The Complete Guide to Loss Prevention in Manufacturing

This guide is general information for Canadian manufacturing and processing operations. It is not legal advice, and it is not a substitute for a site-specific hazard assessment. Requirements vary by province, by facility, and by process.

Most manufacturing plants already have a safety program. Far fewer have a loss prevention program — and the difference shows up in the numbers that reach the executive team.

A safety program asks whether anyone got hurt. A loss prevention program asks a broader question: what did this plant lose this year, and why. Lost production hours. A guard bypassed on a converting line. A dust collector that failed a housekeeping inspection three quarters running. A contractor crew that arrived without valid tickets and pushed a shutdown two days past window. An insurer that repriced coverage after a hot work incident.

Those are all losses. They have different owners, different budgets, and usually different reporting lines — which is precisely why they rarely get managed as one system.

This guide sets out how to build that system. It is written for the people who actually carry it: project managers running capital and maintenance work, and maintenance leaders who own the equipment, the contractors, and the outage schedule.

What loss prevention means in a manufacturing plant

Loss prevention is the practice of establishing policies, procedures, and business practices that reduce an organization's exposure to financial, production, inventory, and property loss — and the likelihood those losses occur.

The term causes confusion because it means something narrower in retail, where "loss prevention" is largely shrinkage and theft. In an industrial setting the definition is wider. A plant loses value through five channels, and a credible program addresses all five:

1. People. Injury, illness, and exposure. This is the channel with statutory force behind it, and the one that drives the others when it fails.

2. Production. Unplanned downtime, extended outages, missed turnaround windows, throughput lost to rework. The single largest dollar figure in most plants, and the one least often attributed to a safety root cause.

3. Property and equipment. Fire, explosion, mechanical damage, water damage, damage caused during maintenance activity itself.

4. Product and inventory. Contamination, spoilage, foreign material in the process stream, damaged stock, unaccounted materials movement.

5. Compliance and reputation. Orders, stop-work notices, administrative penalties, premium increases, failed customer audits, loss of a certification a customer contractually requires.

The reason to treat these together is that the same underlying conditions produce all five. A poorly executed lockout is a fatality risk, a downtime risk, an equipment damage risk, and a regulatory risk simultaneously. Managing it once, well, addresses the whole column.

Key distinction for plant leadership: Loss prevention is not a department. It is a management system that connects hazard data, maintenance planning, emergency capability, workforce competency, and audit evidence into one loop. Most plants already own every input. Few have connected them.

The regulatory floor in Canada

Manufacturing in Canada is regulated provincially in almost all cases. Federal jurisdiction under the Canada Labour Code Part II and the Canada Occupational Health and Safety Regulations applies to a narrow set of operations and rarely to conventional manufacturing.

Your program has to satisfy the province you operate in. If you run multiple plants across provinces, you are satisfying several at once, and they differ in structure more than in intent.

British Columbia — the Workers Compensation Act and the OHS Regulation (B.C. Reg. 296/97), administered by WorkSafeBC. Part 3 sets the requirement for a formal occupational health and safety program based on workforce size and risk classification. Part 5 covers chemical and biological agents. Part 8 covers personal protective equipment. Part 9 covers confined spaces. Part 10 covers de-energization and lockout. Part 12 covers tools, machinery, and equipment. Part 32 covers evacuation and rescue.

Alberta — the Occupational Health and Safety Act and OHS Code, administered by Alberta Occupational Health and Safety. Part 2 covers hazard assessment, elimination, and control. Part 5 covers confined spaces. Part 7 covers emergency preparedness and response. Part 15 covers managing the control of hazardous energy. Part 25 covers tools, equipment, and machinery. A written health and safety program is required at defined workforce thresholds under the Act.

Saskatchewan — The Saskatchewan Employment Act and The Occupational Health and Safety Regulations, 2020, administered by the Ministry of Labour Relations and Workplace Safety.

Ontario — the Occupational Health and Safety Act and O. Reg. 851, Industrial Establishments, administered by the Ministry of Labour, Immigration, Training and Skills Development. O. Reg. 851 is the core machine guarding, material handling, and lockout regulation for Ontario plants, and joint health and safety committee obligations attach at defined workforce sizes.

Quebec — the Loi sur la santé et la sécurité du travail and the Règlement sur la santé et la sécurité du travail, administered by the CNESST.

The consensus standards that fill the gaps

Regulations set the floor. They are usually performance-based, which means they tell you the outcome required without telling you the method. Auditors, insurers, and courts look to consensus standards to judge whether your method was reasonable.

The ones that matter most in a manufacturing loss prevention program:

On NFPA 660 specifically: if your plant generates any combustible dust — wood, grain, food, plastics, metals, pharmaceuticals, coatings — this is the document you now work from. <cite index="5-1">On December 6, 2024, NFPA 660, Standard for Combustible Dusts and Particulate Solids, replaced the six previously separate NFPA combustible dust standards: NFPA 61 for agriculture and food processing, 484 for metals, 652 for fundamentals, 654 for chemicals and other materials, 655 for sulfur, and 664 for woodworking and wood processing.</cite> <cite index="3-1">The first ten chapters are fundamentals that apply to all facilities — hazard identification, the dust hazard analysis, management systems, and hazard prevention and mitigation — with a new tenth chapter added for emergency planning and response, and commodity-specific chapters beginning at Chapter 21.</cite>

Two practical notes. First, NFPA 660 is a consensus standard, not law in itself — <cite index="10-1">it becomes enforceable when a fire code adopts it, when an insurer makes it a condition of coverage, or when it is used as evidence of a recognized hazard</cite>. Second, jurisdictions and insurers that adopted a specific legacy edition keep referencing it until they update their adoption, so confirm the applicable document with your authority having jurisdiction and your insurer rather than assuming the consolidation applies to you automatically.

Step 1 — Build the hazard assessment your program actually runs on

Everything downstream depends on this. A weak hazard assessment produces a program that looks complete on paper and fails at the first serious event.

Most plants have hazard assessments. Most of those assessments have the same three defects.

They are task-based only. A job hazard analysis for "changing the die" is necessary and insufficient. It will not surface the hazard that lives between tasks — the interaction of a maintenance crew, a partially energized line, a contractor working overhead, and a production supervisor who needs the line back.

They are stale. The assessment was written when the equipment was installed. Since then the plant has added a conveyor, changed a solvent, reconfigured a mezzanine, and started running a second shift. None of that is in the document.

They have no owner in operations. They sit with HSE. Maintenance planning does not consult them. Project management does not consult them. The assessment describes a plant that no longer exists and nobody notices, because nobody reads it.

What a program-grade assessment includes

Work through your facility in layers rather than by department:

Process and equipment hazards. Machine guarding and access points on every production line. Stored energy across all sources — electrical, pneumatic, hydraulic, thermal, chemical, gravitational, and residual. Pressure systems. Automated and robotic cells. Material handling and mobile equipment traffic paths.

Chemical and physical exposure. Solvents, process chemicals, welding fume, silica, wood or grain dust, noise, heat stress, vibration. This layer is where you determine whether you need quantitative exposure data. If you are estimating rather than sampling, you do not have a defensible position — industrial hygiene and exposure monitoring closes that gap.

Fire and explosion. Ignition sources against fuel sources, mapped. Hot work locations and permit control. Dust accumulation surfaces, including the ones above sightline — beams, cable trays, ductwork, light fixtures. Flammable storage. Suppression system condition and inspection currency. Fire inspection services give you a documented baseline here rather than a walkthrough impression.

Confined and restricted spaces. A full inventory, classified, with entry procedures and a rescue plan attached to each. Silos, hoppers, mixers, tanks, pits, ductwork, and process vessels. Note that many plants have spaces that meet the regulatory definition and have never been formally identified as such.

Work at height and access. Mezzanines, platforms, roof access, elevated maintenance points, fixed and portable ladders.

Contractor and interface hazards. What happens when an outside crew works alongside production. This is consistently under-assessed and consistently over-represented in serious incidents.

Business continuity exposure. Single points of failure. Long-lead-time equipment with no spare. Utilities. One process step that stops the whole plant.

Rank it so it can be actioned

Score every identified hazard on likelihood and consequence, and consequence should include production loss, not injury severity alone. That is the modification that makes a risk register useful to a maintenance leader instead of an HSE artifact. Apply the hierarchy of controls in order — elimination, substitution, engineering, administrative, PPE — and record which level of control you selected and why.

Assign every controlled hazard a named owner with a date. An unowned control is a finding waiting to be written.

Step 2 — Engineer out the losses that repeat

Once you have the register, attack the categories that produce the highest combined injury and downtime cost. In manufacturing, these five recur across nearly every facility.

Machine guarding and safeguarding

The most cited category in most provinces and the most common source of serious injury in manufacturing. Fixed guards get removed for maintenance and not replaced. Interlocks get defeated because a cycle time improves without them. Guarding gets designed for the original machine configuration and never revisited after a retrofit.

Build a guarding inventory tied to your equipment register, inspect it on a defined frequency rather than opportunistically, and treat a defeated interlock as a reportable event internally — not a workaround.

Control of hazardous energy

Lockout failures are the highest-consequence maintenance hazard in a plant. The common program defects are consistent: procedures written generically rather than equipment-specifically, no accounting for stored or residual energy, group lockout that breaks down at shift change, and contractors operating under their own procedure rather than yours.

CSA Z460:20 permits alternative control methods where full lockout prevents completion of a task that is integral to the production process — but only on the basis of a documented risk assessment. That is a narrow, evidenced exception, not a general permission. If your plant uses alternative methods, the risk assessment supporting each one needs to exist and be current.

Combustible dust

If the plant produces dust, the dust hazard analysis is the controlling document. It identifies where combustible dust exists, where it can accumulate, what could ignite it, and what controls apply — engineering controls including collection, venting, suppression, isolation, and bonding, plus housekeeping, maintenance, and training.

The failure mode is almost always housekeeping drift. The DHA is completed, controls are specified, and eighteen months later accumulation is back on the overhead surfaces because cleaning frequency was never written into a maintenance work order with an inspection behind it. [VERIFY WITH OPERATIONS — confirm Irwin's role in DHA support before publishing.]

Hot work and permits

Hot work causes disproportionate property loss. The controls are well known — permit, fire watch before, during, and after, area clearance, suppression availability, no-hot-work zones. The failure is administrative: permits issued by someone without authority, fire watch abandoned early, or maintenance work reclassified to avoid the permit process.

If your fire watch coverage collapses during outages because everyone is committed elsewhere, that is a staffing problem with a staffing solution, not a discipline problem.

Confined space entry

A silo, hopper, or process vessel entry in a manufacturing plant carries the same risk profile as one in a refinery, but plants often manage it with less infrastructure. The entry permit, atmospheric monitoring, attendant coverage, and a rescue plan that can actually be executed in the time available are all required — and the rescue plan is the element most often theoretical.

For plants running high-frequency entries during outages, digital confined space monitoring provides continuous gas detection, access control, camera visibility, and a documented entry record without relying entirely on manual attendant coverage.

Step 3 — Make the emergency plan executable

An emergency response plan that has never been tested is a document, not a capability. The test is simple: if the event happened on night shift, during a shutdown, with a reduced crew and contractors on site — could this plan be executed by the people actually present?

What the plan has to cover

Manufacturing sites need scenario-specific response, not a generic evacuation procedure:

• Fire and explosion, including dust deflagration if applicable
• Chemical release, spill, and uncontrolled reaction
• Medical emergency and trauma, including entrapment in machinery
• Confined space rescue, non-entry and entry
• Work-at-height rescue, including suspension trauma response
• Utility failure and loss of critical services
• Severe weather and external events
• Evacuation, muster, and accounting for all personnel — including contractors and visitors

Where plans break

Response time versus actual capability. Municipal fire response time to your plant is a known number. Get it. If a rescue has to begin before that number expires, you need on-site capability, not a phone call. Many industrial rescues require a start measured in minutes.

No trained on-site responders on every shift. The plan names a team. Half of them work days. Check the roster against the shift schedule.

Rescue planning treated as generic. A rescue plan for a specific silo, in a specific configuration, with specific access limitations, cannot be written in the abstract. It requires walking the space.

Contractors excluded from the plan. During a shutdown, contractor headcount can exceed plant headcount. If your muster and accounting process does not include them, it does not work when it matters.

No drills. A drill is not a fire alarm test. It is a scenario run against the actual plan with the actual people, with observers recording where it broke, followed by corrective action.

Where a plant cannot maintain trained response capability across all shifts — which is most plants, most of the time — on-site medical and first aid coverage and industrial firefighting services can be staffed to the periods of elevated risk rather than year-round. That is typically shutdowns, turnarounds, major maintenance, and capital project phases.

Step 4 — Train to competency, and prove it

Training is where loss prevention programs most often satisfy the letter of the requirement and miss the intent. A worker who has a valid ticket and cannot perform the task safely is a documented liability, not a controlled risk.

Build a training matrix, not a training list

Map every role in the plant against required competencies, and drive the requirements from your hazard assessment rather than from what has historically been booked. A useful matrix specifies, for each role:

• Regulatory training required by the province
• Site-specific and equipment-specific competency
• Refresher frequency and expiry
• Method of verification — knowledge test, practical demonstration, supervised field observation
• Who is qualified to sign off

Manufacturing-specific requirements to cover

WHMIS 2015. Lockout and hazardous energy control, including any alternative methods in use. Machine-specific operation and safeguarding. Confined space entry, attendant, and supervisor roles. Fall protection. Respiratory protection, including fit testing. Powered mobile equipment and forklift. Fire extinguisher and fire watch. First aid to the level your province requires for your workforce size and hazard rating. Emergency response roles. Incident investigation, for supervisors. Supervisor competency in due diligence obligations.

Respiratory protection deserves a specific flag. A respirator provides its rated protection only when it seals on the individual wearing it. If your plant has workers in respiratory protection without current fit test records, you have an exposure control gap and an audit finding. Respirator fit testing is a discrete, schedulable service — including mobile delivery to the plant, which avoids pulling crews off site.

Contractor competency is your exposure too

Where you are the owner, the constructor, or the prime contractor, contractor qualifications are part of your program. Verify tickets before mobilization, not on arrival. Deliver a site-specific orientation that covers your hazards, your permits, your muster points, and your reporting expectations. Track expiry centrally so a certification lapse is caught before it becomes a stop-work.

Manual tracking of this at scale does not hold. When a plant has several hundred workers and contractors with rolling certification expiries, a spreadsheet fails quietly — the Wombat safety software partnership exists to move certification tracking, digital forms, inspections, and corrective actions onto a system that flags expiries before they lapse.

Step 5 — Control site access, materials movement, and physical loss

This is the channel most often left out of a safety-led program, and it is where a genuine loss prevention program differs from an OHS program.

Access control and personnel accounting. Knowing who is on site, where, and under what authorization is simultaneously a safety control and a loss control. It underpins muster accounting, contractor management, and materials accountability.

Materials and inventory movement. Documented movement of high-value equipment, tooling, and consumables — particularly during outages, when the volume of people and material moving through the plant spikes and normal controls loosen.

Foreign material exclusion. In food, pharmaceutical, and precision manufacturing, FME is both a product quality control and an equipment protection control. Establishing and monitoring FME zones during maintenance activity prevents contamination events that can force a recall or destroy a production run.

Equipment condition and inspection currency. Fall protection, rescue equipment, gas detection, fire suppression, and lifting equipment all carry inspection requirements. Expired or uninspected equipment fails in two ways — it may not perform, and its condition is a finding regardless of whether it was used. Equipment inspections on a documented schedule remove both.

Irwin's provides trained personnel for site access monitoring, personnel recording, and associated emergency response and medical support on private industrial property. Irwin's is not a registered security company and does not provide security services on public property. [VERIFY WITH OPERATIONS AND LEGAL — wording of this paragraph.]

Step 6 — Measure, audit, and close the loop

A program without measurement degrades. The question is what you measure.

Stop leading with lagging indicators

Recordable injury frequency, lost-time claims, and incident counts tell you what already failed. They are required, they are useful for trending, and they are useless for prevention because they arrive too late to act on. A plant can report zero recordables in a quarter where every control it depends on quietly eroded.

Track leading indicators that predict loss

Audit at three levels

Internal self-assessment, quarterly, run by the plant. Cheap, frequent, and catches drift.

Internal audit against a standard, annually, run by someone independent of the area audited. Structured against CSA Z45001, your COR requirements, or your customer's audit protocol.

External audit, at the interval your certification or customer requires. This is the one with consequences attached.

The purpose of an audit is not the score. It is the corrective action register that comes out of it, with owners and dates, tracked to closure. A plant that audits well and closes findings slowly has an expensive documentation habit, not a program.

Investigate to cause, not to blame

Every significant incident and high-potential near miss gets investigated to root cause. CSA Z1005 sets out the methodology. The test of a good investigation is whether the corrective action changes a condition or a system — if the action is "retrain the worker" and nothing about the equipment, procedure, or supervision changed, the cause was not found.

Building the program: a practical sequence

For a plant starting from a partial program, this sequence front-loads the work that everything else depends on.

Months 1–2 — Baseline. Gap assessment against your provincial requirements and CSA Z45001. Inventory what already exists. Identify what is missing, what is out of date, and what exists but is not being used.

Months 2–4 — Hazard assessment. Complete the facility-wide assessment across all layers. Build the risk register. Rank with production consequence included.

Months 4–6 — Priority controls. Address the highest-ranked hazards. Machine guarding remediation, lockout procedure rewrite, dust hazard analysis, permit system repair.

Months 5–7 — Emergency planning. Write scenario-specific plans. Confirm on-site response capability against actual response times. Identify where external coverage is required.

Months 6–9 — Training. Build the matrix. Close the gaps. Establish the tracking mechanism.

Months 8–10 — Systems. Move documentation, inspections, corrective actions, and certification tracking onto a platform. Define the leading indicator set.

Months 10–12 — Verify. First internal audit. First full drill cycle. Corrective actions assigned and tracked. Management review.

Ongoing. Reassess after every process change, equipment change, incident, and at a defined periodic interval regardless of change.

Every date above needs a named owner, and the program needs a single accountable executive. Programs that are everyone's responsibility are nobody's.

Where an external partner makes sense

Not every element of this belongs in-house. Three tests usually settle it.

Is the requirement continuous or episodic? Fire watch, standby rescue, and additional safety coverage spike during shutdowns and capital work and drop to near zero between. Staffing permanently for a peak is expensive; scaling for it is not.

Does it require capability you cannot maintain? Quantitative exposure monitoring, fit testing, technical rescue, and industrial firefighting all require equipment, currency, and practice. A plant that needs these four times a year will not maintain them well.

Does independence matter? Gap assessments, internal audits, and incident investigations carry more weight — internally and externally — when the person conducting them does not report to the area being reviewed.

Irwin's supports Canadian industrial operations across loss prevention services for industrial worksites, occupational health and safety consulting, emergency response, safety training, occupational testing, equipment, and labour support. [VERIFY WITH OPERATIONS — confirm manufacturing sector coverage before publishing; the current service page lists power generation and distribution, mining and metals, forest products, oil and gas, and pipeline construction.]

Frequently asked questions

What is loss prevention in manufacturing? Loss prevention in manufacturing is the set of policies, procedures, and business practices that reduce a facility's exposure to financial, production, inventory, property, and compliance loss. It covers worker injury, unplanned downtime, equipment and property damage, product and inventory loss, and regulatory exposure as a single connected system rather than as separate departmental concerns.

How is loss prevention different from a health and safety program? A health and safety program focuses on preventing worker injury and illness and meeting regulatory obligations. A loss prevention program includes that and extends to production loss, equipment and property damage, inventory and product loss, and business continuity. The same hazard controls often serve both, but loss prevention measures success across a wider set of outcomes and typically reports to operations as well as HSE.

Is a written loss prevention program legally required in Canada? There is no standalone federal or provincial requirement titled "loss prevention program." However, most provinces require a written occupational health and safety program at defined workforce thresholds and risk classifications, along with hazard assessment, emergency response planning, worker training, and incident investigation. A loss prevention program is generally built on top of those statutory requirements. Confirm the specific obligations for your province and workforce size with the applicable regulator.

How often should a manufacturing hazard assessment be updated? Reassess whenever a process, substance, piece of equipment, or work method changes, after any significant incident or near miss, and at a defined periodic interval regardless of change. Many facilities use an annual review cycle as a baseline, with event-triggered reassessment in between. Check your provincial requirement, as some jurisdictions specify the trigger conditions.

What does NFPA 660 mean for my plant? If your operation generates combustible dust, NFPA 660 is the consolidated standard that replaced six previous NFPA combustible dust standards effective December 6, 2024. It requires a dust hazard analysis identifying where combustible dust exists and what controls apply. NFPA 660 is a consensus standard, so its legal force depends on adoption by your authority having jurisdiction, your insurer's requirements, or its use as evidence of a recognized hazard. Confirm which document applies to your facility with your AHJ and insurer.

Which leading indicators best predict loss in a plant? The most useful are corrective action closure rate against target dates, percentage of workforce holding current required training, hazard assessments current versus overdue, permit audit pass rate, preventive maintenance schedule adherence, and near-miss reporting volume. These move before an incident occurs, which is what makes them actionable.

Can a manufacturing plant outsource emergency response? Yes, and many do for specific capabilities or specific periods. On-site medical, fire watch, standby rescue, and industrial firefighting are commonly contracted for shutdowns, turnarounds, and capital projects when internal capability cannot be maintained across all shifts. Confirm that any external provider's personnel qualifications, response capability, and coverage hours match what your emergency plan assumes.

Start where the exposure is

If you are building a loss prevention program from a partial base, the sequence matters more than the speed. A current hazard assessment makes every subsequent decision defensible. Without it, you are allocating budget against assumptions.

Talk to Irwin's Safety about a loss prevention program assessment. We will review your current hazard assessment, emergency plan, training matrix, and audit position against your provincial requirements — and identify the gaps that carry the most exposure.

This guide is general information and does not constitute legal or regulatory advice. Requirements are jurisdiction-specific and depend on your facility, processes, and hazard assessment. Confirm current obligations with the regulator in your province and seek professional review for site-specific decisions.