Rogue seasonality in supply chains: an investigation and a measurement approach

Abstract

Purpose – Shukla et al. (2012) proposed a signature and index to detect and measure rogue seasonality in supply chains, but which, however, were not effectively validated. The authors have sought to investigate rogue seasonality using control theory and realistic multi echelon systems and rigorously validate these measures, so as to enable their application in practice. The paper aims to discuss these issues. Design/methodology/approach – Frequency domain analysis of single echelon and simulated four echelon Beer game system outputs are used in the investigation, with the simulation incorporating realistic features such as non-linearities from backlogs and batching, hybrid make to order-make to stock ordering system and the shipment variable. Lead time, demand process parameters, ordering parameters and batch size are varied in the simulation to rigorously assess the validity of the index. Findings – The signature based on the cluster profiles of variables, specifically whether the variables cluster together with or away from exogenous demand, was validated. However, a threshold for the proportion of variables that could be clustered with exogenous demand and the system still being classified as exhibiting rogue seasonality, would require to be specified. The index, which is derived by quantifying the cluster profile relationships, was found to be a valid and robust indicator of the intensity of rogue seasonality, and which did not need any adjustments of the kind discussed for the signature. The greater effectiveness of the frequency domain in comparison to time for deriving the signature and index was demonstrated. Practical implications – This work enables speedy assessment of rogue seasonality in supply chains which in turn ensures appropriate and timely action to minimize its adverse consequences. Originality/value – Detailed and specific investigation on rogue seasonality using control theory and Beer game simulation and rigorous validation of the signature and index using these methods.