ABSTRACT Human milk replacement formulas (HMRFs) are foods designed for babies in their early months of age. An accelerated storage test showed that dark color increased at a higher rate than oxidized flavor. Additionally, a survey among mothers who were giving HMRFs to their babies showed that mothers rarely tasted the product. Thus, the salient sensory attribute they would pay attention to when preparing their babies' bottles was color. If a portion was rejected by the babies, mothers attributed the rejection to stomach trouble and not to an inappropriate flavor. Thus, as a result of the color having increased at a higher rate than flavor changes and being the salient attribute that mothers would pay attention to, it was chosen as the critical shelf life descriptor. The sensory cutoff point was determined from the correlation between mothers' acceptability scores for samples with increasing color and the dark color of the same samples measured by a trained panel. A high correlation was obtained between dark color measurements from a trained panel and CIE-LAB values (L*, a* and b*). The cutoff point was used to determine shelf life based on sensory attributes of a commercial HMRF stored at 25, 37 and 45C, resulting in 150, 43 and 21 days, respectively. Activation energy is calculated. PRACTICAL APPLICATIONS Sensory shelf life determination in foods for babies is not an easy task. This work offers a methodological practice to develop a decision criterion for when the end user cannot be consulted directly.