Multi-level supply chains are the norm in many industries. The need to establish such complex supply chains with central and regional warehouses has been reinforced by several trends in recent times. Companies have to respond to increasingly individualized customer requirements, which has led to an almost explosive growth in the range of articles in recent years. As a result, warehouses must provide ever greater capacities for picking and storage. In addition, customers' requirements for delivery speed and article availability are constantly growing. At the same time, free warehouse space and land have become increasingly scarce commodities in recent years, especially in metropolitan regions.

These trends mean that along with the regionally installed warehouse levels, it may make sense to set up and operate a central warehouse, or, conversely, to introduce regional warehouse levels starting from the central warehouse.

By setting up central warehouses, companies can free up capacity in regional locations, reduce inventory costs through risk pooling, and bundle the product portfolio, which makes the use of modern warehousing technologies economically viable. In addition, this can result in cost advantages through transport and purchasing bundling.

In addition to these advantages, however, the extra stage in the supply chain also creates additional complexities. For example, transports and cross-docking activities have to be organized and financed, and customer and article groups have to be reassigned in terms of sales.

Due to the multi-level nature of the supply chain network, the question also arises as to which article should be placed in which warehouse? This question, which sounds simple at first, is very complex to answer, but also has great potential. A holistic approach to article allocation ensures that enormous cost optimization potential can be leveraged in the supply chain network.

When introducing a holistic approach to article allocation in a multi-level network, several objectives must be taken into account. For example, articles should be allocated in a cost-optimized manner. This means that transports between warehouse locations should be reduced, warehouse systems should be used efficiently, and inventories should be lowered through risk pooling. In addition, supplier conditions and quantity discounts can be utilized with good allocation. It is not only cost advantages that can result from efficient item allocation. At the same time, articles should be placed in the supply chain in such a way that article availability is increased and fast delivery of articles with short delivery times is ensured. The following figure shows the basic goals of item allocation.

In the strategic planning of a central warehouse, assumptions are first made to roughly divide the range of articles into regional warehouse articles and central warehouse articles. For example, small and valuable articles and those with low sales are centralized. These may be better suited to technical systems on a case-by-case basis, due to lower transport costs and large inventory cost effects. If these assumptions are correct and target-oriented in the strategy development, additional new challenges arise as soon as the network is planned in more detail:

Cost drivers in handling and transport for individual articles must be mapped in detail, since articles with different sizes and sales patterns can vary greatly in their specific costs. This item-specific cost analysis is ensured by setting up a total cost of ownership model.

Numerous cost effects do not arise solely at the level of the individual item. They depend on the effects of transport bundling and purchasing conditions. This results in additional complexities when evaluating costs. Based on experience, these can no longer be calculated in Excel models alone, but require detailed consideration in simulation models.

In order to establish the developed solution in the operative business, more than good calculation models are required. At the same time, it is important to ensure that target systems, processes, and responsibilities are structured in such a way that an optimal item allocation is also permanently established in the supply chain.

With the help of the total cost of ownership model, all costs that an item causes from the supplier to the customer are mapped holistically. The art of depicting the logistical reality in the model is to find the optimal level of detail. Not every "special feature" in handling has to be represented in the model. Instead, it is important that the main processes which are significant cost drivers in reality are reflected in the model. The main processes should include essential handling steps in the warehouse, from goods receipt, to picking, to loading at the ramp. At the same time, transport and inventory costs are also mapped in the model.

The right level of detail also includes the fact that articles are valued differently in terms of costs depending on their relevant properties. Article dimensions and volumes play an important role, both for individual transport costs and for handling costs in the warehouse. Sales volumes and fluctuations are also taken into account, as well as life cycles and item values.

In this way, it can be determined at the item level whether allocation in the central warehouse or regional warehouse makes more sense. An article with a large volume and throughput would cause high costs on longer transports if allocated in the central warehouse, therefore regional storage tends to be the more cost-effective option for such an article. For articles with low throughputs, low volumes and high article values, on the other hand, it can make sense to allocate them in the central warehouse. Here you benefit from risk-pooling effects, which shrink the total stock in the network while maintaining the same service. Additionally, transport costs for small articles are not as significant. Figure 2 schematically shows the effects of the relevant article properties on the allocation decision.

In the real world, allocation decisions can no longer be considered only at the level of the individual item. Particularly in retail, cross-article dependencies arise at the supplier level due to supplier conditions and quantity discounts. For example, the optimal location for an article from a particular supplier may well be the regional warehouse level, while all other articles from the supplier should be placed in the central warehouse. If all these articles are combined in one location, certain quantity thresholds may be exceeded, and significant cost advantages may result from quantity discounts.

While the mapping of costs for individual articles is still feasible in Excel tables, modern simulation models offer the advantage of mapping the aforementioned dependencies and providing decision-making bases for complex issues. In such models, decisions can be made at the item and supplier level and their effects on storage space allocation and costs can be calculated as a whole in the network over longer periods of time.

Once all cost components are mapped depending on the storage level and the article properties, simulation runs can begin. Here, possible article allocations are simulated and tested for the resulting total costs. Even a one-time improved article allocation offers great cost optimization potential.

However, the full potential can only be realized if a model can also be used operationally. It should therefore have access to currently available sales data and also be able to map changes in the product range. Solutions are available here that link to existing systems and make the simulation results usable not only strategically, but also operationally. By accessing current data it is possible, for example, to plan changes in article allocation during the year. This is particularly useful for seasonal articles, for which sales fluctuate strongly over the course of the year. A simulation tool can also help with article allocation when introducing new articles and assortments, such as deciding when an article should be moved from the central warehouse to a regional warehouse.

The integration of an item's allocation must not only take place with regard to existing systems, but also organizationally. In most companies, several departments have interests in allocation decisions. Sales departments like to allocate stock to regional warehouses, in order to be close to the customer. Particularly if the sales department is divided up regionally, a shift of stock towards the central warehouse can lead to conflicts. For example, the question of which department has access to which stock must be clarified. Regional warehouse managers may also have an interest in centralizing certain articles or keeping them in the regional warehouse. For example, sales-oriented targets can lead to high-priced articles being kept in the regional warehouse, even though it is precisely these articles that have the highest inventory optimization potential. When implementing article allocations, it is therefore important to ensure that the focus is on the overall ideal. To ensure that the newly gained cost transparency also leads to better decisions when allocating articles, targets as well as processes and (organizational) structures must be adjusted.