Undeclared allergens are consistently one of the leading causes of food recalls in North America, and most of those aren't label mistakes — they're cross-contact from a line that wasn't fully cleaned between an allergen-containing product and the next run. Changeover cleaning is where allergen control is won or lost. And the single most common gap we see is a team that confuses validating the procedure with verifying a specific changeover. They're not the same, and an auditor will ask you to prove both.
Validation vs verification — the distinction that trips people
Validation answers the question: "Does this cleaning procedure, performed correctly, actually remove the allergen to a safe level?" It's a one-time, science-based study you do when you design or change the procedure — using allergen-specific testing to prove the method works before you rely on it.
Verification answers a different question: "Did this specific changeover, on this line, right now, actually get done and get done right?" It's the routine check you perform at every changeover — visual inspection plus a rapid test — to confirm the validated procedure was executed effectively this time.
Put bluntly: you validate the recipe once; you verify the meal every time you cook it. Plants that only do one of the two have a hole. A validated procedure with no per-changeover verification means you're trusting that it was followed. Per-changeover swabbing with no validation means you're testing against a method nobody proved works.
What a changeover cleaning procedure must cover
Before you can validate anything, the procedure itself has to be complete and specific to the equipment. A defensible allergen changeover SSOP names the "from" and "to" products, calls out every product-contact surface including the hard-to-reach ones (dead legs, augers, seals, overhead structures that shed onto the line), and specifies disassembly, the cleaning chemistry and concentration, water temperature, contact time, rinse, and the acceptance criteria. Dry-cleaning environments — bakery, snack, powder — need their own approach, since water isn't an option and controlled removal plus segregation does the work.
Validation — proving the procedure works
Validation is a designed study, not a one-off swab. Run the full cleaning procedure after a worst-case allergen-containing product, then test the cleaned surfaces for the specific allergen protein using an allergen-specific method — commonly a lateral-flow device or ELISA calibrated to the target allergen. Sample the hardest-to-clean locations, not the easy flat surfaces, and repeat across enough changeovers to show the result is consistent, not lucky. Document the target allergen, the test method and its limit of detection, the sample sites, the results, and the conclusion that the procedure reliably removes the allergen. Re-validate whenever the product, equipment, chemistry, or procedure changes.
One important nuance: general cleanliness tests like ATP measure organic residue, not a specific allergen. ATP is excellent for verifying overall cleaning effectiveness and is fast and cheap — but it does not confirm the absence of a particular allergen protein. Validation of allergen removal needs an allergen-specific test.
Verification — proving it worked this time
At every changeover, verification is a two-step gate. First, a documented visual inspection of the disassembled, cleaned equipment by a trained person — no visible residue, product, or film on any product-contact surface. Second, a rapid test appropriate to your validated program: an allergen-specific lateral-flow swab on the highest-risk sites, and/or a protein-residue or ATP swab as your program defines. The rule that makes it defensible: production does not start until verification passes. A fail doesn't get a note and a shrug — it sends the line back for re-cleaning and re-testing, and it's recorded.
A changeover verification checklist
The steps below are the ones we see hold up under audit. Adapt the specifics to your validated procedure.
1. Confirm the changeover need. Record the "from" product and its allergens and the "to" product. Confirm whether a full allergen clean is required (it is any time the incoming product doesn't declare the outgoing allergen).
2. Perform the validated cleaning procedure. Follow the current SSOP for that equipment. Disassemble to the level the validation required. Don't shortcut the hard-to-reach sites — that's where residue hides.
3. Visual inspection, documented. A trained inspector checks every product-contact surface. Any visible residue means re-clean before going further. Record who inspected and the result.
4. Rapid verification test. Swab the pre-defined highest-risk sites using your validated method. Record the site, the test, the result, and the time.
5. Pass/fail decision — before startup. Pass: release the line, with the record attached to the run. Fail: re-clean, re-inspect, re-test. No product runs on a fail.
6. Close the loop. Every fail is a corrective action — root cause identified, procedure or training adjusted, and a check that the fix held. Recurring fails on the same asset mean the validated procedure or the equipment design needs another look.
Where changeovers go wrong
The failures are consistent. Verification results recorded after the line already started running — which defeats the entire point. Swabbing only the easy flat surfaces while the seals and dead legs go untouched. Using ATP as if it confirmed allergen removal. A "complete" changeover with no record of the verification swab result. And the quiet one: a validated procedure that drifted — the chemistry, concentration, or disassembly on the floor no longer matches the method that was actually validated.
Making it defensible
The goal is a changeover that can't be marked complete without the verification result attached, and a line that can't start on a fail. On paper that discipline is hard to enforce — the record lives on a clipboard that reaches QA hours later, by which point the run is underway. When the changeover procedure, the verification result, and the release decision live in one system, the sequence is enforced by design: the task surfaces the validated procedure, the verification result is captured at the point of work, a fail opens a corrective action automatically, and the timestamped, signed record is exactly what an auditor asks to see. That's the difference between hoping the changeover was clean and being able to prove it.
Make every changeover provable
MySitesSupervisor runs allergen changeovers as guided verification programs — the line can't be released until the visual check and verification result are captured, a fail rolls straight into a CAPA, and every record is timestamped and audit-ready for SQF, BRC, and FSMA.
Schedule a 30-min demo