Microbiological safety is where a beautifully formulated cosmetic can quietly fail. A product can meet every claim on the label and still be unsafe if its bioburden is uncontrolled — and for buyers of natural ingredients, the raw materials are often the largest single source of that risk. ISO 17516 is the standard that puts numbers to the problem, and understanding it changes how you read a certificate of analysis.
What ISO 17516 actually specifies
ISO 17516 sets the microbiological limits for cosmetic products. It does two things: it caps the total aerobic mesophilic count — the general population of bacteria, yeast and mould a sample can carry — and it requires that certain named organisms be absent altogether. Crucially, these are finished-product criteria. The standard tells you whether a batch of cosmetic is fit to release; it does not, by itself, set a limit for the plant extract or hydrosol that went into it.
The two product categories
The standard splits products into two categories according to how and where they are used.
| Category | Product type | Total aerobic mesophilic limit |
|---|---|---|
| Category 1 | Children under 3, eye-area, mucous membranes | ≤ 10² CFU/g or CFU/mL |
| Category 2 | All other products | ≤ 10³ CFU/g or CFU/mL |
The logic is exposure. A product used around the eyes or on a young child's skin, or one applied to mucous membranes, has a lower tolerance for contamination, so its ceiling is ten times stricter than the general limit. Knowing which category a formula falls into is the first step in setting an internal ingredient specification, because a Category 1 product leaves far less headroom for incoming bioburden. If a botanical extract arrives close to the Category 2 limit on its own, it may simply be incompatible with a Category 1 formula unless its bioburden is reduced first.
The four organisms that must be absent
Alongside the count limits, ISO 17516 requires the absence of four specified microorganisms in 0.1 g or 0.1 mL of product: Pseudomonas aeruginosa, Staphylococcus aureus, Candida albicans and Escherichia coli. These are treated as pass/fail — there is no acceptable low count, only presence or absence in the test portion. Each has a history in cosmetic contamination incidents, from Pseudomonas thriving in water-rich systems to E. coli signalling faecal contamination in the supply chain.
The methods behind the numbers
The limits mean nothing without agreed test methods, and ISO 17516 sits on top of a family of them. Enumeration of aerobic mesophilic bacteria follows ISO 21149; yeast and mould counts follow ISO 16212. Detection of the specified organisms uses dedicated methods — ISO 22717 for P. aeruginosa, ISO 22718 for S. aureus, ISO 18416 for C. albicans and ISO 21150 for E. coli. Separately, the preservation efficacy or challenge test, ISO 11930, evaluates whether the finished formula can resist a deliberate microbial challenge over time. Release testing and challenge testing are complementary: one proves the batch is clean today, the other proves the formula stays clean in use.
Where natural ingredients raise the risk
Natural and botanical materials are exactly the ones that stress a microbiological programme. Plant waters, hydrosols, aqueous extracts and low-preservative naturals offer both nutrients and moisture, and they may arrive carrying an agricultural bioburden from the plant material and its processing. The governing variable is water activity (a_w) — the free water available to microorganisms. Anhydrous oils and high-solids extracts with low a_w are largely self-protecting, whereas a botanical water with a_w near 1.0 is an open invitation to growth unless it is preserved, filtered or otherwise stabilised. A natural claim that limits preservative options makes this harder still, not easier.
What to require on a CoA
Distinguish the two levels clearly. For an ingredient, ask for the total aerobic mesophilic count, the yeast and mould count, confirmation that the four specified organisms are absent, and — for any water-containing grade — the water activity, each with the method cited. Set acceptance criteria tighter than the finished-product limits, because ingredient bioburden accumulates into the blend. For a finished product, the same counts are judged against the correct ISO 17516 category limit, with the challenge-test result on file. Read this way, a CoA stops being a formality and becomes the evidence that a natural ingredient can be formulated safely and defended under audit.