This article was originally published on LinkedIn on 7 December 2025. It has been moved to the riskfacilitator Insights library so the website remains the permanent source.[1]
Primary sources used to verify and contextualise this archive: [2]
A recent post by Ben Hutchinson (PhD) resurfaced a powerful idea from Don Norman’s The Design of Everyday Things, and it’s worth pausing on. Not because the concept is new, Norman wrote it years ago, and Trevor Kletz wrote it decades before that, but because our industry still hasn’t fully absorbed the lesson.
We continue to talk about “human error” as if it is a meaningful explanation for why things go wrong. It isn’t. It never was. And clinging to it holds organisations back from real improvement.
When error rates are high, the problem isn’t people, it’s design
Norman challenged the often-quoted claim that 75–95% of accidents are caused by human error. He argued that if error were genuinely responsible for that many failures, the underlying issue must be something far more systemic:
If humans were really at fault that frequently, the design itself is broken.
In other words, “human error” is not a cause. It is evidence, a signal that the system is mismatched with human capability, context, or the realities of operational work.
This reframing is one of the most important conceptual shifts in contemporary risk, safety, and human-centred design.
Why our response to failure reveals the real problem
Norman highlights a contradiction that anyone in safety will recognise.
- When a bridge collapses, we analyse the engineering, question our assumptions, and update the standards so it doesn’t happen again.
- When a person slips, presses the wrong button, misjudges timing, or misses a cue… we blame the individual and carry on as if the broader system needs no change.
The first response drives learning. The second achieves nothing.
This is the legacy of treating humans as the cause of failure rather than the source of adaptability that allows systems to function at all.
Trevor Kletz warned us decades ago: “Try to change situations, not people.”
Kletz understood something essential, no operator works in isolation. Behind every action are layers of:
- managerial decisions
- design decisions
- environmental constraints
- competing objectives
- resource limitations
- time pressures
- ambiguous procedures
- tools that may or may not match the task
- assumptions about how work is “supposed” to be done
Operators are simply the last link in a long decision chain.
Kletz observed that everyone in that chain can make errors, yet the person at the sharp end is often the only one held accountable. And by blaming them, we lose the opportunity to see upstream influences that created the conditions for failure.
This is why the blanket term “human error” is so unhelpful. As Kletz put it: Saying an incident was caused by human error is about as informative as saying a fall was caused by gravity.
True, but it tells us nothing about what needs to change.
Modern safety thinking has caught up, but practice is lagging
Hollnagel, Dekker, Conklin, and other modern thinkers have reinforced what Norman and Kletz knew intuitively:
- People do not cause most failures.
- People prevent failure constantly by adapting, adjusting, and bridging gaps in the design.
- Variability is not the enemy, it is the defining feature of real work.
- Work-as-Done is the only accurate representation of reality.
- Systems must be designed with human performance variability in mind.
This is the foundation of Safety-II and Resilience Engineering: We must study why things go right far more often than why they go wrong. The same adaptations that lead to success can, under different pressures, contribute to incidents.
The variable is not human competence. The variable is the context.
Design is the real leverage point for change
If we accept that human error is mostly a symptom, not a cause, then the question becomes:
How do we design systems, tools, environments, and workflows that support human performance instead of undermining it?
This requires us to think deeply about:
1. Interfaces and cues
Are they intuitive? Are they aligned with human perception and cognition?
2. Workflows and procedures
Do they reflect Work-as-Done or just Work-as-Imagined?
3. Environmental factors
Lighting, noise, distractions, time pressure, fatigue - all shape performance.
4. Decision-making demands
How many decisions are we forcing onto operators, and are they reasonable given the pace and complexity of work?
5. Systems of safety
Are controls coherent and integrated, or are they siloed and reactive?
6. Organisational priorities
Do we unintentionally reward speed over safety? Driving outcomes faster than the system can realistically support.
These are the areas with true leverage. This is where progress happens. This is how we prevent repetition.
Where leaders must shift their mindset
If organisations want fewer incidents, fewer surprises, and stronger predictability, they need to embrace a central truth:
Human performance is shaped by design. People can only perform as well as the system allows them to.
This means leaders must stop asking: “Who made the mistake?”
And start asking: “What in the system made this mistake possible?” “What made it likely?” “What helped this person succeed the other 364 days of the year?”
Blame protects the status quo. Design improves it.
A mature approach to risk and safety
Human error will always exist because humans will always exist. But error is predictable, manageable, and designable-for.
Stronger guardrails, clearer cues, better interfaces, more realistic procedures, smarter systems of safety, and an honest understanding of Work-as-Done all play a role.
The future of safety is not about eliminating error, it’s about designing systems where error does not lead to harm.
This is where the industry needs to go. And the more we challenge the mythology of “human error,” the faster we’ll get there.
References
- Paul Chivers, Human Error Isn’t the Cause, It’s the Clue. Why We Need to Redesign the Systems We Expect People to Operate, LinkedIn, originally published 7 December 2025.
- Safe Work Australia, Good work design, accessed 10 August 2026.