Sourcing orders – deciding when and from where an order should ship – is a hidden gold mine for many companies. Both cost reductions and service improvements are available from just slight changes in the way sourcing decisions are made. Let’s examine a hypothetical but all to common organization, how they source orders and its consequences.

A manufacturer of electrical products distributes a line of about 4,000 items to distributors nation wide. The company has four regional distribution centers. Each customer has an assigned warehouse. The order processing system works something like this.

Upon getting an order, look for products available at the assigned warehouse. If all the items are available, source the whole order to that warehouse, ship the order and invoice the customer.
If none of the products are available, establish a back order. Keep checking for product availability (Number 1 above) tomorrow, the next day and so on.
If some of the products are available and not others, ship those that are available and back order the rest. Process the back order as in 1 or 2 above.
This is the way a host of order processing systems work. What’s wrong with it? Only two things – it cost too much and provides poor service.

To begin with there is nothing worse than a company holding products that their customers have ordered. The customer wants them, the company actually wants to ship them, book the sale and collect the revenue. Yet company after company commits the sin of holding onto orders when it would be to everyone’s advantage to ship them.

Here are some examples.
First, the assignment of territories. Few companies do this well. Territories should be based on some rationale such as low cost or quick service. Low cost territories need to take into account the costs of product acquisition too, not just the cost from the warehouse to the customer. Therefore a warehouse attached to a plant, for example, would have a larger territory than it would if it were supplied from a plant some distance away.

Short lead-time territories, on the other hand, are highly dependent of transit times. These, in turn, are extremely varied. Take Next Day airfreight for example. FedEx gets these parcels from any place in the US to any place in the US overnight – it doesn’t matter where the order is souced from in this case. Moreover, most Next Day air tariffs are not dependent on distance, so it doesn’t matter cost-wise in this case either.

Ground parcels are a different matter. LTL too. With LTL you have the added complication that most carriers have relatively stable patterns from their terminals. Sometimes they service a location 500 miles away better that the one 200 miles away.

Now let’s tackle the very notion of a territory. Why should there be a fixed territory? Why not let the system decide when and from where to ship the order? When you have fixed territories you segregate the inventory to serve only one territory.

Moreover, it’s a smaller territory than the whole country. This has a well-known consequence – it requires more inventory! This is due to the fact that making smaller territories makes more slow movers. If you cut the demand for all products in half say, as you would by carving one territory into two, then you automatically make more slow movers.

If inventory is in one place, it can’t be in the wrong place. However, there’s a way around the territorialization problem. That’s to make all inventory available to all customers.

Consider Figure 1. The territories shown are those that are closest to their respective warehouses. Let’s suppose for the moment that this represents the transit time too. Now consider Figure 2. The area on both sides of the dividing line between two warehouses is “almost” as close to one as it is to another.

Moreover, because carriers typically make delivers once a day the shaded areas probably represents the same actual transit time, in days, from one warehouse as the other. Consequently, there’s a large area that can be served from multiple warehouses in the same amount of time. In this example Chicago, Indianapolis, Cincinnati and Atlanta can be served from Memphis or Scranton; Dallas from Houston or Memphis and Denver from Houston, Memphis or LA. Lots of flexibility.

What about cost-based territories? The story there is similar but slightly more complex, at least for ground based transportation – TL, LTL and Parcels. These tariffs are a little more complex than distance relationships. Consequently, the regions of “approximately equal cost” are typically larger than shown in Figure 2. But the idea is the same – there’s lots of flexibility.

So far the idea relative to our electrical products manufacturer is that the notion of a fixed warehouse assignment is not a good one. There are probably many options of sourcing customers’ orders from multiple warehouses that provide the same cost and transit time. Therefore, the smart thing to do is to check for product availability in these more distant warehouses.

Think of the additional orders that would be processed in this way. Think of the decrease in back orders, the increased cash flow, and the improved customer satisfaction!

What about other situations? Let’s look at the case where the customer is clearly closer to a given warehouse, where the transportation cost is a slam-dunk.

If all the items ordered are available from the “best” warehouse, it’s clear that it should be processed from that warehouse. But there’s lots of times when this isn’t the case. Think of a company that has 85% service (line file rates). If the average number of lines ordered is just four, then fully half the orders will not be complete – they’ll create back orders.

Suppose then that an order can’t be shipped complete from the assigned warehouse. Suppose also that it could be shipped complete from some other warehouse, further away.

This is illustrated in Figure 3. Here we have to compare two (actually, at least two) shipments from the assigned warehouse that couldn’t ship complete to one shipment from the next warehouse.

A little knowledge of the ground LTL tariffs helps. Generally speaking, ground tariffs are quite dependent on weight and only slightly dependent on distance. Therefore it’s likely that the one shipment from Houston is cheaper that the two from LA and the one shipment from Scranton to Chicago is cheaper than the two from Memphis.

Think of it this way…as if you were the customer. Would you like a 700 pound shipment today and a 500 pound shipment sometime (!) in the future both from LA, or would you like one 1,200 pound shipment today from Houston?

A few simple steps serve to summarize the above points:
Rationalize territories (customer/warehouse assignments) based on cost, distance, lead-time, and so on.
Don’t stick to rigid fixed territories. Consider shipping from other warehouse when products are available there.
Pick off the “slam dunks” the cases where it’s both less expensive and faster to ship orders from an unassigned warehouse than from the assigned.
When it’s not a slam dunk (higher cost and better service or cheaper but worse service), provide the capability to make trade-off decisions in real time, preferably with the customers input and wishes.
Happy sourcing.