Unsafe Acts vs Unsafe Conditions: Spot and Report Both

An unsafe act is a behaviour that creates risk — bypassing a machine guard, skipping PPE, texting while reversing a forklift. An unsafe condition is a hazardous state of the workplace itself — a blocked fire exit, frayed wiring, an oil-slicked floor. The distinction matters because the fixes differ: conditions are corrected by repair, design, and housekeeping, while acts are corrected by coaching, supervision, and — most often — by fixing the condition that made the act tempting. A good reporting programme captures both, without turning act reports into blame.

Two categories, one incident

Walk through almost any incident and you'll find both types tangled together. A worker steps over a spill (act) that had sat unmopped for an hour (condition) because the spill kit was empty (condition) and the shift was short-staffed (organisational condition), and a supervisor had walked past it too (act). Pull one thread and the others come with it.

That's the first practical lesson: the categories are a sorting tool for observation and fix selection, not a verdict on whose fault something was. You classify what you saw so you know which corrective lever to reach for — not so you know whom to blame.

Examples worth training people on

Abstract definitions don't change what people notice on the floor. Concrete examples do. Use these as a starting set and replace them with examples from your own operation within the first quarter.

Common unsafe acts

• Operating equipment without training or authorisation.
• Removing, bypassing, or defeating a guard or interlock.
• Not wearing required PPE, or wearing it incorrectly.
• Using a phone while driving, operating machinery, or walking through vehicle routes.
• Standing or walking under a suspended load.
• Improvised working at height — chairs, pallets, racking as ladders.
• Rushing manual lifts, or lifting loads clearly beyond safe capacity.
• Leaving blades, hot equipment, or chemicals unattended in shared areas.

Common unsafe conditions

• Wet, greasy, damaged, or obstructed floors and walkways.
• Blocked or locked emergency exits and fire equipment.
• Missing or damaged machine guards.
• Exposed wiring, damaged cords, overloaded sockets.
• Poor lighting in work or transit areas.
• Overloaded, damaged, or unsecured racking and shelving.
• Defective tools or vehicles still in service.
• Missing labels on chemical containers.

Why does the distinction change the fix?

Because each category responds to different controls, and applying the wrong one wastes effort.

Two failure patterns dominate. For conditions, it's the sign instead of the fix — a "mind the step" notice where a repaired step should be. Signage is a holding measure, not a control. For acts, it's retraining as reflex — sending someone back through a course they passed last year, when the real cause is a production target that can't be hit with the guard in place.

The trap in "unsafe act": stopping at the person

Here is the uncomfortable pattern experienced safety people learn to expect: investigate an unsafe act honestly and you usually find an unsafe condition — physical or organisational — standing behind it. The operator bypassed the guard because it jams four times an hour. The team skips harnesses because retrieving them adds fifteen minutes and the schedule doesn't allow it. The new hire copied exactly what the veteran next to them did.

So when an act is observed, ask what makes this act make sense to this person today? before reaching for discipline. Reserve consequences for genuine recklessness — impairment, sabotage, deliberate repeat violations after fair correction — and treat everything else as diagnostic information about your systems. This is the core stance behind observation programmes, explored further in our guide to behaviour-based safety.

None of this means acts go unaddressed. It means the conversation happens in the moment, respectfully, and the report feeds a fix rather than a file.

Building the reporting habit for both

Conditions are the easy half: they hold still. Routine workplace safety inspections catch most of them, and any team member should be able to raise one between inspections in under two minutes, ideally with a photo.

Acts are harder because they're momentary and because reporting a colleague feels like informing. Three design choices soften this:

1. Report the behaviour and situation, never the name — "guard bypassed on line 2 during changeover" is actionable without being an accusation.
2. Encourage self-reporting by treating an honestly reported own mistake as a contribution, not a confession.
3. Fold act-spotting into structured, scheduled observations by trained observers, so it's a normal activity rather than surveillance.

Close calls sit across both categories, and the same no-blame machinery applies — a mature near-miss reporting programme will capture unsafe acts and conditions that never quite became incidents.

A one-week exercise to calibrate your team

Try this before building anything elaborate. For one week, ask every supervisor to note three things per day: one unsafe condition, one unsafe act, and one safe act worth reinforcing. No names on the act reports. At the end of the week, sort the pile.

Most teams discover the same things: conditions cluster in a handful of neglected areas; the acts are nearly all rational responses to pressure or awkward equipment; and praising the safe acts was the part supervisors found most unnatural — which tells you something about the prevailing culture. The sorted pile becomes your first corrective-action backlog, already prioritised by frequency.

Tracking both types without a paper mountain

The follow-through is where programmes sink: condition reports need work orders, act patterns need trend visibility, and both need closure evidence. Task10x supports this with ad-hoc report templates any team member can fill from a phone browser with photos, scheduled safety inspections per location, and automatic corrective actions — assigned, due-dated, and tracked to closure with photo proof — plus dashboards showing open actions by location so recurring hazards stand out. How this fits frontline teams is outlined on the use-cases page.

Teach the distinction, publish local examples, protect reporters, and fix conditions fast. The acts largely take care of themselves once the conditions stop inviting them.