Preservative Efficacy Testing for Cosmetics: What the Challenge Test Reveals Before Your Product Ships

A personal care product that looks completely stable on the shelf can harbor a Pseudomonas aeruginosa population after less than 28 days of normal consumer use — if the preservative system isn’t doing what you think it’s doing. That’s not a hypothetical. It’s precisely the scenario that preservative efficacy testing is designed to catch before your product ships.

And yet brands routinely launch without it. Sometimes because their contract manufacturer assured them the formula is “industry standard.” Sometimes because the test feels expensive relative to a small first run. Sometimes because no one asked. The result is products that pass visual inspection and basic stability panels but harbor a quiet preservation failure that only becomes visible when a consumer files a complaint — or worse, a dermatologist traces an infection back to your eye cream.

Here’s the regulatory reality: the FDA doesn’t require preservative efficacy testing for finished cosmetics the way it does for OTC drug products. There’s no mandatory pre-market approval for most cosmetics under 21 CFR. That gap means brands can legally ship a product that would fail a challenge test, right up until things go wrong at the consumer level. A cosmetic testing laboratory can close that gap before it becomes your problem.

What a Preservative Efficacy Test Actually Measures

The preservative efficacy test (PET) — also called a challenge test or antimicrobial effectiveness test — tells you how well your preservative system performs under real-world microbial stress. The procedure is deliberate: your finished formula is inoculated with specific microorganisms at defined concentrations, then monitored over time to see how quickly and completely your preservative system kills or inhibits them.

The two dominant frameworks for cosmetics in the US market are USP <51> Antimicrobial Effectiveness Testing and ISO 11930:2019 — Evaluation of the Antimicrobial Protection of a Cosmetic Product. ISO 11930 has become the more commonly requested standard for finished cosmetics, while USP <51> remains the go-to for cosmetic products carrying drug claims — SPF moisturizers, antiperspirant/deodorant hybrids, medicated lip treatments. Both methods share the same core logic: inoculate, incubate, count. But they differ in their acceptance criteria, which matters more than most formulators realize going in.

A full challenge test runs for 28 days. Microbial plate counts are pulled at four timepoints — Day 2, Day 7, Day 14, and Day 28 — and the trajectory of reduction across those timepoints determines a pass or fail. Any lab promising a faster result is either running predictive models (not a substitute for a real test) or skipping timepoints, which will get your COA rejected by most retail buyers and Amazon compliance reviewers.

The Five Test Organisms — And Why Each One Was Chosen

Standard cosmetic challenge testing uses five reference organisms, each selected because it represents a realistic contamination risk during manufacturing or consumer use:

Staphylococcus aureus (ATCC 6538) — A Gram-positive bacterium found on human skin and mucosal surfaces. The most clinically relevant contaminant in personal care products that contact broken or compromised skin.

Pseudomonas aeruginosa (ATCC 9027) — A Gram-negative bacterium with a notorious ability to survive in low-nutrient aqueous environments. The reason eye product contamination is treated as a Category A risk — P. aeruginosa eye infections can cause corneal perforation in severe cases.

Escherichia coli (ATCC 8739) — A Gram-negative enteric organism. Less common as a cosmetic contaminant in practice but included as a broad-spectrum stress test for your preservative system.

Candida albicans (ATCC 10231) — A yeast associated with skin and mucosal infections. Particularly relevant for water-based emulsions and products used in occluded, warm environments.

Aspergillus brasiliensis (ATCC 16404) — A mold. Slow-growing and often invisible until a product has already failed at the spore level. The organism that makes “borderline” preservation failures particularly dangerous in packaged products with long shelf lives.

Each organism is prepared to an inoculation level of approximately 1×10⁵ to 1×10⁶ colony-forming units per milliliter (CFU/mL) and added to separate aliquots of the finished formula. The four-timepoint count structure tracks whether the preservative system is actively reducing the population, holding it static, or — the failure mode — allowing recovery after an initial kill phase.

Criterion A vs. Criterion B: Where Most Brands Get Surprised

This is the section that catches formulators off guard. Under ISO 11930, there are two acceptance criteria, and they are not interchangeable:

Criterion A is the default, stricter threshold. For bacteria, it requires a reduction of at least 2 log units (a 99% population decrease) from the initial inoculum count by Day 7, with no recovery of bacterial counts through Day 28. For yeasts and molds, no increase from the initial count through Day 28. Criterion A applies by default to all product categories and is mandatory for eye-area products, single-dose products, and products intended for use on damaged skin.

Criterion B is a reduced-stringency threshold that requires a 2-log bacterial reduction by Day 14 (rather than Day 7) and no increase in yeast or mold counts through Day 28. ISO 11930 permits the use of Criterion B only when Criterion A “cannot be achieved for technical reasons” and when that limitation is formally justified in the product safety assessment. It is not a fallback you can invoke because your formula didn’t quite make Day 7.

In practice, this distinction matters a great deal for Amazon and retail compliance. A COA showing a Criterion B pass with no documented technical justification will frequently be flagged by compliance reviewers who are familiar enough with ISO 11930 to know the difference. If your formula can only meet Criterion B, you need to document why in your safety file — and “the preservative is natural” is not a recognized technical justification under the standard.

Under USP <51>, the structure is category-based rather than risk-tiered. Most finished cosmetics fall under Category 2 (topical, non-sterile preparations), which requires a 2-log reduction in S. aureus, P. aeruginosa, and E. coli by Day 14, with no increase in Candida or Aspergillus counts through 28 days. The Day 7 data is collected but not a pass/fail gate under Category 2, which is why some brands prefer USP <51> for products with naturally preserved systems — though ISO 11930 has broader international acceptance for cosmetics specifically.

Why Clean Beauty Formulas Are Failing Challenge Tests at Higher Rates

The clean beauty trend has pushed formulators away from traditional broad-spectrum preservatives — parabens, phenoxyethanol at full concentration, formaldehyde-releasing agents like DMDM hydantoin — toward alternatives that are more consumer-friendly on an ingredient label but often less effective in practice.

We’re seeing this in the challenge test data. Water-heavy formulas — hydrosols, essence toners, aloe vera gels, hyaluronic acid serums — using natural or “clean” preservative systems are overrepresented among borderline and failed challenge tests. The problem usually isn’t the preservative ingredient itself. It’s one of three interaction issues:

pH incompatibility. Many organic acid-based preservatives — sorbic acid, benzoic acid, levulinic acid — lose the majority of their antimicrobial effectiveness above pH 6.0 because the active form (the protonated acid) is no longer dominant in solution. Brands that market a “pH-balanced” or “gentle” formula at neutral pH may be inadvertently eliminating the efficacy of their preservative chemistry in the same formulation decision.

Removal of chelator synergy. Disodium EDTA is not a preservative on its own, but it dramatically potentiates the activity of other preservatives against Gram-negative bacteria by chelating divalent metal ions in the outer membrane, making the cell more permeable. Removing EDTA in pursuit of a “clean” ingredient list — without compensating with another chelator like phytic acid or gluconic acid — is one of the single most common root causes of P. aeruginosa failures in challenge tests we see in the lab.

Underestimated water activity. Formulas that include high-water botanical extracts, hydrolyzed plant proteins, or sugar-based humectants at significant concentrations can have a water activity (aw) well above the 0.90 threshold above which most spoilage organisms thrive. Natural extracts also introduce trace minerals and organic compounds that can interact with preservative molecules and reduce their free concentration in solution. A formula modeled at aw 0.87 on paper may test at 0.93 in the finished product with full botanical loading.

None of this means natural preservation is impossible. It means formulating a naturally preserved product that passes Criterion A of ISO 11930 requires deliberate testing — ideally at the prototype stage, not after you’ve committed to a production run.

Getting Your Formula Ready for a Preservative Efficacy Test

A few details that meaningfully affect your test results and turnaround:

Submit at final pH and viscosity. Challenge tests run on an intermediate version of your formula may not reflect how the finished product behaves. A pH shift of 0.5 units or a rheology modifier addition can measurably alter preservative efficacy. We recommend submitting the same batch you’d send for stability testing.

Confirm which standard and which criterion you need before ordering. If you’re selling eye-area products, your COA should document ISO 11930 Criterion A compliance — not just a general “pass.” Retailers and Amazon compliance reviewers increasingly distinguish between the two. Ordering the wrong evaluation tier means a delayed launch while you re-test under the correct criterion.

Consider a bracketed temperature test if you have shipping concerns. Standard PET runs at 20–25°C. If your product will travel through climates where temperatures regularly reach 30–35°C, a supplementary test at 30°C can reveal whether thermal stress reduces your preservative system’s efficacy — some phenoxyethanol-based systems show measurably reduced performance at elevated temperature.

Budget the full 28–35 days. The 28-day incubation is not a formality — it’s the only way to catch late-stage microbial recovery after an initial kill phase. Provisional results at Day 14 can look like a pass while a Candida rebound is already underway. For results that will hold up to retailer and regulatory scrutiny, you need the complete data package.

At Qalitex, we run ISO 11930 and USP <51> challenge tests as part of our standard Cosmetic Testing panel. Most brands receive a complete data report — including log reduction tables for all five organisms across all four timepoints, plus raw plate count data — within 30–32 days of sample receipt. That’s the format retailers and Amazon compliance reviewers expect to see, and it’s structured so your regulatory team can slot it directly into your product safety assessment without additional interpretation.

If you’re not sure which standard applies to your product type, that’s a conversation worth having before you order testing. The method selection affects your timeline, your pass/fail bar, and the downstream usability of your COA.

---

Written by Nour Abochama, Vice President of Operations, Qalitex Laboratories. Learn more about our team

Talk to our team about your testing needs. Contact us

Related from our network

• FDA Audit Readiness and Regulatory Documentation Best Practices — Aurora TIC specializes in FDA compliance consulting, process validation, and audit preparation for regulated industries.
• Raw Material Supplier Qualification and COA Verification for Cosmetic Ingredients — Ayah Labs provides global B2B raw material testing and supplier qualification services, including identity and purity testing for cosmetic actives.