On receipt of a customer order the warehouse must perform checks such as verifying that inventory is available to ship. Then the warehouse must produce pick lists to guide the order-picking. Finally, it must produce any necessary shipping documentation and schedule the order-picking and shipping. These activities are typically accomplished by a warehouse management system, a large software system that co¨ordinates the activities of the warehouse. This is all part of the support to expedite the sending of the product to the customer.

    Order-picking typically accounts for about 55% of warehouse operating costs; and order-picking itself may be further broken like this [21]:

Activity                                         |  % Order-picking time
Traveling                                      |  55%
Searching                                     |  15%
Extracting                                     |  10%
Paperwork and other activities    |  20%

    Notice that traveling comprises the greatest part of the expense of order-picking, which is itself the most expensive part of warehouse operating expenses. Much of the design of the order-picking process is directed to reducing this unproductive time.

    The outbound processes of the warehouse are initiated by receipt of a customer order, which may be thought of as a shopping list. Each entry on the list is referred to as an order-line and typically consists of the item and quantity requested. The warehouse management system (WMS) then checks the order against available inventory and identifies any shortages. In addition, the WMS may re¨organize the list to match the layout and operations of the warehouse for greater efficiency. For example, if a customer has ordered 15 of a particular item, the warehouse management system (WMS) may check to see how the item is packaged. If 12 of the item comprise a carton, the WMS may convert the order-line for 15 eaches to two pick-lines, one for 1 carton and the other for 3 eaches. In many warehouses, each-picking and carton-picking are separate processes, and the pick-lines are diverted appropriately.

    Pick-lines are instructions to the order-pickers, telling them where and what to pick and in what quantity and units of measure. Each pick-line (or, more briefly, pick or line) represents a location to be visited, and since travel is the largest labor cost in a typical warehouse, the number of pick-lines is an indication of the labor required.

    Note that a pick (line) may require more than one grab if, for example, several items of a sku are to be retrieved for an order. Generally, this represents a much smaller proportion of the labor, because it is controllable by appropriate packaging (for example, pick one carton rather than 12 eaches).

    The WMS organizes pick-lines into pick-lists to achieve still more efficiencies, so that an order-picker may be able to concentrate on one area of the warehouse and so reduce travel. In addition, the WMS may sequence the pick-lines so that the locations to be visited appear in the sequence in which they will normally be encountered as the picker moves through the warehouse. (This will be explored in more detail in Chapter 10).

    The pick-list may be a physical sheet of paper, or merely a sequence of requests communicated by a stream of printed shipping labels, or by light, RF, or voice transmission.

    The most labor-intensive order-picking is the picking of less-than-carton quantities, referred to typically as broken-case or split-case picking. Broken-case picking is labor-intensive because it requires handling the smallest units of measure in the warehouse and this is generally resistant to automation because of the size and variety of skus to be handled. In contrast, carton-picking (picking full cartons) can sometimes be automated because of the relative uniformity of cartons, which are almost always rectangular and packed to resist damage.

    The pick face is that 2-dimensional surface, the front of storage, from which skus are extracted. This is how the skus are presented to the order picker. In general, the more different skus presented per area of the pick face, the less travel required per pick. An informal measure of this is sku density, which counts the number of skus available per unit of area on the pick-face. If a warehouse has a suitably high sku density then it will likely achieve a high value of pick density, or number of picks achieved per unit of area on the pick face, and so require less travel per pick.

    Sometimes it is useful to interpret the informal measures sku density and pick density as measuring skus or picks per unit of distance along the aisle traveled by an order-picker. One can then talk about, for example, the pick density of an order. An order that is of high pick density does not require much travel per pick and so is expected to be relatively economical to retrieve: we are paying only for the actual cost of retrieval and not for travel. On the other hand, small orders that require visits to widely dispersed locations may be expensive to retrieve because there is more travel per pick.

    Pick density depends on the orders and so we cannot know it precisely in advance of order receipt. However, it is generally true that pick density can be improved by ensuring high sku density, which is number of skus per foot of travel.

    Pick density can be increased, at least locally, by storing the most popular skus together. Then order-pickers can make more picks in a small area, which means less walking.

    Another way to increase the pick density is to batch orders; that is, have each worker retrieve many orders in one trip. However, this requires that the items be sorted into orders either while picking or else downstream. In the first case, the pickers are slowed down because they must carry a container for each order and they must sort the items as they pick, which is time-consuming and can lead to errors. If the items are sorted downstream, space and labor must be devoted to this additional process. In both cases even more work and space may be required if, in addition, the orders themselves must be sorted to arrive at the trailer in reverse sequence of delivery.

    It is generally economic to batch single-line orders. These orders are easy to manage since there is no need to sort while picking and they can frequently be picked directly into a shipping container.

    Very large orders can offer similar economies, at least if the skus are small enough so that a single picker can accumulate everything requested. A single worker can pick that order with little walking per pick and with no sortation.

    The challenge is to economically pick the orders of intermediate size; that is, more than two pick-lines but too few to sufficiently amortize the cost of walking. Roughly speaking, it is better to batch orders when the costs of work to separate the orders and the costs of additional space are less than the extra walking incurred if orders are not batched. It is almost always better to batch single-line orders because no sortation is required. Very large orders do not need to be batched because they will have sufficient pick density on their own. The problem then is with orders of medium-size.

    To sustain order-picking product must also be replenished. Restockers move skus in larger units of measure (cartons, pallets) and so a few restockers can keep many pickers supplied. A rule of thumb is one restocker to every five pickers; but this will depend on the particular patterns of flow.

    A restock is more expensive than a pick because the restocker must generally retrieve product from bulk storage and then prepare each pallet or case for picking. For example, he may remove shrink-wrap from a pallet so individual cases can be retrieved; or he may cut individual cases open so individual pieces can be retrieved.