Bakery and confectionery testing

The headline contaminant risk in this industry is not made in your plant. Aflatoxins come in with nuts, cereal flours, cocoa and spices, and they are heat-stable — baking does not destroy them, and blending a contaminated lot into a larger batch dilutes the number without removing the hazard. Once a hot lot is in the mixer, every downstream option is bad.

That is why the test belongs at goods-in, against the limits in SFDA.FD CAC 193, before release to production. A certificate of analysis from the supplier is a claim; a result on the delivered lot is verification. Where a limit depends on the specific commodity and its end use, the laboratory confirms the applicable figure in writing before testing begins.

Nuts and nut pastes

Aflatoxin verification on delivered lots, not supplier paperwork alone

Cereals and flours

Aflatoxins plus moisture at intake — wet flour is a mould inoculum

Spices and cocoa

Small inclusion rates, disproportionate aflatoxin risk

An aflatoxin failure in the finished batch is a disposal decision. The same failure at goods-in is a rejected pallet.

Mould is the failure mode that ends bakery shelf life, and water activity — not moisture content alone — is what sets the clock. Two cakes at the same moisture can have very different water activities depending on sugars and humectants, and only one of them will hold its declared life. If you formulate against moisture and never measure water activity, you are estimating the one number that decides the outcome.

The expiry date itself is a regulated claim under SFDA.FD 150-1 and SFDA.FD/GSO 150-2, and microbiological criteria sit in GSO 1016:2015. A shelf-life study replaces ‘we have always given it six months’ with data: product held under defined conditions, tested at planned points, mould and water activity tracked until the evidence supports the date on the pack.

Unlike aflatoxin, acrylamide is a process contaminant — it forms during baking and frying, driven by recipe and by time-and-temperature. That makes it controllable, but also means every recipe or process change can move the number: a new flour supplier, a darker bake, a reformulated dough. Monitoring gives you a baseline per product, so a change shows up as a shift you can act on rather than a surprise in someone else’s audit. Benchmark and limit values vary by product category; the laboratory confirms what applies to your product in writing.

Multi-ingredient products are where calculated panels go wrong. A database calculation is only as stable as its inputs, and in a bakery the inputs move: a supplier reformulates a margarine, rework goes back into the mix at a variable rate, inclusion levels of nuts or chocolate vary piece to piece, and baking loss changes the concentration of everything that remains. The panel printed at launch quietly stops describing the product being made.

Analysed values anchor the declaration under GSO 2750:2024, sitting inside the general labelling requirements of GSO 9:2022. The practical rhythm is analysis at launch and after any recipe or supplier change — the two moments the calculation is most likely to have drifted.

Bakeries run nuts, sesame, milk, egg and gluten on shared lines, so allergen control is usually your highest-consequence risk. We will be straight with you: allergen quantification by ELISA is not in our current catalogue, and we will not sell you a test we do not run. What we can do is discuss where analytical verification genuinely fits in your allergen matrix and cleaning validation, and put in writing what we can and cannot support — the same discipline behind our accredited work. Our testing accreditation under ISO/IEC 17025:2017, certificate N-T-00607, covers food and agricultural microbiology; you should expect that level of candour about scope from any laboratory you use.