Abstract
A long-term stability study of a manufactured substance, material, or object (a product) is a known task in analytical chemistry (testing) laboratories serving applied science, industry and trade. It is also an important part of development of reference materials as measurement standards. The main outcome of a long-term stability study is a conformity decision at the point on the time scale when the product property value is expected to violate its tolerance/specification limit. The period of time (the interval on the time scale) from the product fabrication up to this ‘tipping point’, determined with the specified confidence probability, is recognized as the product shelf life or lifetime. After the tipping point, the product is deemed unfit for its intended use. Using a Bayesian approach, the present position paper (a perspective) considers tipping point analysis as a statistical tool to estimate the shelf life of a substance, material, or object, taking into account risks of false decisions in its conformity assessment.