Practically every industrial organisation has safety training, and measures whether people complete it. Sooner or later the same question comes up in all of them: why do we keep seeing the same incidents?
The cause rarely sits with the workforce. They are not unwilling, not stupid and not careless. It sits with training built to be ticked off rather than to change behaviour.
These are the five patterns that give it away.
1. Completion is your only measure of success
If the only thing you measure is whether somebody finished the training, you are measuring the wrong thing. Completion says nothing about understanding, nothing about changed behaviour and nothing about what somebody does when it really counts.
An employee who clicks “next” through twenty slides of ATEX procedures has completed the training. So has the employee who afterwards picks up the wrong tool in an ATEX zone.
Effective safety training measures behaviour: decisions in scenarios, recognising situations before they escalate, the right response under pressure. That is measurable, but it demands a different approach than a completion record.
2. The same incidents keep coming back, framed differently
If your incident register shows patterns, the same kind of mistake, the same step skipped, the same situation turning out differently than expected, that is a signal the training does not address that situation.
It sounds obvious. In practice, though, the link between incident analysis and training content is weak. Training is tracked in the LMS, incidents in a separate system, and the analysis of “what could training have changed here?” rarely happens in any structured way.
Good safety training is built on incident data and near-miss analysis. The scenarios people rehearse are the scenarios that can occur in the installation, not generic textbook situations.
3. Training is detached from the workplace
Classroom training in a meeting room, e-learning on an office laptop, a simulator at a training centre three hours down the road. All of these formats share one problem: the distance to the workplace is too great.
Transfer, the extent to which what you learn in training also gets applied for real, depends directly on how closely the training environment approaches the real one. A generic simulation of a chemical installation transfers poorly for the operator working on one specific installation in Rotterdam.
Training that works is built on the real installation, the real procedures and the real risks. Not as a nice extra, but as a basic requirement.
4. People know that they have to train, not why
Ask somebody why they took the LOTOTO training. The answer is almost always “HR required it” or “I had to renew my certificate.” Rarely “Because I want to understand how I protect myself and my colleagues during this kind of maintenance.”
That is not a culture problem, it is a training design problem. Training offered as a compliance obligation, with no context about why the knowledge matters, treats people as administrative units. They behave accordingly.
Effective training starts with why. Not as a motivational speech, but as a substantive part of the training: what can go wrong, what the consequence is, how you prevent it. Concrete case material from your own sector works better for that than abstract statistics.
5. The training calendar drives it, the risk does not
If your safety training sits on a fixed calendar, annual refresher training, quarterly checks, mandatory modules on contract renewal, then that calendar is probably based on regulatory requirements rather than on risk.
That is understandable, because regulation dictates frequency. But it produces a system in which somebody who lived through an incident three months ago does not repeat the training for another nine, while somebody who has performed a procedure faultlessly for years takes the same training as somebody just starting.
Risk-driven training looks at it differently: who carries which risk, what has changed in the installation or the procedure, where have near-misses been recorded? Those answers drive who trains on what and when, rather than the calendar.
What does work
The five patterns above share a cause: training designed for compliance rather than for competence. Those two goals are not mutually exclusive, but they demand different approaches.
Training that changes behaviour has four characteristics:
- Specific: built on the organisation’s real installation, procedures and risks
- Practice-led: people act, decide and make mistakes in a safe environment
- Measurable: behaviour gets measured, not only completion
- Continuous: knowledge is kept active through spaced repetition, not polished up once a year
That asks more than an LMS and a catalogue of e-learning. It also delivers more: fewer incidents, higher competence, and training your people understand the reason for.
Those four requirements are hard to meet with separate trainings and a good deal easier once there is a structure underneath. How to build one is in our guide on setting up or renewing a corporate academy. What a practice-led setup looks like in the field is under scenario training. The generic-simulation example comes from chemical and process.