without the Overhead.
References
Health and Safety Executive (2026). Monitor performance and review progress. Available at: https://www.hse.gov.uk/involvement/consult/monitorperformance.htm
