07 Nov Viewpoint: Data sharing as an important success factor for logistics optimization
By Eirini Spiliotopoulou, assistant professor of Supply Chain Management at VU University Amsterdam
Technological developments are making inventory management increasingly easier. But can technology overcome inefficiencies introduced via human errors and decision-making processes? Eirini Spiliotopoulou, assistant professor of Supply Chain Management at VU University Amsterdam, discusses her research on how firms should solicit information from local managers to set inventory levels efficiently.
Consider a company that has multiple stores that are replenished from a common distribution center (such as ZARA or Albert Heijn). The inventory is held centrally and sent to the stores based on customer demand.
In such distribution systems, it is often the case that each store is allocated a minimum inventory level to ensure income from operations, local competitive power or some customer service level. Allocating dedicated stocks is a common practice, not only in the retail sector but also in the service parts industry where expensive, low-demand parts benefit from (virtual) stock centralization (i.e., inventory pooling).
Each of these stores is managed by a local (store) manager that has more information about his store’s demand (knowledge or feel about market conditions, customer preferences, and so on). The central planner looks to the regional managers to share their more accurate, local demand information to set inventory levels.
However, soliciting credible local demand information to inform the inventory decision is challenging. Discrepancy in incentives may result in unreliable forecast sharing, leading to suboptimal inventory decisions and lower total system performance. Forecast manipulation is widespread in many industries: stores often order more than they need to gain a favorable allocation or inflate their forecasts to assure supply.
In our research, our team used game-theory models to formally show that incentives are misaligned and truthful information sharing is not an equilibrium. Dedicated inventories do not ensure credible information sharing.
We explored two potential ways to improve the system. The first was for local managers directly place orders (rather than simply passing demand forecasts). In the second option, a pricing scheme for inventory reallocations is imposed.
Under direct ordering, local managers place orders based on their true forecasts, although the requested order quantities are not system optimal (e.g., some inventory pooling benefits are lost). Hence, a trade-off exists between central coordination and local demand information.
The second option showed somewhat different results. We found that an inventory transfer pricing mechanism that restores demand truth-telling and maximizes total company profits may exist under certain conditions. This policy can have practical benefit in cases where there is frequent information asymmetry across decision makers. However, coordinating prices do not exist when locations have different profit margins and reallocation costs are negligible.
Comparing the two solutions, numerical study shows that for high margin products direct ordering results in lower inventories, while the reverse is true for low margin products.
Hence, when coordinating prices (best solution) are not possible, direct ordering is the preferred choice for the company if local managers’ knowledge about store demand is quite important and profit margins are high (for example fashion items). This approach simultaneously maximizes profit and minimizes inventories. When the value of local information is low (for example groceries) central coordination of ordering even on limited information results in higher profits, especially when profit margins across stores differ (for example due to location specific store operating costs).
Eirini Spiliotopoulou is assistant professor of Supply Chain Management at VU University Amsterdam. She obtained her PhD and M. Eng. at MIT-Zaragoza International Logistics Program. Her main research interests are supply chain coordination and information sharing, studying both economic incentives and behavioral factors. Her work has been published in leading academic journals in her field.