[Herbnerd]

I had this problem recently with an Aloe Vera juice dietary supplement I developed for a customer. Everything looked fine on the shelf ( quickest way to see if anything goes wrong was to put on a shelf in bright sun… Mould shows up very quickly).
Sent this away for preservative efficacy tests and 3/4 products came back fine. The fourth passed after a tweak to the preservatives.

So, we felt fine to release for sale whilst doing on-going stability.

Under accelerated stability the Vit C disappeared within hours in three of the four product. The fourth was still there but lower than the levels claimed.

The difference between the fourth product and the others was addition of Vitamin E (dl-alpha tocopheryl acetate).

The quickest solution to the problem was to add vitamin E to all products and increase the overage of the vitamin C by 120%. Ie 90 mg/serve claimed input amount was 198 mg/serve.

Admittedly I was working with pharmaceuticals rather than cosmetics, but I am sure there will be a cross-over of information that could be useful. We did consider using Ferulic acid, but we wanted a solution that didn’t require product relisting.