Working on Heights Safety: Your 2026 Australian Guide

Monday morning on a live site usually looks the same. A delivery is late. One subcontractor has turned up with the wrong gear. Another says their SWMS is on the ute. The client wants progress photos by lunch. At the same time, someone needs to get onto a roof, a mezzanine edge, a scaffold deck, or an elevated platform to keep the job moving.

That is where working on heights safety either holds up or falls apart.

The paperwork matters, but the true test is what happens when trades overlap, access changes by the hour, and production pressure starts pushing people toward the quickest option instead of the safest one. In Australia, that pressure sits on top of a stubborn problem. Approximately 7,800 serious injury claims in 2023 to 2024 were specifically attributed to falls from height, and over the past decade falls have accounted for around 25 to 30% of all serious workers' compensation claims according to WAHA's summary of Safe Work Australia data.

Most sites do not fail because nobody knew the rules. They fail because the controls were too generic, the equipment register was spread across phones and folders, or nobody pinned down who was responsible once three subcontractors were working in the same area. That is the gap this guide addresses.

Establishing Your Working at Heights Framework

A workable framework starts before the first person clips onto anything. If the first genuine discussion about height risk happens during the pre-start, you are already behind.

The legal side is clear enough. Under Australia's WHS framework, duty holders must manage risks using the hierarchy of controls. On paper, that sounds straightforward. On a live construction or industrial site, it gets messy fast because access routes shift, workfaces change, and one trade's temporary fix becomes the next trade's exposure point.

Start with who owns the risk

On multi-contractor sites, confusion about responsibility is common. The principal contractor may control the site, but each PCBU still carries duties for the work they direct, the plant they bring, and the workers they put into the task.

That means you need written clarity on:

• Work area control: Who signs off the area before height work starts.
• Access equipment ownership: Who supplies, tags, inspects, and removes EWPs, ladders, harnesses, static lines, nets, and temporary edge protection.
• Interface risks: Who manages overlapping trades below or beside the task.
• Rescue readiness: Who provides the rescue kit and who is trained to use it.

If those points are left vague, the site ends up relying on assumptions. Assumptions do not hold up after an incident.

Build the risk assessment around the actual task

A lot of site documents describe hazards in broad terms. "Fall from height" is not a risk assessment. It is a category.

A useful assessment breaks the task down into what workers will really do. For example:

1. How they access the area Roof hatch, internal stair, scaffold stair, EWP basket, fixed ladder, temporary platform.

2. Where the exposure starts Roof edge, brittle surface, incomplete deck, penetration, open stair void, loading platform, conveyor gantry.

3. What changes during the job Material movement, weather, shift handover, power lead routing, removed handrails, mobile plant movement, lighting changes.

4. What can go wrong during recovery A worker can be protected during the task but still exposed during setup, repositioning, or pack-down.

That level of detail matters because most failures happen in transitions, not in the neat version of the task written in the office.

Tip: If your risk assessment reads the same for roofing, plant maintenance, and façade work, it is too generic to manage working on heights safety properly.

Make the framework usable on site

The strongest framework is one supervisors can use in two minutes at the workface. It should tell them what to check, what cannot proceed, and who has authority to stop the task.

I look for three things in any site system:

Digital systems help when they replace confusion, not when they add another layer. A decent health and safety management system should let you tie the risk assessment, permits, inspections, and subcontractor records to the same task so the supervisor is not chasing five versions of the truth.

What a defensible starting point looks like

A sound framework is not a thick procedure manual. It is a repeatable way to answer a few hard questions before work starts:

• Can this task be done without leaving the ground?
• If not, what is the highest-level control that is practical here?
• Who checks the setup before the worker is exposed?
• What changes would stop the task immediately?
• If someone falls or becomes suspended, who does what next?

If your team can answer those questions quickly and consistently, you have a framework that works on site, not just in an audit folder.

Applying the Hierarchy of Controls in Practice

The hierarchy of controls gets repeated on every induction, but sites still jump straight to harnesses. That is usually where the trouble starts.

A harness feels like action. It is visible, easy to issue, and simple to write into a SWMS. But in working on heights safety, it is often the last line of defence, not the first decision. The better question is always this: how do we stop the worker getting into a fall position in the first place?

Elimination first, even when the program is tight

Elimination is not theoretical. It means changing the job so the work happens from the ground or from a safer permanent position.

Examples that work in the field include:

• Pre-assembly at ground level: Build frames, handrail sections, pipe spools, or service brackets before lifting them into place.
• Remote inspection methods: Use camera poles or drones where inspections do not need a person exposed at height.
• Design changes: Move plant controls, valves, or service points to an accessible level during install instead of accepting future maintenance at height.

These decisions often need to be made early. Once the structure is up and the program is running hot, crews stop looking for elimination options and start asking which harness to issue.

Substitution and engineering controls do the heavy lifting

Substitution means replacing a risky method with a safer one. Engineering controls mean changing the physical setup so the hazard is contained.

That usually looks like this on Australian sites:

The reason supervisors should fight for these controls is simple. They do not depend on one worker clipping on correctly, choosing the right anchor, adjusting lanyard length, and maintaining clearance every minute of the task.

A practical summary of the hierarchy of control measures is useful, but the main site rule is even simpler. If the chosen control still leaves the worker exposed to falling, keep looking higher up the hierarchy.

Key takeaway: The cheapest control at purchase is often the most expensive control to supervise.

Administrative controls matter, but they cannot rescue a weak setup

Administrative controls include permits, SWMS, sequencing, spotters, exclusion zones, and supervision. They matter because even good physical controls can be defeated by poor coordination.

On mixed-trade sites, I want admin controls to answer a few basic questions:

• Who is allowed into the area below the work
• What plant movements are restricted during the task
• What weather trigger stops the job
• Who checks the setup after lunch, shift change, or a break in the task

Admin controls are useful when they support solid engineering. They are weak when they are expected to compensate for missing edge protection or poor access planning.

PPE is not the plan

PPE includes full body harnesses, lanyards, inertia reels, helmets, and connected components. It has a place, but relying on PPE as the primary control usually creates more management work and more failure points.

Before signing off on PPE as the main control, check these issues:

• Anchor suitability: Is there a rated anchor in the right location for the way the worker will move?
• Clearance: If a fall occurs, will the worker hit a lower level, structure, or plant?
• Swing fall exposure: Can the worker pendulum into steel, cladding, or an edge?
• Rescue time: Who retrieves the worker, with what kit, and from where?

If those answers are weak, the system is weak.

The trade-off nobody likes discussing

Higher-level controls can look slower during planning. Guardrails need installation. Nets require design, timing, and coordination. EWPs need space and traffic planning. Drones need competent operators and a suitable inspection method.

But those controls usually save time once work starts because they remove constant stop-start supervision. You are not checking every individual clip-on point or arguing over whether a ladder is good enough for just one quick task.

The sites that manage working on heights safety well do one thing consistently. They make it harder to start the task with the wrong control.

Selecting and Managing Fall Protection Equipment

Equipment decisions get muddled because sites often mix up fall prevention and fall arrest. They are not the same, and treating them as interchangeable is where many jobs come unstuck.

Fall prevention stops the worker reaching a place where they can fall. Fall arrest accepts that a fall may happen and tries to stop it becoming fatal. If you choose arrest when prevention was possible, you have already accepted more risk than necessary.

Prevention versus arrest on real sites

This is the simplest way to separate them.

On a manufacturing site with fixed plant, prevention often wins because the environment is stable. You can install permanent walkways, handrails, gates, and anchor systems tied to known maintenance points.

On house builds and other dynamic residential work, the answer changes daily. Frame stages move fast, subcontractors rotate, and edges appear and disappear. That is exactly why prevention needs to be considered early, not dropped because the site is changing.

Subcontractor gear is where control breaks down

One contractor brings harnesses with current tags. Another turns up with faded webbing and no clear record. A third has compatible lanyards but no agreed inspection history. If nobody controls that mix, the principal contractor ends up with exposure whether the gear belongs to them or not.

That problem is not theoretical. A Master Builders Australia 2025 report noted that 35% of falls in Victoria and NSW involved subcontractors who lacked unified digital tracking for equipment. Further analysis from Curtin University shows that for certain residential projects, safety nets can be 2.5x more effective than guardrails, yet adoption lags at 18% according to the source provided in this brief at PMC.

That points to two practical issues. First, equipment records scattered across subcontractors create blind spots. Second, some sites keep defaulting to the control they know rather than the control that best suits the stage of work.

Tip: Do not accept "tag looks current" as proof of serviceability. On a busy site, visual reassurance and documented verification are not the same thing.

Build one register, not five versions of one

If multiple contractors are working at height, run a single site equipment register for anything used under your permit or within your controlled area. Ownership can stay with each contractor, but visibility cannot.

A workable register should capture:

• Item identity: Unique ID, type, brand, and serial or internal number
• Current status: In service, quarantined, under inspection, out of service
• Inspection record: Pre-use check outcome, formal inspection date, inspector name
• User link: Which contractor or crew is using it
• Location: Roof zone, scaffold bay, plant room, façade stage, or storage point

The point is not bureaucracy. The point is making sure a supervisor can answer, on the spot, whether the harness, reel, net, rail, or anchor system in front of them is approved for use.

What to inspect before every use

Pre-use checks should be brief, specific, and mandatory. Workers need to know what a failed item looks like.

A practical check includes:

1. Webbing and stitching Cuts, fraying, burns, pulled threads, contamination, UV damage.

2. Hardware Distortion, corrosion, sharp edges, gate function, locking action.

3. Shock packs and lanyards Signs of deployment, tearing, knotting, unauthorised modification.

4. Labels and identification Legible tags and traceable item ID.

5. Anchor and connection path Suitable point, no edge damage risk, no incompatible connectors, no trip hazard in the line.

If any one of those fails, remove it from service. Not after the shift. Immediately.

Equipment compatibility gets missed

A lot of incidents do not come from a single failed item. They come from systems that were never checked as a whole. Different karabiners, anchor slings, reels, beam clamps, static lines, and temporary anchors may all be individually serviceable but still unsuitable together.

This is especially important in specialist trades. For façade access and suspended work, it helps to review task-specific gear lists from adjacent industries. A resource like high-rise window cleaning equipment is useful because it shows how access, anchorage, descent control, and retrieval components need to work as a system, not as random kit pulled from a gang box.

Set a hard quarantine rule

Every site needs one fundamental rule. If an item is questionable, it is quarantined. No debate at the workface. No "just this once". No putting it back in the crate for someone else to discover later.

Use a simple process:

• Red tag it
• Remove it from the work area
• Record who found it and where
• Assign review to a competent person
• Close out the record before any return to service

That discipline matters more than fancy forms. Good equipment management is mostly about refusing to let doubtful gear drift around the site until somebody uses it.

Implementing Safe Work Procedures and Training

Good procedures are blunt instruments in the best sense. They leave little room for guesswork when the site gets noisy, rushed, or crowded.

The trouble is that many SWMS and permits still read like they were written to satisfy a file check. They use broad hazards, generic controls, and copied wording that could apply to any job in any suburb. That kind of document does not help a leading hand decide whether a task should start at 6:30 am on a wet roof with two other trades below.

What a good SWMS includes

A useful SWMS is tied to the site, stage, and sequence of work. It should tell the crew what the hazards are at this location, with this access method, using this equipment, around these other trades.

I want to see these features:

• Task sequence that matches the actual job Setup, access, work, repositioning, breaks, weather interruption, pack-down.

• Controls linked to each exposure point Not just "wear harness". Identify edge protection, exclusion zones, anchor method, rescue arrangements, and stop-work triggers.

• Interface controls How the task is separated from mobile plant, overhead lifting, deliveries, or workers below.

• Site-specific sign-off The person approving the SWMS has seen the work area.

A poor SWMS usually has one tell. If the work area changes and the document still reads fine, it is too vague.

Permit to work for the jobs that can turn quickly

Not every task at height needs a complicated permit process. But non-routine, unusual, or compounded risks do.

Typical examples include maintenance over running plant, access near fragile surfaces, edge work during shutdowns, façade work above public interface zones, or any job where rescue would be difficult.

A permit should force the supervisor and permit issuer to stop and confirm:

Permits work when they are short, clear, and checked in the field. They fail when they become office paperwork detached from the job.

Key takeaway: If the permit issuer cannot stand in the work area and verify the controls, the permit is only half done.

Rescue planning is part of the job, not an appendix

A lot of sites talk about rescue only after they have chosen harness-based controls. That is backwards.

If a worker can be suspended after a fall, the rescue method has to be practical for that exact location. Saying "call emergency services" is not a rescue plan for most construction or industrial tasks. It is escalation after your own system has failed to retrieve the worker promptly.

A basic rescue plan should identify:

• Likely fall path and final position
• Who leads the rescue
• Who contacts emergency services
• What equipment is needed
• How the casualty will be reached
• How the area below is secured
• Where first aid support will be staged

Then test it. Walk the route. Unpack the kit. Check whether the ladder, EWP, rope kit, or retrieval device can reach the suspended worker in the space available.

Train for competence, not certificates alone

A card in a wallet proves attendance. It does not prove the worker can set up a system properly on your site.

For working on heights safety, supervisors need evidence of competence in context. That means observing people doing the job, not just reading their ticket.

Useful checks include:

• Practical demonstration: Fit harness, inspect lanyard, select anchor, maintain tie-off, explain exclusion zone.
• Site induction verification: Confirm they understand your access rules, permit process, and emergency arrangements.
• Trade-specific review: Roofers, riggers, maintenance fitters, cladders, and cleaning contractors often face different exposures even when all are "working at heights".
• Refresher triggers: New equipment, changed work method, incident, near miss, or long gap since last task.

For broad site capability building, structured onsite training and assessment is useful because it checks what workers can do in the environment where they will be exposed.

Do not ignore ladders just because bigger risks exist

Ladders are involved in plenty of avoidable incidents because crews treat them as low-risk by default. They are often used for short tasks that then run longer, require overreaching, or involve tools and materials that should have triggered another access method.

When ladders are permitted, the details matter. One simple example is setup angle. A practical guide on finding the correct angle for a ladder is useful for training because it translates a common rule into something crews can check on site.

The same principle applies across all access methods. Workers need cues they can use at the workface, not just generic instructions from a slide deck.

Monitoring Your Program and Driving Improvement

If your review process only starts after an incident, you are using lagging information too late.

The best monitoring on a high-risk site is simple and repetitive. Supervisors inspect work areas. Leading hands report what changed. Permits are checked against what is happening. Equipment issues are closed out, not left hanging in a register nobody reads.

Watch for drift, not just breaches

Most failures do not begin with a dramatic non-compliance. They start with drift.

A crew removes a handrail to land materials and leaves it off for the next task. A permit is issued for one access method, but the team swaps to another after lunch. A subcontractor brings replacement harnesses that were never added to the site register. None of that looks major in isolation. Stacked together, it creates the conditions for a serious event.

That is why inspections should focus on live controls, not paperwork alone. Ask:

• Is the selected control still the right one for the task as it exists now
• Has anything changed in access, edges, weather, sequencing, or overlapping work
• Can the crew explain the rescue method without guessing
• Are fatigue and time pressure starting to affect decisions

Use incidents and near misses properly

A near miss is useful only if it changes the system. If the same themes keep appearing, you are looking at a program weakness, not a worker issue.

This matters in Western Australia in particular. A 15% rise in incidents has been linked to human factors like fatigue, and a 2024 University of WA study found 62% of roofers reported fatigue-induced errors while only 12% of H&S programs include modules on cognitive factors, according to the source provided in this brief at Safety by Design.

That is a useful reminder that monitoring should include more than PPE checks. Review shift length, sequencing pressure, heat, rework, and how often crews are being asked to make judgement calls late in the day.

Tip: If one crew's paperwork is perfect but they constantly need last-minute clarifications, the issue may be fatigue, rushed planning, or poor handover rather than form quality.

Digital oversight helps when it closes gaps

The practical value of digital tools is visibility. One dashboard for permits, inspections, subcontractor documents, corrective actions, and equipment status is far better than a folder, a spreadsheet, and six text messages.

What matters is not the software itself. It is whether the site team can see weak points early enough to intervene. If the pattern shows repeated issues in one work area, one contractor package, or one stage of the build, you can act before the next near miss becomes an injury.

Safety Space gives construction, manufacturing, and industrial teams one place to manage permits, inspections, subcontractor records, training, and corrective actions without chasing paper across multiple sites. If you need tighter control over working on heights safety and clearer visibility of what is happening at the workface, take a look at Safety Space.