Most logistics professionals are familiar with designing warehouse networks where the objective is to determine how many warehouses a company should have and where they should be located. Establishing the network, however, is not enough. It leaves open too many other questions. Moreover, just adding warehouses to an existing network could actually decrease service to the customer to say nothing of the added cost it would take to operate.
Many logistics executives routinely question their warehouse networks. They often ask how many warehouses they should have and where they should be located. Perhaps they believe their network should contract to fewer warehouses than they currently have or they might believe warehouses should be added.
In either case this is a common question. Large companies seem to analyze their networks about every five years. They typically undertake a network design project to determine if their warehouses are properly positioned. A vast number of articles have been written on how to perform this analysis. Many companies hire consultants to do the project and normally specialized software is used to perform “the network” analysis.
In all but a handful of cases these analyses are not enough. They address the network locations but not all the elements of the supply chain that provide service, and in most cases, they don’t even address the most important of these elements or their complete effect on service to customers.
In most supply chains the lead-time provided customers can be divided into two components – inventory availability and product acquisition time. But acquisition time is only relevant when the inventory is unavailable.
When inventory is available, the time to get product from the warehouse to the customer is just about fixed. It consists of the time to process the order plus the time to transport it to the customer. These times are generally fixed. They don’t vary much. Moreover, customers are generally aware of and used to them. They have ordered products before and know what to expect. Even if it’s a first time order, the new customer has a reasonable feel for how long it should take to process an order and transport it to them.
On the other hand, think about the case when the inventory is unavailable. That’s when the acquisition time becomes important. In that case the customer’s lead-time will include the added time to get the product back in stock or the time to process and ship the product from some other location – another warehouse, a manufacturing plant (with its potential time needed to make the product), the supplier, whatever.
The overall lead-time the customer is subject to depends on inventory availability – often thought of as fill rate. In many supply chains this dependency is substantial. Suppose, for example, a warehouse processes all the orders for which it has inventory in one day and that the average transit time is an additional day. Then on orders for which the inventory is available, the customer’s lead-time is two days.
Now suppose that the inventory availability is 90% and the average time to acquire out-of-stock product is ten days. Then the overall lead-time to customers is 3.0 days (90% of 2 days plus 10% of 12 days). In this case one full day (33% of the total) of the overall lead-time is composed of time incurred due to the inventory availability, or unavailability, if you like.
So we can list the elements that make up the customer’s lead-time – at least in our simplified example. They are the transit time (one day, on average, in our example), the order processing time (also one day), inventory availability or the likelihood of inventory being in stock (90%), and the acquisition time (10 days). Let’s ask which of these elements is dependent on the network, or where warehouses are located.
The answer is that the network, per se, substantially impacts only one of these elements – the transit time from the warehouse to the customer. This transit time generally depends on the distance from the warehouse to the customer, which is directly dependent on where they are located. In most supply chains the average distance (and, consequently, the transit time) decreases as warehouses are added to the network. So, in our example, the network only impacts one day of the 3-day average customer lead-time. That’s only a third of the total!
To make matters worse, think about the capability of the network to decrease transit times by adding more warehouses. In markets that serve customers that are distributed approximately like the US population (which is most, by the way), adding more warehouses impacts the average distance only slightly.
The following table illustrates this. Shown below are the average distances in well-designed networks.
In a four-warehouse network, for example, the most the transit time can be reduced by adding a fifth warehouse is 15.9%. Moreover the transit time is only part of the customer’s lead-time – a third in this case. The added warehouse would reduce the overall customer lead-time by about 5%!
What do we conclude from all this? We conclude that the warehouse network itself is a relatively weak lever in providing service to customers and is certainly overrated in logistics circles. Other levers that may be far more effective are order processing times, inventory availability and acquisition time – especially the combination of all three.
To be sure, the network may have some effect on these components, however, that effect is small and, in at least one case, it’s a contrary effect. Let’s explain. As the number of warehouses increases, inventory availability goes down causing lead-times to increase. (This effect stems from demand variability increasing as territories get smaller.) Depending on the relationships of the two effects (the decreasing transit time and the increasing unavailability), adding more warehouses will actually lengthen the customer’s lead-time.
Most companies account for this effect (decreasing fill rates as inventory is stored in more places) by adding more inventory. But that takes more capital which could be spent in other ways – quicker order processing, faster transit, and lower acquisition times, to name a few.
The “bottom line”: A supply chain’s raison d’être is to provide short lead-times to customers. The traditional network analysis that looks at how many and where warehouse are located is not enough. As a matter of fact adding warehouses might be counter-productive – actually increasing customers’ lead-times. To add insult to injury it would cost more too.
The solution: Warehouse network designers must consider more than just where warehouses are located. They should account for all the elements of the customer’s lead-time. These elements consist of at least the four above: transit time, order processing time, inventory availability, and acquisition time.