Cosmetic Stability Testing and Shelf Life in a China OEM

Stability testing is how a cosmetic OEM proves your product stays safe, effective and unchanged for the shelf life printed on the pack. It runs two ways: real-time testing, where samples are held at normal storage conditions for the full claimed life, and accelerated testing, where stress conditions such as elevated temperature are used to predict problems faster. A credible shelf-life claim rests on real-time data, with accelerated results as an early-warning screen. When you brief a China factory, the questions that matter are simple: what protocol will they run, against what specification, and will you receive the report. Get those agreed before sampling and your launch timeline and your label claim both rest on evidence rather than a promise.

What does cosmetic stability testing actually check?

Stability testing watches a finished formula over time and asks whether it still meets its specification. The checks fall into a few groups. Physical: does the colour, odour, viscosity, pH or texture drift; does an emulsion separate, a cream weep or a suspension settle. Chemical: do active ingredients degrade, does the fragrance shift, does the pH move out of range. Microbiological: does the preservative system hold up, which is confirmed separately through preservative efficacy ("challenge") testing. Packaging compatibility: does the product interact with its container — staining, swelling, scent pickup or leaching — because a formula is only stable inside the pack it actually ships in.

The international reference point is ISO/TR 18811, the guidance document on stability testing of cosmetic products. It is a guideline rather than a fixed pass/fail recipe, which is the key thing buyers miss: there is no single legally mandated test duration or temperature. The manufacturer designs a protocol appropriate to the formula, the packaging and the markets it sells into, then documents the rationale. That flexibility is exactly why you should see the protocol, not just the conclusion.

Real-time vs accelerated: how long does it take?

Real-time (or "long-term") testing holds samples at the product's intended storage condition for the entire claimed shelf life. If you want to print a multi-year life, real-time data ultimately has to back it — there is no shortcut that fully replaces it. That is why mature brands begin stability studies early and keep them running in parallel with launch.

Accelerated testing exists to give you answers sooner. Samples are held at elevated temperature — a commonly used condition is around 40 °C — and often at additional points such as a chilled condition and room temperature, then pulled and tested at intervals over a period of weeks to a few months. The logic is that heat speeds up many degradation reactions, so a formula that survives the stress is unlikely to fail quickly in normal storage. Freeze-thaw cycling and a light-exposure check are frequently added to mimic real-world abuse during shipping and on-shelf display.

The trap is treating accelerated data as a finished shelf life. It is a predictor and a screening tool, not proof. A formula can pass an accelerated panel and still drift over real time, and some changes (certain fragrance or natural-extract shifts) simply do not accelerate cleanly with heat. Use accelerated results to make a confident provisional claim and to catch failures before tooling is committed; let real-time data confirm the number you ultimately print. A disciplined OEM will tell you which of your data is predictive and which is confirmed.

Shelf life vs PAO: which number goes on the pack?

Two different numbers describe how long a cosmetic lasts, and buyers routinely conflate them. Shelf life is how long an unopened product stays within specification from manufacture. Under the EU Cosmetic Products Regulation (Regulation (EC) No 1223/2009), products with a minimum durability of 30 months or less must show a best-before date; the rules differ by market, so confirm what each destination requires.

PAO — Period After Opening — answers a different question: once a consumer opens the jar, how many months does it stay good? It appears as the open-jar symbol with a figure such as "12M". Products durable for more than 30 months typically carry a PAO instead of a best-before date. PAO is informed by in-use stability and the strength of the preservative system, not by the unopened shelf-life study alone — which is why a product can have a long shelf life but a short PAO, or vice versa. When you brief an OEM, state both the shelf-life target and the PAO you want to print, because they are supported by different evidence and the factory needs to design the study to deliver both.

How do you write stability into a China OEM brief?

Stability is cheap to specify up front and expensive to retrofit after tooling. Put it in the brief in plain terms. State the markets you will sell into so the OEM can match the protocol to the right regulatory regime. Name the storage conditions your product will actually face — a sunscreen for hot climates and a balm for temperate retail are not the same study. Ask explicitly for the test the product will live in, including the final primary packaging, not a neutral lab jar, because packaging compatibility is where otherwise-stable formulas fail. Confirm whether preservative efficacy testing is included, and who runs it. And settle the deliverable: you want the protocol, the pull schedule, the acceptance criteria and the signed report — not a verbal "it's stable."

This is also where a factory's documentation discipline shows. An OEM with real GMP personal-care experience, like Aozi Cosmetics in Hebei, treats a stability protocol and its report as standard project paperwork rather than an extra you have to chase — and that paper trail is what your own importer and regulatory file will lean on later. Before you commit, verify those systems the same way you would any supplier: a documented quality process is a procurement question, not a leap of faith. Our China factory audit checklist covers how to check that a factory's quality and documentation systems are real rather than claimed.

How do you read a stability report?

When the report arrives, read it like a buyer, not a spectator. Start with the protocol section: which conditions were used, how many time points were pulled, and what the acceptance criteria were. A report that lists only "pass" with no criteria is asserting a conclusion, not evidencing one — you want to see the target ranges for pH, viscosity, colour and odour and the actual measured values at each pull, so you can judge whether a result sat comfortably in range or scraped the edge.

Next, check what was tested against what was promised. The study should name the formula version, the batch, and the exact packaging. If accelerated and real-time conditions were both run, the report should separate them and say plainly which findings are predictive and which are confirmed. Look for the supporting work too: preservative efficacy results, and where relevant a packaging-compatibility note. A drift that stays inside the acceptance range is normal and expected; what matters is a trend heading toward a limit, which is your early signal that the shelf life or PAO needs to be conservative.

Finally, weigh the gaps honestly. A formula with only a short accelerated dataset and no real-time history can still launch, but the claim it supports should be provisional and the real-time study should continue after shipping. A good OEM will state these limits for you rather than paper over them — and that candour is itself a quality signal worth more than a glossy one-page summary. If you are comparing suppliers, the way each one talks about the weaknesses in its own data tells you as much as the data itself.

How do you verify stability claims before you ship?

A report is a starting point, not the end. Tie stability evidence to what actually leaves the line. The retained reference samples from your stability study should match the production batch in formula, fill and packaging — if the study was run on a lab-scale batch in different packaging, its conclusions are weaker than they look. Check that batch records, certificates of analysis and the stability file are internally consistent, and that the best-before or PAO marking on the pack matches the data behind it.

Then confirm it physically before goods sail. A pre-shipment inspection is the moment to check labelling, batch coding, fill levels and packaging integrity against the spec the stability study assumed. Our guide to product inspection in China before shipment walks through what to verify at that gate. Stability testing and inspection are two halves of the same assurance: one proves the formula holds over time, the other proves the units in the carton are the ones the data describes. Treat them as a single evidence chain — formula, study, batch record and inspected goods all pointing to the same specification — and the shelf life on your label becomes a claim you can stand behind in front of a retailer, an importer or a regulator.

ChinaMakersHub connects global buyers with verified manufacturers across China's Greater Bay Area and beyond. Submit an inquiry to get introduced to vetted factories in your category.