Accounting for deficit and a known demand quantile in ABC-XYZ analysis

Abstract

Product deficit typically appears as a result of inadequate planning and unforeseen supply chain disruptions. A traditional approach to demand distribution estimation does not account for possible deficit; deficit generates right censored data. If the censoring is ignored, the demand estimation becomes negatively biased. Consequently, all analytics based on the demand distribution also becomes biased. In this article, the well-known survival analysis Kaplan-Meier estimator is applied to demand data with observed deficit. In addition to the reduction of bias associated with the use of the KaplanMeier estimator, we also modify it with a known quantile of the demand distribution. Monte-Carlo simulation studies show that the use of additional information in the presence of censored observations substantially improves estimation quality. The estimators modified with quantile information led to new product grouping in the ABC-XYZ analysis. An illustrative example shows the impact of new estimators on ABC-XYZ grouping.