Short answer: A challenge test inoculates a finished formula with five organisms and measures how fast the preservative system kills them. Formulas usually fail for physical reasons — pH, partitioning, packaging — not because too little preservative was added.
A preservative system is the only part of a formula whose job is to do nothing visible. It does not thicken, scent or hydrate; it holds the product in a stable microbiological state for two or three years, in a bathroom, with fingers going in and out. Because it has no sensory signature, it is also the part most often designed by spreadsheet — a percentage copied from a similar product — and then discovered to be wrong at the point where a challenge test comes back with a bacterial count that will not fall.
What the test actually measures
Preservative efficacy testing — the challenge test — asks a single question in a structured way: if this product is contaminated, does the preservative system reduce the contamination and keep it reduced?
The reference method is ISO 11930, the standard for evaluating the antimicrobial protection of a cosmetic product. It is not a pass-or-fail certificate for the formula alone; it is a procedure for interpreting data from a preservative efficacy test, or from a microbiological risk assessment, or both. That wording matters, because it places the test inside a wider argument about risk rather than at the end of a checklist.
The method works on the finished product, in its final packaging where the protocol requires it, not on a bench sample of the preservative blend. That is the first place expectations diverge: a supplier's data on a preservative concentrate tells you about the concentrate, not about your emulsion with 18% oil phase, a xanthan gum network and a pH of 5.8.
The method in plain terms
The sequence is straightforward, and the discipline sits in the details.
Organisms. Five are used as standard: Staphylococcus aureus, Escherichia coli and Pseudomonas aeruginosa for the Gram-positive and Gram-negative bacterial challenge, Candida albicans for yeast, and Aspergillus brasiliensis for mould. The mould is the one that ends careers, because fungal reduction is slower and more formula-dependent than bacterial reduction.
Inoculum. Each organism is introduced at a defined load, typically in the 10⁵ to 10⁶ colony-forming units per gram range, into separate aliquots of product. Separate, not pooled — pooling hides which organism the system cannot handle.
Sampling. Counts are taken at day 0 and then at intervals through day 28, commonly day 7, 14 and 28. The shape of the curve matters as much as the endpoint: a system that reduces bacteria by four logs at day 7 and then lets the count climb back by day 28 has failed, even though the day-7 number looks excellent.
Criteria. ISO 11930 sets two acceptance profiles, A and B. Criterion A is the stricter of the two and is the one a formulator should aim at; criterion B allows more time to reach the same reduction and is often paired with a packaging or use-pattern argument. In broad terms both look for at least a three-log reduction of bacteria and at least a one-log reduction of yeast and mould within the defined windows, with no increase thereafter. Read the current standard text before writing a specification — the numbers belong to the standard, not to a blog table.
Exemptions. ISO 29621 describes when a product can be treated as low microbiological risk — anhydrous systems, extreme pH, very low water activity — and therefore assessed by risk assessment rather than by challenge test. The exemption is an argument, not a shortcut: it has to be documented product by product.
Why a preserved formula still fails
Failures cluster into a handful of physical mechanisms. None of them is solved by adding another 0.1% of the same preservative, which is why the first response to a failed test is usually the wrong one.
| Failure mode | What is happening | What to change |
|---|---|---|
| pH outside the active window | Organic acids are only antimicrobial in their undissociated form; benzoic and sorbic acid lose most of their effect above roughly pH 5 | Lower pH, or switch to a system whose activity does not depend on pH |
| Partitioning into the oil phase | Lipophilic preservatives migrate into the emulsion droplets and leave the water phase, where microbes live | Rebalance the system toward the aqueous phase, or increase the effective aqueous concentration |
| Binding to surfactants and polymers | Micelles, PEG-based emulsifiers and some gums associate with preservative molecules and reduce free concentration | Reduce the binding load, or use a system less prone to association |
| Chelator depletion | EDTA improves Gram-negative kill, but binds metals in the formula and can be saturated | Adjust chelator level to the formula's metal load |
| Raw-material bioburden | Botanical extracts, gums and proteins arrive carrying organisms and a nutrient load | Set incoming specs, consider filtered or irradiated grades |
| Packaging and use pattern | An open jar is inoculated every time a finger enters; an airless pump is not | Match the system to the pack, or change the pack |
| Water activity creep | Humectants and glycols raise osmotic stress slowly as water leaves the product | Verify at end of shelf life, not only at release |
The table is also a diagnostic order. Packaging and use pattern are cheapest to change at concept stage and most expensive after tooling; pH is free to change and often overlooked because the formula "feels" fine.
The preservative-free claim and where it breaks
Anhydrous balms, oil serums and powder formats can legitimately sit outside the challenge-test regime, because without water there is no microbial habitat. That is a real formulation achievement, not a marketing trick, and ISO 29621 gives the framework for arguing it.
The claim breaks when the product's water activity drifts, when the consumer introduces water — wet fingers, a damp bathroom, a reusable applicator — or when the "preservative-free" language is applied to a product that contains a multifunctional ingredient doing preservative work. A glycol at 3% in a water-containing formula is a preservative by function, whatever the INCI list implies.
Two habits keep this honest. First, write the claim to the product's real life, including in-use conditions. Second, test the product at the end of its shelf life as well as at release, because a system that passes at day 0 and fails at month 24 on an accelerated sample is a recall waiting for a date.
What the test does not tell you
A passing report is narrower evidence than most brand teams assume. It does not prove the product stays preserved for three years — that is a stability programme question, and the two are usually run as separate studies with separate endpoints. It does not prove the manufacturing process is hygienic; a clean formula made in a dirty batch still ships contaminated. It does not cover every organism, only the five in the method plus whatever the lab adds for cause. And it tests one batch, at one point in its life, unless the protocol says otherwise.
What it does give you is a defensible answer to the question a retailer, a notified body or an importer will eventually ask: on what basis do you say this product is preserved? For a product in a emulsion system, that answer also needs the process record — homogenisation temperature, cooling rate, and the point at which the preservative was added. A preservative added at 65 °C into an emulsion has been partly spent before the batch is cool.
What to ask for in the document pack
Ask for the report, not the summary line. The report should name the method and its edition, the organisms, the inoculum level, the sampling days, the criterion applied, the batch number and the packaging format tested. If the summary says "passed" without naming the criterion, you cannot tell whether it passed A or scraped through B.
Ask also whether the test was run on the oldest batch available or on a fresh pilot batch, and whether the lab tested the product in its commercial pack. Those two questions separate a document produced for a filing from a document produced for a decision.
FAQ
Q: What is a preservative efficacy test in cosmetics? A: It is a laboratory study that inoculates a finished cosmetic product with five standard organisms — three bacteria, one yeast and one mould — and measures how quickly the preservative system reduces them over 28 days. ISO 11930 is the reference method for interpreting the results.
Q: What is the difference between ISO 11930 criteria A and B? A: Both require a defined log reduction of bacteria, yeast and mould with no regrowth, but criterion A reaches the targets sooner and is the stricter profile. Criterion B allows more time and is often supported by a packaging or use-pattern argument. Aim for A unless there is a documented reason not to.
Q: Why did my formula fail the challenge test when it has a preservative? A: The usual causes are physical rather than dose-related: pH outside the preservative's active window, preservative partitioning into the oil phase, binding to surfactants and gums, high raw-material bioburden, or packaging that re-inoculates the product at every use.
Q: Can a product be exempt from challenge testing? A: Yes, when it qualifies as low microbiological risk under ISO 29621 — for example an anhydrous balm or a product at extreme pH. The exemption has to be justified by a documented risk assessment for that specific product, and it does not survive the introduction of water at the point of use.