Warehouses may be categorized by type, which is primarily defined by the customers they serve. Here are some of the more important distinctions:

A retail distribution center
typically supplies product to retail stores, such as Wal-Mart or Target. The immediate customer of the distribution center is a retail store, which is likely to be a regular or even captive customer, receiving shipments on regularly scheduled days. A typical order might comprise hundreds or thousands of items; and because the distribution center might serve hundreds of stores, the flow of product is huge. The suite of products changes with customer tastes and marketing plans; but because the orders are typically known a day or more in advance, it is possible to plan ahead. Some product may be pushed from the distribution center to the stores, especially in support of marketing campaigns.

A service parts distribution center
is among the most challenging of facilities to manage. They hold spare parts for expensive capital equipment, such as automobiles, airplanes, computer systems, or medical equipment. Consequently, one facility may represent a huge investment in inventory: tens or even hundreds of thousands of parts, some very expensive. (A typical automobile contains almost 10,000 parts.) Because of the large number of parts, total activity in the DC may be statistically predictable, but the demand for any particular part is relatively small and therefore hard to predict. This means that the variance of demand can be large and so relatively large quantities of safety stock must be held, especially since there can be usually long lead times to replenish parts to the warehouse. Indeed, sometimes there is as much safety stock as cycle stock, and so, in aggregate, these skus require much space. This in turn increases travel distances and makes order-picking less efficient.

A typical service parts warehouse manages two distinct order streams: stock orders, by which dealers replenish their shelves; and emergency orders, in which an equipment owner or independent repair shop urgently requires a few special parts to repair a broken piece of capital equipment. Stock orders tend to be large and fairly predictable replenishments of popular consumables, while emergency orders are typically small (two to three pick-lines), unpredictable, and urgent, because expensive capital equipment is likely waiting for repair. Emergency orders are typically for items that are ordered infrequently (otherwise they could have been provided by the dealer from stock inventory). Such orders—a few, slow-moving items that must be picked immediately—are relatively expensive to handle. Worse, customers ordering for repair might order before they are absolutely sure which parts need replacement; and so there can be a significant percentage of returns to be handled at the warehouse.

For most product in a service parts warehouse there are not sufficiently reliable patterns of movement to justify special processes, but one can layout stock to be more space efficient by storing similar sizes together, thereby reducing travel. Furthermore, one can hedge chances of having to travel long distances. For example, it can be advantageous, especially for emergency orders, to store products together that are likely to be ordered together.

Another complication is that the life cycle of a service part is unusual, with three stages of product life, as shown in the next Figure.

Early failures are generally due to manufacturing imperfections; mid-life failures are generally due to random events that stress the part beyond its tolerance; and end-of-life failures are due to expected wearing out of the product. Demand for product generally reflects this pattern, and creates challenges in the warehouse. For example, there is little time to ramp up availability of new product at the start of its life cycle. Also, parts are more frequently requested at the end of the product life cycle, and so it is easy for the warehouse to be stuck with obsolete merchandise. Finally, it may be necessary to relocate product as its popularity changes.

A catalog fulfillment or e-commerce distribution center
typically receives small orders from individuals by phone, fax, or the Internet. Orders are typically small, for only 1–3 items, but there may be many such orders, and they are to be filled and shipped immediately after receipt. Because customer orders require instant response, such distributors typically try to shape demand by offering special prices for ordering at certain times or in certain quantities or for accepting more variable delivery dates.

A 3PL warehouse
is one to which a company might outsource its warehousing operations. The 3PL provider might service multiple customers from one facility, thereby gaining economies of scale or complementary seasons that the customers would be unable to achieve on their own. 3PL facilities may also be contracted as overflow facilities to handle surges in product flow.

A perishables warehouse
may handle food, fresh flowers, vaccines, or other product requiring refrigeration to protect its very short shelf life. They are typically one link in an extended cold chain, along which perishable product is rushed to the consumer. Such DCs are distinctive in that product dwells within for very short times, frequently only hours. Also, there is a great emphasis on using space effectively because, with refrigeration, it is so expensive. They face many challenges in inventory management, including requirements to ship product according to FIFO (First-In-First-Out) or FEFO (First-Expired-First-Out). Also, there are typically many restrictions on how product is handled. For example, chicken cannot be stacked on top of anything else, to protect against juices dripping onto product below and contaminating it. Finally, appropriate temperatures must be maintained and this can be different for different kind of products. A typical food DC operates separate areas for ambient temperatures, chilled (around 2 degrees C, 35 degrees F), and frozen product (-18 degrees C, around 0 degrees F). To protect stored product, it is important to avoid bringing in anything warmer. This type of warehouse will become more common as China, India, Brazil, and other rapidly industrializing countries build a middle class, which will increasingly want fresh fruit, vegetables, meat, and dairy.

While there are many types of warehouses in the supply chain, one of the main reminders is that there is a systematic way to think about a warehouse system regardless of the industry in which it operates.

What we need to realize is that the selection of equipment and the organization of material flow are largely determined by :

• Inventory characteristics, such as the number of products, their sizes, and turn rates;
• Throughput and service requirements, including the number of lines and orders shipped per day;
• The footprint of the building and capital cost of equipment;
• The cost of labor.

We also see that there is a regional difference in how to organize warehouse operations. Here is a brief survey:

North America

North America is driven by mass consumption. Think Wal-Mart. This enables huge economies of scale and, indeed, the trend has been for ever larger distribution centers and ever accelerating rates of product flow. As telecommunications enables better coordination along the supply chain, the uniformity of market and of distribution infrastructure allows fewer, more centralized and therefore larger distribution centers.

The Amazon.com distribution center shown in the following figure is typical:

One level, with conveyors and sortation equipment but little other significant automation. Such warehouses are generally built in the countryside surrounding major metropolitan areas, so that land is cheap but there is still ready access to large markets.

The fairly high costs of labor are held down somewhat by constant immigration into the US and Canada.

Warehouses in North America are coordinated by increasingly sophisticated warehouse management systems and so very rich data sets are available with which to evaluate and refine performance.

East Asia

Business in Asia has traditionally been based on personal relationships and less on computational models. Because of this tradition, data is not robust and not widely available; consequently the opportunities to improve operations by science are not fully developed at present.

In general, the most active economic areas are separated by lots of water, which means lots of product conveyed by air (for high-value or time-sensitive products) or ship (for bulky items or commodities). For both air and sea cargo, the large fixed costs increase incentives to consolidate freight. Consequently one expects to see the emergence of strong regional hubs, such as Singapore and Hong Kong, to support this consolidation.

India

India, like many developing countries, has both inexpensive land and low labor costs. Capital costs are relatively high in relation to the cost of labor and so there is less pressure to install specialized storage, even pallet rack. And because labor costs are low there is less incentive to increase efficiency. For example, it is not an attractive proposition to reduce labor costs by picking from flow rack: The labor savings cannot justify the cost of the rack or the forklift trucks.

In addition, warehouses in India distribute mainly to the local economy and so supply a market that is not wealthy. Consequently, the SKU’s are not likely to be high cost items and so there is not much savings available from reducing inventories by precise timing. Consequently information technologies cannot generate much savings.

Finally, inefficiencies in transport make India in effect a collection of smaller markets. These inefficiencies include the physical, such as roads in less-than-ideal condition, as well as the administrative. For example, each state within India levies customs duties on freight transported across the border. This slows interstate commerce and increases the expense. Such factors increase the costs of transportation and so favor a strategy of having more, smaller distribution centers rather than fewer, larger ones, where the volume of activity could better justify capital investment. The national government is attempting to revise its tax structure fix these inefficiencies.

Figure : The relatively low cost of labor, high cost of capital, and artificially small market mean that this warehouse in India may be economically efficient. (Photo courtesy of Rohan Reddy)

India is increasingly becoming a global sourcing hub and so suitable distribution centers are being built in around large ports, such as Mumbai (Bombay). However, land can be expensive there. Caleb Tan of Menlo Worldwide Logistics has observed prices comparable to those of Singapore or Hong Kong. Apparently this is due to a lack of land because of encroachment of slums as more and more people migrate from the countryside to economically vibrant areas.

China

A distinctive feature of the logistics scene in China is the seemingly boundless supply of very low cost of labor together with relatively inexpensive land. Consequently warehouses tend to be large, low buildings as in North America; but with some striking differences. For example, it is not unusual as of this writing to find a warehouse of 250,000 square feet with a single fork lift truck. The reason is that equipment is expensive but labor is cheap.

Figure : In the US warehouse on the left, cartons of beer have been palletized because labor is expensive compared to capital. The reduction in labor is worth the expense of a forklift plus the additional storage space. In the Chinese warehouse on the right, cartons have been stacked by hand and must be unstacked by hand; but labor is cheap and capital is expensive.

Despite cheap labor, China does have some capital-intensive warehouses, with the latest information technology and storage equipment. Such warehouses are most likely devoted to the distribution of high-value goods for export. Because such goods, such as consumer electronics, have high-value and short life-cycle, the warehouses can justify their equipment by substantial reductions in inventory costs.

The very different costs in the US and China sometimes leads to behavior that makes sense locally but may make the supply chain inefficient. For example, The Home Depot receives some Chinese-built product at its Import Distribution Center in Savannah, Georgia, USA. The shipping department in the warehouse in China de-palletizes freight in order to pack each trailer as tightly as possible for the drive to the sea port. Thus an expenditure of relatively cheap labor will reduce the relatively significant costs of equipment and transportation. But this means the product arrives in the US as loose cartons in containers and so The Home Depot must re-palletize the cartons before storage in deep, drive-in pallet rack. And most of it will, shortly after, be de-palletized once more when it is picked as cartons for shipment to stores.

Singapore, Hong Kong, Japan

Some economic powers such as Singapore, Japan, and Hong Kong suffer from limited space so land is much more expensive than elsewhere. Consequently, many of the warehouses are high-rise, such as shown here:

Figure :Multi-story warehouses are common in Singapore, Hong Kong, and Japan where land is expensive.

In addition, as first-world economies, labor in these places is expensive and so warehouses here are more likely to be automated. Freight elevators are likely to be bottlenecks to material flow in these facilities.

Space constraints have led to an interesting type of warehouse in Hong Kong and in Singapore: A multi-floor facility with no automation or elevators, but, instead, a spiral truck ramp so that trailers may be docked at any floor

Figure : Multi-story warehouse with no automation but accessibility provided by a spiral truck ramp.

In effect, each floor becomes a ground floor—but the cost is that significant land area, determined in part by the turning radius of a truck, is lost to the ramp and unavailable for storage. That storage space must be reclaimed up above.

This design is a clever and efficient way of using space if each floor is occupied by an independent tenant. But one must be careful if some tenant occupies multiple floors for then it may become necessary to shuttle trucks among floors. For example, a multi-story warehouse with spiral ramp would be unsuitable as a local distribution center: Trailers departing with small shipments for many customers must be loaded in reverse sequence of delivery to avoid double-handling. But since the load may not match the layout of product amongst the floors, this could require much shuttling of the trailer among floors to load.

Similarly, this might be an inefficient warehouse in which to receive shipments of diverse items that may be stored on different floors: Because rear-entry trailers restrict the sequence in which freight can be accessed, a trailer may have to shuttle among the floors to deliver all the freight to the appropriate places. Alternatively, the trailer would have to be loaded to match the allocation of product among floors at the warehouse. In either case, extra work is required.

Central and South America

This is a region of developing markets that are separated by geography such as the Amazonian rain forest and the Andes mountains. Consequently, markets tend to be relatively small and inappropriate for significant capital investment in warehouses.

Data may not be available or may not be transmitted to supply chain partners. In part this is because there has not been a strong, reliable stock market and so wealth has generally been invested in family businesses, which are less inclined to share data.

Labor costs are relatively low, but some segments of the workforce enjoy strong political protections, which results in labor inflexibility. Consequently, while labor can be cheap, it can be hard to shrink or re-deploy a workforce. Most of labor is devoted to handling small quantities, because retail stores are small and customers purchase tiny amounts daily.

Space is the main concern. Infrastructure is not well-developed and so, to reduce travel on bad roads, warehouses must locate close to the customer. But customers are concentrated in a few, very large cities such as Mexico City, Sao Paolo, and Lima, which are congested and where space is expensive.

Most warehouses in this region hold goods for domestic distribution. Because markets are not highly developed, inventory tends to be of less expensive goods and so the costs of moving inventory slowly are relatively low. (In contrast, China has many distribution centers supporting export of high-value goods, such as consumer electronics. The high cost of holding such inventory, justifies investment in facilities that enable rapid movement of product. This also holds for Mexico, which can get product quickly to market in the US and so able to justify advanced distribution centers.)

The following photograph reflects the low labor costs relative to the costs of capital in this region.

Figure: A ladder is much cheaper than a person-aboard truck, though much slower. (Note the product stored as loose cartons.)

Europe

Warehouses in Europe, especially in Germany and France, are shaped by the relatively high labor costs and inflexibility of the work force. These facts push designers to find engineering solutions rather than social solutions to logistics challenges. For example, there is a greater inclination to use automation than in comparable facilities in North America.

In the past, the economies of Europe were separate. More recently the economies are integrating into a common market, which will create economies of scale, which will likely lead to larger warehouses. However, urban areas, many of which have grown out of ancient towns, will still present challenges to the efficient flow of product. All this is reflected in Figure,

Figure: A highly automated distribution center in Germany. (Photo courtesy of Kai Wittek)

which shows a distribution center of a major drugstore chain in Germany. The multi-story portion of the building, visible in the background, houses a high-rise automated storage-and-retrieval system. This is not so much to conserve space as to reduce labor costs. (Unlike the Singapore warehouses, this facility is tall only where the AS/RS is installed.)

According to the Rossman web page, the AS/RS is 30 meters high, the aisles are 127 meters long and the facility provides 14,000 pallet positions. The delivery trucks in the foreground are relatively small, at least compared to the 48-foot trailers common in North America. The small trucks are necessary to deliver to stores in the centers of cities, the ground plans of which may have been laid out centuries ago and cannot accommodate large trucks. But the warehouse is large, reflecting the extent of the market it serves.

 

References : Copyright 1998–2011 John J. BARTHOLDI, III and Steven T. HACKMAN. All rights reserved. This material may be freely copied for educational purposes—but not for resale—as long as the authors’ names and the copyright notice appear clearly on the copies. Corresponding author: john.bartholdi@gatech.edu