Will a Centralised DC Help Lower Inventory Levels?
The conventional wisdom on inventory management is that centralising inventory in a single place will lower inventory levels (primarily safety stock levels). While this is true in some cases, when certain criteria are met, the reality might be quite different.
First, in order for the inventory levels to be lowered at all, the products that are stored at the central warehouse must be sold at in regions supported by more than one of the locations being consolidated. If this is not true, then there is no demand consolidation, and consequently, no improvement in forecast accuracy or demand variability through consolidation. This tends to be more of an issue in country specific warehouses where each country has different labeling requirements.
Most practitioners understand this. In fact, there are ways to make this true, through SKU harmonisation or postponement of labeling, which has been used to transform products that were country specific.
A fact remains that centralisation of a portion of a portfolio is not usually economical – so that if the centralisation does take place, all SKUs, harmonised or not, are likely to be centralised.
The second point is a little more subtle. The lowering of the inventory levels is also dependent upon the correlation of forecast error across the locations. In simplistic supply chain models, a normal curve is used, with the assumption that the standard deviation will tell you how much safety stock is required to meet a certain service level (on a dosage, unit or case fill basis). As the forecast is only an estimate, the assumption is, that it is as likely to be low, as it is to be high. Therefore, by aggregating demand, forecasts that are low are combined with forecasts that are high, thereby reducing the expected error.
The problem with this logic occurs if the correlation of the forecast errors is greater than zero. In other words, if actual demand is higher than forecast in one region, how likely is it that the demand is higher than forecast in another region? If there is no relationship across these regions, then the theory above works. But, if the same wrong assumptions are made by forecasters for the different regions, then the errors are additive. In the worst case, there is a correlation of "1" and the safety stock required in the central location is the sum of the safety stocks that are needed in each of the locations being consolidated (no inventory savings).
Third, a central location will typically add some transportation lead time for the shipment of orders; some import/export process lead time might also be necessary as well as some labeling time if postponement of labeling is one of the methods being used (to address the concerns described above). As customer requirements are unlikely to change due to a centralisation of a supplier’s distribution network, this will reduce flexibility and lead time to react to demand and, consequently, increase safety stock levels (as safety stock is based upon variability over the lead time, which is increased as the time between receiving an order and needing to ship it is decreased). Moreover, this may also have the effect of increasing either inventory levels or days sales outstanding (DSO), depending upon accounting practices and recognition of the sale, as the product will be in transit to the customer for a longer period. Depending upon the location of the central warehouse with respect to sourcing facilities, this might be countered by a reduction in lead time from supplier/factory to distribution centre.
Note that the above is not a complete view of the reasons to move to a central distribution facility – there may be other logistical or transportation savings, or even potentially tax savings (especially if some value is added to the product at this facility and the facility is in a low tax jurisdiction). Also, the centralisation of other functions, such as supply chain planning, distribution and customer service can have other added benefits and might be tied to this type of project. Nevertheless, to assume that "centralisation" will automatically mean lower inventory levels is not always correct, and requires some analysis by those who will be held accountable for inventory levels.