Article Preview
Top1. Introduction
The supply chain management is increasingly including both theoretical methods and practical recommendations in order to draw more benefits for all supply chain links (Min et al., 2019). Nowadays, a strong tendency to analyze the sustainable performance of supply chains has taken place in the fields of research and industry. While it is not always evident to derive continuous supply chain savings, the directive line in the supply chain management is often the proposal of adapted strategies and techniques that may contribute to supply chain sustainability, such as efficiency-based, innovation-based, or closed-loop strategies (Khan et al., 2022a).
While the benefits of information sharing policies and systems’ integration in the supply chain field such as supply resilience and sustainability are well-known (Khan et al., 2022b), there is a lack of study of how practitioners may achieve higher global performance without complete integration and critical information sharing. Indeed, privacy policies are continuously persisting in many industries. As it was reported in some papers, financial constraints, lack of information systems’ compatibility, lack of trust, unwilling of sharing risks and rewards, and information protection laws in different countries, are always leading to sub-optimal solutions which affect the whole supply chain’s performance (Cetindamar et al., 2005; Cai et al., 2010). In this context, collaborative forecasting is one of the most distinguished organisational tools that allow performance improvements such as sustainability improvement (Shoukohyar & Seddigh, 2020). Collaborative forecasting refers to the situation where supply chain actors collaborate, either by coordinating forecasting methods or sharing forecast information, in order to achieve higher supply chain’s performance. Many authors discussed the impacts of collaborative forecasting on different supply chain industries such as the renewable energy sector (Nam et al., 2020), the defense sector (Kim et al., 2015) or the food sector (Eksoz et al., 2014).
One of the most recent collaborative forecasting strategies is called Downstream Demand Inference (DDI). DDI emerged in the supply chain theory in order to better control the forecasting processes in decentralized supply chains, where demand information is not shared. It proposes balanced solutions for actors who specifically don’t want to dislock their strategic demand information. DDI refers to a collaborative forecasting strategy mainly between two supply chain links: an upstream actor and a downstream actor receiving the demand of a customer. Some works already provided an initial vision on this coordination strategy (Ali and Boylan, 2012; Ali et al., 2017). More specifically, the demand inference refers to the mathematical deduction of the customer’s demand arriving at the downstream level, based only on the downstream orders’ process arriving at the upstream level. The mathematical deduction supposes that the propagation of the demand across the supply chain is unique, which is true under some known assumptions. The works of Ali and Boylan (2011; 2012) showed that DDI is not possible with Single Exponential Smoothing (SES) or Minimum Mean Squared Error (MMSE) forecasting methods. Next, the papers of Ali and Boylan (2011) and Tliche et al. (2019; 2020) showed that the adoption of either Simple Moving Average (SMA) or Weighted Moving Average (WMA) forecasting methods at the downstream level, insures unique demand propagation at the upstream level. This leads to the following conclusion: either with SMA or WMA, for a well-known orders’ process at the upstream actor, there exists one unique demand process arriving at the downstream actor. Consequently, it is possible for an upstream actor to estimate and reconstruct the demand process arriving at his formal downstream actor without need of demand information sharing. This research has opened the door to promising new directions on which the authors have aligned.