
Every formulator knows the stability study: appearance, viscosity, pH, colour, odour, microbial count, challenge test. It runs at room temperature and under accelerated conditions, the table gets filled in, the PAO is set. It is well-established practice and it works for what it was designed to do: show that the product stays intact. It does not show that it is still active.
Physical stability says the emulsion does not separate. Chemical stability says the assay of the active has not dropped below a set threshold. Microbiological stability says the product does not get contaminated. The fourth — efficacy stability — says the active still does what it was put in the formula to do. It is the only one that supports the claim, and it is the one almost nobody measures.
The distinction is not academic. Analytical assay counts molecules; an efficacy test measures what they do. A gap can open between the two, and it opens above all when the active is biological.
For a protein, an enzyme or a peptide, function depends on conformation and not only on presence: partial denaturation leaves the assay value almost unchanged and takes the activity to zero. An extract or a postbiotic is a mixture, and the component carrying the effect can degrade before the others while the analytical marker chosen for control stays stable.
On top of that there is everything the formula does to the active over time: oxidation, hydrolysis, pH drift, interaction with surfactants and chelators, adsorption onto the packaging walls. These are slow phenomena, invisible at T0, and none of the other three checks catches them.
Without inventing anything: you repeat the same assay that generated the claim on the samples of the stability study already under way. Same matrix, same protocol, same controls, at the time points that matter — typically T0, three months accelerated and end of shelf life at room temperature. The comparison is internal to your own data, so no absolute reference value is needed: what is needed is that the activity level at the end of the period is still inside the range the claim was built on.
There is one practical condition, and it has to be decided in advance: that the samples are planned and kept. That is why this check belongs at the start of the stability study, not when the product is already on the shelf.
The claim is born at T0. Efficacy was shown on the fresh sample, often on the pure active, and then carried over to the finished product for its whole commercial life. It is the most common extrapolation and the least documented.
The formula changes later. A preservative swapped, a pH adjusted by two tenths, new packaging for marketing reasons: changes that pass as minor and that, for biological activity, may not be minor at all.
The samples are gone. The most banal and most frequent case: when the question arrives — from a customer, a distributor, an authority — the stability study is closed and the samples no longer exist.
Little, if the assay already exists: it means running it two or three more times on samples that are in stability anyway. The real cost is the other one: finding out after the fact that the claim is no longer supportable, once the product is launched, the artwork printed and the dossier delivered.
Does the claim hold for the whole life of the product, or only for the sample that was tested? That question decides whether this check is needed. Which assay generated the claim, and can it be repeated on the finished formula? And are stability samples already planned at the right time points, or do they have to be added now?
In cosmetic projects we treat efficacy stability as an extension of the stability study, not as a separate study: sampling times are agreed against the plan that already exists, and the efficacy assay is re-run on those same samples. The result is one more table in the dossier — the only one that says for how long the claim is true.
The page on microbiology and stability describes the checks available on the product; the page on cosmetic effects collects the efficacy assays that can be repeated over time. If the stability study has already started, the later sampling points can still be adjusted.
Technical content and news from our laboratories, sent when there is something worth saying.