Every capsule brief eventually produces the same question: what shelf life can you claim? And most brands arrive expecting a number that comes straight out of an accelerated study. It does not work that way, and understanding why saves a good deal of argument later.
Accelerated testing is genuinely useful. It is also the most over-interpreted document in a cosmetic launch, and for a softgel it has one specific blind spot that is worth knowing before you commission it.
Accelerated testing measures change, not a date
An accelerated study puts the finished packed product into stressed conditions - commonly elevated temperature with controlled humidity - and pulls samples at intervals to measure what has changed. Typical cosmetic regimes borrow from pharmaceutical practice: a long-term arm at ambient conditions, and an accelerated arm at a higher temperature and humidity for a defined period.
What comes out is a set of measurements at several time points. Colour, odour, viscosity, peroxide value, active content where there is one, shell condition, microbial status. What the study does not produce is a calendar expiry date, because translating stressed conditions back to ambient storage is an estimate rather than a measurement.
There are rules of thumb in circulation - the idea that a set number of months at elevated temperature corresponds to some multiple at ambient. They are used as screening heuristics and they are not a substitute for real-time data. If you want to print a two-year or three-year date, you need real-time data to support it, which means opening that arm early.
The softgel blind spot
Here is the specific caveat for capsules. A softgel shell softens as it warms. Push the temperature high enough and you reach conditions where the shell material itself starts behaving differently - capsules can stick together in the pack, deform against each other, or weep at the seam.
That is a real problem and a misleading one. Those failures may be artifacts of the stress condition rather than anything that would ever happen in a warehouse. Meanwhile the failure that is most likely to actually reach a customer - a gradual loss of actives, or a shell that slowly embrittles - is not accelerated in the same way by heat.
So ask your manufacturer directly: what conditions does this study use, and does the shell tolerate them? A study run at a temperature the shell cannot handle tells you nothing useful about either variable. Our humidity control article explains why the shell is this sensitive, and our seam integrity article covers how failures present.
What to measure at each pull
A cosmetics shelf life study is a set of chosen parameters, and you decide which. For a capsule serum the useful list looks like this:
- Appearance and odour - cheap, and catches most real failures. Include the shell, not only the fill.
- Shell condition - softness, brittleness, stickiness, any change in how it opens.
- Fill characteristics - viscosity, colour, separation if the fill is not a single phase.
- Oxidation markers - peroxide value matters for any fill containing polyunsaturated oils.
- Active content - for actives with an assay. See why this matters in our peptide capsule article.
- Microbial status - only relevant if water is present in the fill, but essential if it is. See our preservative article.
- Packaging - any interaction between the fill and the blister or bottle over time.
Measurement intervals vary, but the shape is always the same: a baseline reading, then pulls over months. Three points spread sensibly across the period tell you far more than two clustered at the end, because you are looking for a trend rather than a pass mark.
Real-time and accelerated run together
The practical structure most brands should use: run an accelerated study to screen the formula and packaging quickly, and start a real-time arm at the same conditions you expect in distribution at the same time.
Launch on the accelerated data if it is clean, keeping the claim conservative, and extend the printed shelf life later once the real-time arm has the months behind it. This is how sensible brands get to market without either over-promising or waiting three years.
The mistake is launching on accelerated data with no real-time arm open. Then when someone asks for the supporting data in two years, there is none, and the product is stuck at its launch claim.
The pack is part of the product
A capsule that keeps beautifully in bulk can fail in its pack, because the pack controls what the shell experiences.
A blister with a poor moisture barrier lets the shell equilibrate with the room. In a humid market that softens it; in a dry one it embrittles. A bottle that is opened daily exposes the remaining capsules repeatedly. Neither is visible if you test the capsule on its own.
Test the finished packed product, always. Our packaging compatibility article covers what to commission, and our blister article sets out how barrier is specified.
What a three-year claim actually requires
Long shelf life claims are not freely available. In several markets there is a threshold above which the label has to carry different information - in the EU, for instance, a durability period over thirty months means the pack carries a period-after-opening symbol rather than a use-by date. There is also the practical question of whether you want to print a date at all.
A short printed date is sometimes the better commercial choice. It creates a reason to repurchase, avoids arguments about old stock, and asks less of your data. Our labelling guide covers how this differs by market.
Typical stability conditions
| Condition | Purpose | What it catches |
|---|---|---|
| Long-term, ambient | The real answer | Everything, slowly - the only basis for a long claim |
| Elevated temperature with controlled humidity | Screening | Chemical degradation trends; may create shell artifacts |
| Low temperature | Cold-chain and winter shipping | Clouding, thickening, shell embrittlement |
| Freeze-thaw cycling | Distribution stress | Phase separation; seals that fail after expansion |
| Light exposure | Where the pack is not opaque | Colour change in the fill and the shell |
Note which row is described as the real answer. Everything else is a proxy.
Where the schedule meets reality
Stability takes calendar time, and calendar time does not compress. This is the single most common cause of a capsule launch slipping, because it usually sits at the point in the project where everything else has already been decided.
Start it as early as you can tolerate. A pack-and-forget real-time arm running quietly in the background while you finish artwork costs nothing but the samples, and it is the difference between being able to extend a claim later and not. The full sequence and realistic durations are laid out in our launch timeline article.
Reading a stability report
Three things to check before accepting one.
Is it the finished packed product? Bulk-fill data is not enough, for the reasons above. Are there actual numbers at each pull? A report that says "pass" at each interval tells you nothing about the trend; you want values. Does it cover your market conditions? A study designed for temperate Europe does not cover Southeast Asia.
If any of those three is missing, ask. Our stability testing guide sets out what a properly specified study looks like.
What to ask your manufacturer
- Which conditions do you recommend for this shell, and why?
- Will you run the study on the finished packed capsule?
- Which parameters, at which intervals, reported as values?
- Can we open a real-time arm now, even if we launch on accelerated?
- What claim can we support today, and what would we need to extend it?
The last question is the one that gets asked least and matters most. Knowing the answer at the start shapes what you print.
Where our stability capability fits
We run stability on finished packed capsules across the standard conditions and report values at every pull. Our irregular purple capsule and other shaped formats are tested in their actual production blister, so the barrier and the shell are validated together.
We work with brands on private label (your brand on a proven base), OEM/ODM contract manufacturing, and full custom formula development - whichever matches where your brand is today.
Tell us your target markets and the claim you would like to print, and we will tell you what data supports it today and what would support more. Start at our inquiry page, email capsulebiotech@163.com, or message the right team on WhatsApp: Capsule Series (capsules only) +86 158-7655-5319, Cosmetics Series (serum, hair care, body care, fragrance) +86 158-7655-5319.
General manufacturing information, not regulatory or legal advice. Claim limits vary by market.