How to Test Skincare Formula Stability Before Launch: A Protocol

What "Stable" Actually Means: Five Ways Products Fail

Stability is not one property. It is the absence of five different failures, and a formula can pass four and still be unsellable.

**1. Physical separation.** The emulsion breaks. You see creaming, a layer of oil at the top, syneresis (water weeping out of a gel), or sedimentation of a suspended powder. Usually driven by an inadequate emulsifier system, wrong processing temperature, or electrolyte load.

**2. Chemical change.** Actives degrade, pH drifts, fragrance oxidises, colour shifts. Ascorbic acid browning is the textbook case. A product can look fine and no longer do what the label claims.

**3. Microbiological failure.** The preservative system cannot cope with real-world contamination, especially in jars, or it is adsorbed by the packaging and drops below effective level over time.

**4. Package interaction.** Fragrance or solvents attack plastic. Oils soften a liner. A surfactant lifts printed decoration. A formula swells a dip tube or a gasket until the pump fails. The formula is intact; the product is dead.

**5. Sensory drift.** No measurable failure, but after four months the texture is draggier and the scent is flat. Customers notice this before any instrument does, and repeat purchase is where it shows.

A proper stability programme tests for all five. Many brand owners test only the first and are then surprised by the other four. Before anything else, write down which of these five would be fatal for your product — that decides how much testing is proportionate.

Building the Test Plan: Samples, Conditions and Timepoints

A stability study is only as good as its design, and the design is a conversation you have with your **OEM manufacturer in Thailand** or your contract lab before anything goes into an oven.

**Samples.** Test the pilot batch, not a beaker sample — the pilot is made with production equipment and real packaging. Fill enough units for every condition and every pull point, plus spares for retests. Keep a retained reference sample stored cool and dark; every later judgement is made against it.

**Conditions commonly used in cosmetic stability work:**

• **Accelerated**, typically around 40°C with controlled humidity. The workhorse condition for predicting long-term behaviour.
• **Ambient**, around 25°C, as the real-world control.
• **Cold**, around 4°C, to expose crystallisation and thickener failure.
• **Freeze–thaw cycling**, alternating sub-zero and room temperature over several cycles. Brutal and very revealing for emulsions.
• **Light exposure**, for anything in clear or translucent packaging, or with light-sensitive actives.
• **Elevated stress**, around 45–50°C, as a short screening condition when comparing formula variants.

**Timepoints.** Readings at the start and then at intervals through the study, with the accelerated condition read most often. A three-month accelerated study is a common industry practice for inferring longer-term behaviour, but it is an inference, not a measurement — real-time data at ambient remains the reference.

**One rule above all:** never shorten the study to fit a marketing date. Agree the schedule before you set the launch plan, not after. Durations depend on the formula, the actives and the lab, and are confirmed when the plan is agreed.

What to Measure at Each Pull Point

Decide your measurements and your acceptance limits in advance. Reading results without pre-set limits invites wishful interpretation.

**Every pull, every condition:**

• **Appearance** — colour, clarity, any separation, creaming or sedimentation. Photograph against the retained sample in the same light.
• **Odour** — assessed by the same people each time, against the retained sample. Low-tech, high-value.
• **pH** — at a controlled temperature. Drift is an early warning of chemical change even when the product looks perfect.
• **Viscosity** — same instrument, same spindle, same speed, same temperature. Without that discipline the numbers mean nothing.
• **Centrifugation** — a short spin to expose instability that has not yet become visible. Useful as an early screen.

**At defined pulls:**

• **Active assay** — the actual concentration of the ingredient your claim depends on. Essential if you make an ingredient claim, and the main reason vitamin C products need tighter control.
• **Microbial counts** — total aerobic count, yeasts and moulds, and specified organisms.
• **Fill weight or weight loss** — a steady loss points to evaporation through the pack, not a formula problem.

**Acceptance limits** should be numeric, for example a maximum pH shift, a viscosity band, and a minimum percentage of the declared active remaining. "No significant change" is not a specification; it is an argument waiting to happen.

For **sunscreen OEM** projects there is an additional layer: UV filters and their stabilisers need their own monitoring, and any protection-factor testing is a separate study with its own protocol and timing.

Packaging Compatibility: The Test Most Brands Skip

This is the single most common gap in a first-time brand's testing. The formula is tested in a neutral lab jar, the launch pack is chosen later by the design team, and nobody tests the two together.

Run compatibility in the **exact** pack you will buy — same resin, same decoration method, same closure, same dip tube, same gasket. A "similar" bottle from another supplier is not the same test.

**What to look for over the study:**

• **Weight loss** from the filled unit, indicating permeation or a poor seal.
• **Container deformation** — panelling, swelling, softening, stress cracking at the neck or base.
• **Closure integrity** — does the cap still torque correctly, does the pump still prime and deliver a consistent dose, does the sprayer still atomise.
• **Decoration survival** — ink adhesion, label edge lift, shrink sleeve movement. Test with a wet thumb, because that is the bathroom reality.
• **Component swelling** — gaskets, dip tubes and valves in contact with the formula.
• **Migration and extractables** — relevant for high-solvent or high-fragrance formulas, where a specialist lab is needed.

Test upright and inverted, and include at least one shipping simulation: a vibration or drop test on a filled, cartoned unit. A product that survives an oven and fails a pallet ride has still failed.

This matters disproportionately for alcohol-rich and fragrance-rich products. A **private label perfume** or a high-oil balm will interrogate a pack far harder than a simple lotion, and so will concentrated household formulas in a **home care OEM** programme.

The Microbiological Side: Preservative Efficacy and In-Use Risk

Stability testing tells you whether the formula holds. Preservative testing tells you whether it stays safe once a human being gets their fingers in it. They are different studies and you need both.

**Preservative efficacy testing**, often called a challenge test, is standardised internationally — ISO 11930 is the reference most labs work to. The principle: inoculate the product with specified bacteria, yeast and mould, then measure how effectively the formula reduces those populations over a set period. Run it on each distinct formula, and re-run it whenever you change the preservative system, the pH, the surfactant base, or the packaging format.

**Key points brand owners get wrong:**

• A challenge test on an old formula version does not cover the new one. Any change to pH or preservative load reopens it.
• Packaging format changes the risk profile. A jar invites contamination that an airless pump does not. The same formula may need a different preservative strategy per pack.
• Natural and "preservative-free" positioning does not remove the requirement. It makes the testing more important, not less, and usually narrows the safe pH window.
• Packaging can adsorb preservatives over time, so microbial counts should be monitored in the final pack, not only at the start.

Also consider an **in-use test**: units opened and sampled repeatedly to mimic a consumer's routine. It is not always required, but for jars and for products aimed at sensitive skin it is the closest thing to a real-world rehearsal.

Regulatory expectations around safety assessment differ by market and change over time, so confirm what your destination requires with the relevant authority or a qualified safety assessor.

Reading the Results and Making the Launch Decision

At the end of the study you have a folder of numbers. The decision is yours, not the lab's.

**Judge against your pre-set limits, in this order:**

1. **Any microbiological failure is a stop.** No exceptions, no negotiation, no "it was borderline". 2. **Any package failure is a stop** until the pack is changed and the study restarted on the new pack. 3. **Active below your limit** means you cannot make the claim. Reformulate, change the pack to protect the active, or change the claim. 4. **Physical separation at accelerated conditions** needs a judgement call. Reversible creaming that redisperses on shaking may be acceptable with a label instruction. A broken emulsion is not. 5. **pH or viscosity drift inside your band** is normal. Outside it, treat it as a formula problem even if the product still looks fine.

**Set the shelf life from the data you have**, not from what competitors print. Real-time ambient data is the strongest basis; accelerated data supports an initial estimate that real-time results should later confirm. Period-after-opening is a separate judgement informed by the in-use and preservative work.

**Document everything** in one file: protocol, batch record reference, every raw result, photographs, acceptance limits, and the signed conclusion. This file is what you hand to a regulator, a retailer or a distributor who asks, and what protects you if a complaint arrives two years from now.

The discipline sounds heavy for a first product, and it is cheaper than a recall. If you are about to **start your own brand**, build the testing into the plan from the beginning rather than treating it as a hurdle before launch.