Hazard analysis is where HACCP either stands or falls. I've reviewed hundreds of HACCP studies over the years — in my own sites, for clients, and as part of BRC audits — and the single most common weakness I see isn't at the CCP stage or in the monitoring records. It's here, at Step 6, where the foundational thinking happens. Thin hazard analysis produces thin HACCP systems, and auditors know it the moment they open the folder.
This guide covers Codex Principle 1 and Step 6 of CXC 1-1969 as interpreted through BRC Issue 9 clause 2.7. It's written for people who actually have to do this work, not for people writing management presentations about it.
What Hazard Analysis Actually Is
The formal definition from Codex CXC 1-1969 is the collection and evaluation of information on hazards and conditions leading to their presence to decide which are significant for food safety and therefore need to be addressed in the HACCP plan. That last phrase is critical — not every hazard identified needs a CCP. The purpose of hazard analysis is to determine which hazards are significant enough to require specific control measures, and to specify what those control measures are.
BRC Issue 9 clause 2.7 requires that the hazard analysis covers all biological, chemical, physical, and radiological hazards, including allergens as chemical hazards. In SafetyCore, allergens are assessed within the chemical hazard category — which is technically correct and consistent with how most competent HACCP teams approach it. Allergen cross-contact is a chemical hazard; it's a protein transfer, not a biological contamination event.
The Scope of the Analysis
You conduct hazard analysis at every step of your process flow diagram — from raw material receipt through to despatch. Every step. I've seen systems where hazard analysis stops at primary processing and doesn't consider packaging, storage, loading-out. That's incomplete and an auditor will identify it immediately.
For each step, you're asking: what could go wrong here from a food safety perspective? What hazards could be present, introduced, allowed to grow, or survive at this point? The answers need to be grounded in evidence — published scientific data, industry codes of practice, your own historical data, regulatory guidance, customer specifications. Gut feeling isn't hazard analysis.
The Five Hazard Categories
Codex CXC 1-1969 (2020 revision) requires consideration of five hazard categories:
Biological hazards are the most extensively documented. Pathogenic bacteria (Salmonella, Listeria monocytogenes, E. coli O157:H7, Campylobacter), viruses (norovirus, hepatitis A), parasites (Cryptosporidium, Toxoplasma), and mycotoxins from mould growth. The appropriate biological hazards depend entirely on your product, ingredients, and process. A chilled cooked meat product has a completely different biological hazard profile from a shelf-stable dry bakery product.
Chemical hazards include naturally occurring chemicals (mycotoxins, marine biotoxins, alkaloids), processing contaminants (acrylamide, benzene, polycyclic aromatic hydrocarbons), agricultural residues (pesticides, veterinary drug residues), environmental contaminants (heavy metals, PCBs, dioxins), cleaning chemical residues, and allergens. The allergen dimension is significant — deliberate allergen-containing ingredients are managed through product design, but cross-contact allergens require specific control and must appear in your hazard analysis for relevant process steps.
Physical hazards are foreign bodies with the potential to cause injury — metal fragments, glass, hard plastic, bone, stones, wood splinters. The key word is injury. Not every foreign material is a physical hazard; you need to assess size, hardness, and whether it could realistically injure the intended consumer. BRC Issue 9 and most good HACCP guidance now asks you to consider the intended consumer group here — a product aimed at infants or the elderly has a lower threshold for what constitutes a physical hazard.
Radiological hazards were formally added to the Codex list in the 2020 revision of CXC 1-1969. For most UK food manufacturers this is not a significant hazard in practice, but it must be considered and documented. Your hazard analysis should include a brief statement on why radiological contamination is or isn't a concern for your specific product and supply chain.
Allergens (as chemical hazards) deserve specific treatment because the regulatory and commercial consequences of allergen failure are severe. The 14 major allergens listed in UK food law (cereals containing gluten, crustaceans, eggs, fish, peanuts, soybeans, milk, nuts, celery, mustard, sesame, sulphur dioxide and sulphites, lupin, molluscs) must be assessed at every step where they could be introduced as cross-contact hazards, not just at formulation.
The PIGS Framework
PIGS is a structured way of thinking about how hazards behave at each process step. At every step in your flow diagram, for each hazard, you classify the mechanism as one or more of:
Presence — the hazard is already in the raw material or ingredient arriving at this step. Salmonella in raw poultry. Aflatoxin in groundnuts. Allergen protein in milk powder. The hazard exists regardless of what you do at this step — your job is to control it, not prevent it getting there because it's already there.
Introduction — the hazard gets into the product at this step due to something that happens in your process. Metal contamination from worn equipment. Allergen cross-contact from a shared conveyor. Listeria from a food contact surface that wasn't cleaned effectively. Introduction hazards are controlled by preventing them getting in — design, maintenance, cleaning, hygiene controls.
Growth — the conditions at this step allow a biological hazard to multiply. Bacterial growth at an inadequate chilling step. Mould growth in a product held too long between processing stages. Growth hazards require you to control the conditions that allow multiplication — temperature, time, water activity, pH.
Survival — a kill step exists, but the hazard survives it if the step isn't adequately controlled. Pathogen survival through undercooking. Spore survival through a heat treatment that targets vegetative cells but not spores. Survival hazards are why CCPs exist at kill steps — the CCP ensures the kill step delivers what it needs to deliver.
The reason PIGS matters is that the appropriate control measure depends entirely on the mechanism. You control Introduction differently from Growth, and you control Survival differently from Presence. A HACCP team that doesn't think in these terms tends to apply generic controls rather than targeted ones, and generic controls produce generic results.
Significance Assessment
Once you've identified potential hazards at each step, you assess significance. A significant hazard, as defined by the 2020 Codex revision, is a hazard identified through hazard analysis that needs to be controlled by a CCP or OPRP — not just a general hygiene control. Significance is the output of your risk assessment; it's not a pre-existing category.
The standard methodology is severity × likelihood. Both factors need to be assessed based on evidence, not assumption. Severity is the consequence to the consumer if the hazard is present at an unacceptable level — taking into account the nature of the hazard, the dose-response relationship, and the intended consumer group. Likelihood is the probability of the hazard occurring in the absence of the control measure being assessed — considering your raw material inputs, historical data for your product type, published outbreak data, and conditions in your specific process.
Most teams use a 3×3 or 5×5 risk matrix. The specific matrix you use matters less than applying it consistently and documenting the reasoning. Auditors want to see that you thought about it, not just that you scored it. A 1×1 = 1 score for a hazard with no explanation is not convincing. A 1×1 score with a note saying "Salmonella in pasteurised milk powder — risk of presence in ingredients controlled by supplier approval and CoA, heat treatment at step 8 would eliminate if present, likelihood assessed as low based on pasteurisation status of ingredient""is auditable evidence.
The Difference Between Identifying a Hazard and Determining Significance
This is where many HACCP systems go wrong. Every pathogen, every potential contaminant, every allergen gets listed — and then everything gets scored as significant, producing a system with CCPs at every step. Or alternatively, the list is thin and many hazards never appear at all.
The discipline is in the middle. You identify all potential hazards based on what could realistically occur. You then assess significance honestly based on severity and likelihood. The conclusion "this hazard is not significant at this step because...""is a valid and important output of hazard analysis. It's not corner-cutting; it's rigour. A hazard controlled by an effective PRP doesn't need to be elevated to a CCP — doing so creates work, complexity, and the risk of CCP overload making the system unmanageable.
Equally, a hazard that is genuinely significant — where the severity is high and the likelihood without specific control is realistic — must be designated as significant and assigned a control measure at the appropriate level. The most common hazard analysis failure I see in audits isn't over-classification; it's under-classification — hazards present in real products on real production lines that have been assessed as insignificant because the hazard analysis was done at a desk by someone who hadn't walked the line.
Documenting the Hazard Analysis
BRC Issue 9 requires documented evidence of the hazard analysis. This means a record for each process step showing: the hazards considered, the severity and likelihood scores with rationale, the significance conclusion, and the control measures assigned to significant hazards. A table format works well — process step down the left, hazard categories across the top, with scores and control measure assignment for each cell where a hazard is identified.
The control measure column is the bridge between hazard analysis and the rest of your HACCP system. Significant hazards controlled by CCPs flow into your CCP determination. Those controlled by enhanced PRPs or OPRPs flow into your PRP monitoring system. Those controlled by standard GHPs sit in your prerequisite programme documentation. Every significant hazard must have a home; no hazard should be assessed as significant and then left without an assigned control.
Common Audit Failures in Hazard Analysis
The failures I see repeatedly are: hazard analysis that lists hazards but doesn't assess significance (just a list of things that could go wrong); significance scores applied without rationale (a 3×3 table full of numbers with no explanation of why); hazards missing for certain process steps (packaging step with no physical hazard assessment, despatch with no temperature abuse assessment); allergen cross-contact hazards missing from steps where shared equipment is used; no consideration of radiological hazards (even if the conclusion is "not applicable, rationale: X"); and control measures for significant hazards that are either missing or vague ("controlled by good manufacturing practice""is not a control measure).
BRC Issue 9 clause 2.7 is a fundamental requirement. A non-conformance here is a major non-conformance at minimum. I've seen sites fail certification audits because their hazard analysis was a table of hazards with no significance assessment. Don't let that be you.
A hazard analysis is only as good as the knowledge and honesty that went into it. The question isn't "what do we need to write to satisfy the auditor?""— it's "what are the realistic ways this product could harm someone, and what are we doing about each of them?"
SafetyCore's hazard analysis module walks your team through every process step, capturing PIGS classification, severity and likelihood scores with rationale, and automatically routing significant hazards to your CCP determination or PRP register — so nothing falls through the cracks.
Start Free Trial