Temperature distribution tests are normally performed to identify the coldest location in batch retorts in order to design safe processes. This work presents an approach to analysing temperature variability by considering its potential impact on lethality variability. The procedure is illustrated with the analysis of a four basket water cascading retort in both static and rotary modes, and is based on the numerical calculation of the F-values from the experimental retort temperature histories, neglecting all other sources of variability. The temperature profiles were monitored basket by basket (thirty thermocouples in each basket, one at a time) and also in the whole retort (twenty thermocouples in four baskets). The location of the coldest point depended on whether only the heating phase or the complete cycle were considered. Replicates indicated that batch to batch variability can be a major factor.