What a Serum Stability Test Actually Looks Like (and Why It Takes 12 Weeks)

A private label serum stability testing program is a 12-week clock, and it starts the day your pilot batch is filled — not the day your artwork is approved. That distinction has killed more launch dates than any formulation problem we have seen. The clock is not compressible, the intermediate time points are fixed, and a plan that budgets four weeks for stability is not a plan, it is a hope. Below is the protocol we run in our own lab, the numbers we actually read off the instruments, and the two points where the schedule usually breaks.

What a Private Label Serum Stability Testing Program Actually Covers

A stability study produces four separate data sets, and most brands only ever ask about the first one.

Physical data. Appearance, color, odor, pH and viscosity, read at every time point. Color is measured as a delta-E against a retained T0 reference sample, not by eye — “slightly darker” is not defensible in a PIF.

Chemical data. Active content by HPLC, run in duplicate. This is the number that decides whether your label claim survives the shelf life you printed.

Microbiological data. Total aerobic count and the preservative efficacy test (ISO 11930) at day 0, 14 and 28 on the preserved formula, plus a challenge test if the preservative system is new to the plant.

Container data. The study has to run in the exact pack you will ship: fill weight, closure torque, pump prime count, and the container held upright, inverted and on its side. A formula that passes in a glass beaker and turns in an airless pump has not passed anything.

The plan we hand to brands is 12 units per condition per time point, at T0, T2, T4, T8 and T12 weeks. That sampling plan is the backbone of every private label serum stability testing program we run, and it is locked before the first sample leaves the filler. On a two-year claim, the real-time arm keeps running to 24 and 36 months while the accelerated arm has already answered the launch question.

Why the Clock Is 12 Weeks and Not Four

The 12 weeks is not a house rule, it is arithmetic. Reaction rate roughly doubles for every 10°C of storage temperature above ambient. Running at 40°C instead of 25°C gives you about three doublings of rate, which is why 12 weeks at 40°C and 75% relative humidity is accepted as a reasonable proxy for roughly 24 months at 25°C and 60% RH.

Two consequences follow from that.

First, four weeks is not a short version of twelve weeks — it is a different test. Four weeks will tell you whether two ingredients are grossly incompatible, whether the emulsion breaks, and whether the preservative is being destroyed. It tells you nothing about a two-year claim.

Second, 50°C is not a faster answer, and anyone quoting a shorter private label serum stability testing turnaround by raising the temperature is quoting a different test. Emulsions cloud and separate near the cloud point of their surfactants, waxes melt, and some preservatives partition into the oil phase and stop working at temperatures they will never see in a warehouse. We have watched a perfectly good 5% niacinamide base fail a 50°C stress run in nine days that then passed 12 weeks at 40°C without a single out-of-specification result. The false failure cost that brand a reformulation and 14 days. If a supplier offers you a 50°C shortcut to save time, ask what they do with a failure.

The Matrix We Actually Run

One condition is not a study. This is the full set we hold, and what each arm is there to catch.

Condition	Duration	What it catches	Time points
40°C / 75% RH	12 weeks	Oxidation, active loss, color drift, viscosity change	T0, 2, 4, 8, 12
30°C / 65% RH	12 weeks	Warm-climate shipping: Gulf, ASEAN, West Africa	T0, 4, 8, 12
25°C / 60% RH	24-36 months	The shelf-life claim actually printed on the pack	Quarterly
5°C	12 weeks	Cold-chain handling, winter transit, phase separation	T0, 6, 12
Freeze-thaw, -10°C to 25°C	3 cycles, 24 h each	Repeat cold exposure and thaw crystallisation	After each cycle
Preservative efficacy, ISO 11930	28 days	Microbial challenge and log reduction	Day 0, 14, 28

The 30°C arm is the one brands skip and regret. A serum that is fine at 40°C in a chamber can still sweat and separate in a 45°C container on a Jebel Ali dock, because the chamber holds a constant humidity the container does not. If your buyer is in Dubai, Lagos or Jakarta, that arm is not optional.

What Actually Fails in the Chamber

Failures across the private label serum stability testing runs we have logged over the past three years cluster into five groups, and we can usually name the cause before the assay comes back.

Discoloration. A 15% L-ascorbic acid serum is the classic case: the sample goes pale straw at T4 and orange by T12. The fix is not more vitamin E, it is removing dissolved oxygen before filling, nitrogen blanketing the headspace and an airless pack — the same levers we described in the write-up on why vitamin C serums turn orange. The 15% L-ascorbic acid with ferulic base holds its color inside 12 weeks at 40°C only because all three are in place.

Active loss. Retinol is the worst offender we handle. A free retinol serum labeled 0.5% will routinely assay 0.42% at T0 and 0.30-0.35% by T12, a 25-30% loss. You have two honest answers: dose an overage at manufacture, usually 10-20%, and document it, or move to an encapsulated system. On the retinol serum line we run both, and the 0.3% encapsulated retinol starter holds within 10% of label through the full 12 weeks — at roughly USD 0.40-0.60 more per unit at 1,000 pcs.

pH drift. Most formulas move 0.2-0.5 units over 12 weeks. It sounds harmless until the drift crosses a preservative pKa: sorbate and benzoate sit around 4.76, so a product drifting from 4.4 to 5.1 has quietly lost a large share of its preservative capacity. Buffer to hold the pH, or accept a shorter shelf life.

Viscosity collapse. Carbomer gels lose 20-30% of their viscosity at 40°C through polymer hydrolysis, faster in the presence of electrolytes. A 10% niacinamide serum thickened with carbomer at 4,000 mPa·s can read 2,800 mPa·s at T12. It does not look wrong in a photo, but the dropper now drips and the pump now spits.

Container failure. Piston seals swell in some essential-oil systems, silicone-oil lubricated pumps lose prime after 200 actuations, and black butyl dropper bulbs harden at 5°C and stop recovering. All of these appear only when the product is tested in the pack.

When 12 Weeks Is the Wrong Answer

Twelve weeks is the default, not a religion, and insisting on a full private label serum stability testing run where it buys you nothing is its own kind of expensive.

If you are launching from our existing library — a formula that already has 24 months of real-time data in a 30 ml airless pump — and your only change is the label, you do not need a fresh 12-week accelerated study. You need a bridging study: the parent data, the documented delta, and four to six weeks of confirmatory data in your own pack. That is defensible and it saves a month.

If you are selling a six-month limited edition, claim six months. A shorter claim supported by clean accelerated data plus the first three months of real-time data is legitimate in the EU and the UK. Manufacturers who insist on 24 months for a seasonal drop are usually protecting their file, not your schedule.

The case where you cannot shortcut anything is a new active with no published stability profile plus a new pack. There, the accelerated arm tells you almost nothing about photostability or long-term interaction with the elastomer in the pump. Budget 24 months of real time and design the launch around a 12-month claim, then extend it once the data lands.

Factory View: Protocol, Sampling and the Paper Trail

A private label serum stability testing study is only as strong as the protocol written before the first sample is pulled. Ours is fixed before the pilot runs: batch numbers, fill weight to 0.1 g, closure torque, retention sample count, instrument list and acceptance criteria per attribute.

Sampling is 12 units per condition plus 6 retained at -20°C as the T0 reference, taken from one 500 kg production batch. Where a formula is scaled up, we run the stability arm on the pilot only if the pilot uses the same high-shear homogenizer and the same filler head as production — a bench mix tells you nothing about how a 500 kg batch behaves in the stability chamber. Every chamber has a chart recorder logging ±2°C, and every assay is run on the same HPLC with a system suitability check.

The paper trail is what your importer and your notifier actually need: a COA per batch, an MSDS, the full INCI list, a Certificate of Free Sale, and a stability binder holding the protocol, the raw data and the conclusion. Under EU 1223/2009 that binder is part of the Product Information File, and the CPNP notification is filed by your responsible person, not by us. The whole set sits inside the quality and certification stack behind every batch we release, produced under ISO 22716 and GMPC with filling in an ISO 8 cleanroom. If you are choosing between formulas, the 24-formula library marks which ones already carry full real-time data — starting there saves you the longest wait on the sheet.

Practically: sampling is 7 days, bulk is 35 days once the brief is approved, and the minimum is 1,000 pcs per SKU. Stability runs alongside packaging and artwork approval, so it lands inside the 35 days rather than after them — as long as you start it at pilot and not at purchase order.

Common Questions

How long does private label serum stability testing take?

Twelve weeks for the accelerated arm plus the real-time arm for the claim you want to print. Bridging studies for an existing library formula run four to six weeks. Nothing legitimate gets you a two-year claim in four weeks.

Can I use a stability report from another brand?

No. Stability data is specific to the batch, the container and the fill. What you can reuse is the parent formula data as the justification for a bridging study in your own pack, with your own elastomer and your own closure torque.

What happens if a batch fails at week 8?

We stop the run and diagnose before reformulating. Nine times out of ten it is dissolved oxygen, an overage that was never dosed, or a preservative that has drifted past its working pH. Fixing one of those costs a few days; changing the polymer because nobody measured pH costs a month.

—

Send us your brief with your target market, shelf-life claim and container, and we will come back with a stability protocol, a unit price and a production slot within 48 hours. Sampling is 7 days, bulk is 35 days, and the minimum is 1,000 pcs per SKU.