Job Hazard Analysis: A Step by Step Guide for Manufacturing Teams

A job hazard analysis is a structured method for breaking a task into its individual steps, identifying the hazards present at each step, and deciding on controls before the work begins. It is sometimes called a job safety analysis, and the two terms describe the same exercise. The output is a short, specific document that says what could go wrong during a particular job and what will be done about it.

The value of a job hazard analysis lies in its scope. A site risk assessment considers the plant. A job hazard analysis considers the person standing in front of the machine at two in the morning with a spanner in one hand. That narrowing is what makes it useful, and it is also what makes it easy to do badly, because a task that everybody performs every day is the hardest task to see clearly.

When a Job Hazard Analysis Is Needed

Not every task warrants one, and a site that tries to analyse everything will produce a filing cabinet rather than a safer operation. The jobs that repay the effort share a few characteristics:

• Tasks with a history of incidents, near misses, or first aid cases
• Non routine work such as maintenance, cleaning in place, changeovers, and confined space entry
• Jobs where a single error could cause serious harm, regardless of how rarely it has happened
• New processes, new equipment, or tasks whose procedure has recently changed
• Work that operators frequently perform differently from the way the procedure describes

That last one is the most revealing and the most ignored. Where the written method and the actual method have drifted apart, the hazard sits in the gap between them.

Step One: Choose the Job and Set Its Boundaries

A job hazard analysis needs a clear start and a clear end. “Operating the filling line” is too broad to analyse. “Changing the filling nozzle between batches” has a beginning, an end, and a manageable number of steps. Where a task runs longer than around ten steps, it is usually two tasks wearing one name, and splitting it produces better analysis than compressing it.

Step Two: Involve the People Who Do the Job

The operator who has performed the task four hundred times knows things the procedure does not record. Which guard has to be lifted because the interlock sticks. Where the tool is actually kept. What everybody does when the line is running behind and the correct method takes six minutes longer.

None of that information arrives if the analysis is conducted by a safety adviser at a desk. It arrives when somebody watches the job being done and asks about the parts that look improvised. Poor communication in manufacturing is why so many hazard analyses describe a version of the job that nobody performs.

Step Three: Break the Job Into Steps

Each step describes one action, phrased as a verb: isolate the supply, release residual pressure, remove the guard, lift the nozzle clear. Steps describe what happens, and they leave out how well or how safely it happens, because judgement belongs in the next column.

Ten steps is a reasonable ceiling. Fewer than four usually means the analysis has skipped something, and the skipped part is often the preparation or the return to service, which is where a surprising share of injuries occur.

Step Four: Identify the Hazards at Each Step

For every step, the question is what could cause harm, to whom, and by what mechanism. Stored energy, moving parts, chemical exposure, temperature, height, manual handling, noise, and access are the categories that recur in manufacturing. Environmental hazards belong here too: a spill during nozzle removal is a hazard whether or not anybody is injured by it.

The discipline is specificity. “Chemical exposure” is a category. “Residual caustic in the line discharges toward the operator’s face when the coupling is released” is a hazard, and only the second version tells anybody what control is required.

Step Five: Decide the Controls

Controls are chosen in order of preference, and the order matters because auditors and, more importantly, outcomes both depend on it. Eliminate the hazard if the job can be redesigned to remove it. Substitute a less hazardous material or method. Apply engineering controls that work without anybody remembering to use them, which is the logic behind poka yoke and every physical interlock ever fitted. Then administrative controls such as procedures, permits, and training. Personal protective equipment sits last, because it protects one person, only while worn correctly, and only until it fails.

A job hazard analysis that reaches PPE at every step has documented the hazards without controlling them.

What the Finished Document Looks Like

A job hazard analysis is conventionally recorded in three columns. Keeping it to three is deliberate, because the moment a form grows to nine columns it becomes something people fill in afterwards.

The rows above are illustrative. The point they carry is that a good control is testable. Somebody can walk up during the job and see whether it is in place.

Step Six: Put It Where the Work Happens

An analysis that lives in a folder in the office has controlled nothing. The findings need to reach the people doing the job, at the time they do it, which usually means the procedure is updated, the training is refreshed, and the specific hazards are raised at the shift briefing before the task begins.

Handover deserves particular attention. Where a job spans a shift boundary, the hazards travel with the incomplete work. Half of what makes poor shift turnover dangerous is that the incoming crew inherits a condition, an isolation, or a partially reassembled machine without inheriting the analysis that described its risks.

Step Seven: Review It Before It Goes Stale

A job hazard analysis is a snapshot of a job on the day it was analysed. Plants change. Equipment gets modified, suppliers change formulations, and procedures accumulate workarounds.

Review is triggered by an incident, a near miss, a change to the equipment or method, or the passage of time, and the review after an incident is the one that matters most. Where the analysis missed a hazard, the correction belongs in the document, and structured root cause analysis is what turns an incident into that correction rather than into a memo about being careful.

Making review routine is a leadership problem more than an EHS problem. Where leader standard work includes verifying that controls are present during high risk tasks, the analyses stay current because somebody looks at them weekly. Where it does not, they age until an auditor finds them.

Where Job Hazard Analysis Fails

It fails when it is written to satisfy a requirement rather than to inform a task, and the tell is uniformity. Twenty analyses that name the same hazards and prescribe the same controls describe a template rather than a plant.

It fails when the operator was never asked, which produces a document about the official job instead of the real one.

It fails when the controls are not verified. A 5S audit that finds the jib unavailable and the catch pot full tells a manager that two of the controls above existed only on paper.

And it fails when nobody can find the analyses. Hazards recorded in local spreadsheets, filed differently at each site, cannot be compared, trended, or learned from across a group.

Run Job Hazard Analysis Where the Job Is

EviView is a digital daily management system for regulated manufacturing. Hazards and near misses are raised at the point of work rather than reported after the shift, actions are assigned with an owner and a due date, root cause analysis is captured against the incident that prompted it, and audit dashboards show which controls have been verified and which are overdue. Tier boards carry the outstanding safety actions into the same daily conversation as quality, delivery, and cost, which is the only place they compete for attention on equal terms.

To see how job hazard analysis findings stay live instead of ageing in a folder, book a demo with the EviView team, and reach out with the tasks that worry you most so the conversation starts on your own plant.