24 May 2022
It is fairly uncommon for us to receive complaints, but this complaint about predictive modelling of microbiological shelf life – one of our areas of expertise no less – was justified.
We incorrectly underestimated the increased resistance to preservatives of a particular yeast, in a particular matrix, under specific storage conditions. Our calculation models for this matrix, combined with the specific storage conditions, have since been modified.
But, as the saying goes: every cloud has a silver lining. By relating this complaint to the number of predictive modelling advices of this type over the recent period, we were able to calculate the reliability of our advices. We came to a reliability of more than 98%. Admittedly, a result we are secretly a little proud of.