PICKING STRATEGIES IN STORAGE

1. WHAT IS THE PICKING ACTIVITY AND WHY IS IT IMPORTANT?

In order to understand the picking activity (separation and preparation of orders) it is important to present its insertion among the main storage activities. In a simple way, all types of warehouses have the following functions (Figure 1):

• receipt of products
• Storage of products until needed
• collection of products according to customer orders
• preparation of products for delivery to the customer.

The picking activity can be defined as the activity responsible for collecting the correct mix of products, in their correct quantities from the storage area to satisfy the needs of the consumer. In this way, we will be focusing on the order picking activity, known as order picking, or simply picking, as we will refer to it from now on.

Such activity inside a warehouse is considered as one of the most critical. Depending on the type of warehouse, 30% to 40% of the labor cost is associated with the picking activity. Allied to the cost, the time of this activity substantially influences the order cycle time, that is, the time between the receipt of a customer order and the correct delivery of the products.

The progressive increase in the needs (and demands) of consumers and competition has brought about several consequences for the storage activity. Such consequences can be translated into general trends that can be observed in different sectors:

• Profiling the number of SKUs: Increased customer requirements have increased the number of products companies are currently working with
• Increased Number of Orders: customers began to work more and more on continuous resupply philosophies, with the objective of reducing their stock levels. Smaller batch quantities imply an increase in the number of orders over time
• Concentration in Large Warehouses: the paradigm of local presence is starting to disappear. Companies are starting to adopt an operation with a smaller number of warehouses and points of sale, concentrating inventories and obtaining cost reductions with cargo consolidation.
• Next Day Delivery: With an ever-increasing demand for shorter lead times for customers.

In addition to these trends, companies realized the importance of using the service as a value-added differential in their products. Product quality becomes a prerequisite and services such as home delivery and internet sales have increased the level of demand and productivity of storage and transport activities.

In this way, the picking activity must be flexible to ensure an operation within the needs determined by the customer, using control and monitoring systems that support the levels of service and quality diagnosed.

2. WHAT PRINCIPLES SHOULD I USE TO IMPROVE THE PICKING SYSTEM?

Regardless of the size, volume, stock types, customer needs and types of control systems in the warehouse operation, there are certain principles that apply well to any picking activity.

These are principles that should guide the positioning of products within the storage area and the flow of information and documents.

• Prioritize high-volume products

The first step is to identify the products with the highest turnover. In most cases it will be observed that a small group of products corresponds to a large part of the movement in a warehouse. It is the well-known Pareto Law, where 20% of the products correspond to 80% of the movements. High volume products are considered. About 55% of the products correspond to 95% of the volume handled. These 35% of products are considered medium volume. The remaining 45% of products are considered low-volume products, corresponding to approximately 5% of the total volume handled.

The products with the highest turnover must be in positions that are easier for operators to access and easier to resupply. This idea strongly guides the physical arrangement of products in the warehouse. We can present a basic layout developed according to product turnover (Figure 2).

 

The objective is to prioritize the minimization of the distance between the operator who performs the collection and the products to be collected. Thus, the products with the highest turnover must be placed in the region closest to the separation activity (1). The mats (7) eliminate movement when receiving the list of products and sending them to the dispatch. Interesting to reserve an area (2) for the storage and collection of small and high volume products. An area for receiving (3) products that will feed the regions (1) and (2) must be planned. Similarly, a dispatch area must be dimensioned (4) with enough lines to avoid accumulation or queues in the picking line. The mats that take complete orders from the picking area to the shipping area (5) must be high in height to take advantage of floor space. In region (6) we have all products of small and medium volumes, stored on pallets. This is a very generic example, but the philosophy is applicable in many cases.

• Use of clear and easy-to-operate documentation

A picking document must provide specific instructions for the operator in order to facilitate the activity of picking products. It should only contain the relevant information: product location, description and required quantity. In addition, such information must be highlighted in the document, in order to facilitate reading. A greater concern with the preparation of this documentation reduces the time spent reading and looking for products by the operator, reducing activity time.

• Organize orders according to physical configurations

It is necessary that each order sent to the picking area be configured according to the product location restrictions. That is, in the picking document generation stage, the listings must be “assembled” in order to reduce the operator's movement, in addition to observing the proximity of products.

• Maintain an efficient product location system

An efficient order picking system needs a very accurate product location system. With the standardization of addresses for locating products and using technologies that speed up the identification of a position, it is possible to reduce the time to search for a product to fractions of seconds, speeding up the order picking activity.

• The operator must be evaluated for errors

In order to avoid errors in picking orders (incorrect products or incorrect quantities of products) the operator must be evaluated for the correct picking of orders. Its performance must be measured and any deviation from an acceptable target must be analyzed, identifying whether the cause is in the system or in the operator.

• Avoid counting products during collection

Product counting substantially increases order picking time. Such activity can be avoided with simple solutions such as packaging solutions. For example, if an operator needs to separate 1000 units of a certain product, if such product were grouped in packs of 100 units, this would facilitate his work. Also, it would help to eliminate errors in separation.

• Elimination of paper documents

Any paper document takes too much time in the picking activity. Written information must be read, interpreted, and sometimes compared to some control system, which typically results in errors. There are technologies that are becoming increasingly accessible, reducing and even eliminating the flow of paper, including barcode scanners, voice recognition systems and radio frequency terminals.

Such philosophies guide the planning of the picking activity. However, several technologies and strategies can be used and planning a picking activity is not a simple task.

3. HOW DO I ORGANIZE MY PICKING ACTIVITY?

The order picking activity is intensive in materials movement. Depending on the type of warehouse, around 30% to 40% of its cost is directly associated with order picking.

This is because the number of workers and the time associated with this activity are greater than in other warehousing activities. Through studies of more typical order picking systems, where we have a large use of paper documents and product displacement and collection operations, we have information on how the operator's time is spent in this activity.

Through Graph 1, we observe that most of the time spent by operators is moving to collect and place products on the picking line. In other words, a fundamental objective for a good productivity of an order picking system is the minimization of displacement or movement times.

The way we organize the picking strategy is directly associated with handling time. By picking strategy we mean how we organize the order picking process, planning the number of operators per order, the number of different products picked up in each pick and the periods for scheduling or grouping orders during a shift.

Technological storage solutions such as Carrousels, mini-loads, AS/RS systems (Automatic Storage and Retrieval Systems) and WMS are combinations of equipment and control systems that move, store and collect products with high precision, accuracy and speed, depending on the degree of automation. Such systems also reduce the operator's movement time, as it aims to bring the specific products at the time of separation, in addition to reducing the counting time.

Equipment such as barcodes, optical readers, reduce search and documentation time.

We will go into detail only on picking organization strategies, as this is the initial step in planning a system, leaving alternative technological solutions for a future publication. Generally, during the planning of a picking system, considerations about picking strategies are little evaluated or neglected, starting directly with the adoption of technologies and solution packages. Depending on the type of company's process, this can lead to an investment in unnecessary equipment and systems, which could be solved only with a change in the operators' organizational strategies.

4. WHAT ARE THE STRATEGIES FOR ORGANIZING THE PICKING ACTIVITY?

There are 4 basic procedures for organizing picking. These 4 procedures are characterized as “pure” procedures. Generally, what is observed is a composition or mixture of different strategies, generating mixed picking organization strategies.

Basically, when defining which strategy to use, it is necessary to answer the following questions:

• Operators per order: How many operators should be assigned to complete just one order? Is each order worked on by just one operator, or will we have several operators working on the same order?
• Products per order: should the operator pick one product at a time from the order list, or pick multiple products in one pick?
• Scheduling Periods: How many order fulfillment windows should be done in a shift? Is it necessary to reconcile picking with other activities such as receiving products and shipping?

Below we describe the 4 basic picking activity procedures:

4.1. Discreet Picking

In this procedure, each operator is responsible for one order at a time and takes only one product at a time. There is only one scheduling window per shift. This type of organization has a number of advantages, mainly because it is the simplest, adapting perfectly when all documentation is on paper. The risk of errors in the activity is reduced, as there is only one document for each product picking order. However, it is the least productive procedure, as the operator must complete the entire picking order, the displacement time is much longer than in other procedures. There is only one period for scheduling the picking activity.

We can understand the different picking strategies through simple examples. Following the following figure, suppose that the order picking activity is working with only 4 products (P1, P2, P3 and P4). The picking line has 3 operators fully allocated to this activity. We then have 3 orders arriving, composed of a mix and quantity of different products.

In the discrete picking strategy, the first operator would pick the first order (Order 1). He would then be responsible for initiating and completing the picking of all products contained in that order. I would select 10 quantities of the first product, 20 of the second and 5 of the third, placing them in the box for the next operation. At the same time, the second operator would be responsible for the second order, collecting products 1, 3 and 4 in their respective quantities (one at a time). Similarly, the third operator would be responsible for the third order.

It is important to note that each order is initiated and completed by only one operator and that only one product is picked at a time. The first operator to finish his work, which in the example would be the third operator, would take the next order (Order 4, not shown).

4.2. Picking by zone

In this form of organization, storage areas are divided into zones. Each zone has certain products. Each operator of the picking activity is related to one of these zones.

When an order arrives, each operator takes all the product lines related to that order that are part of his work zone. If the order is complete, it can be shipped. Otherwise, it will go to the next picking zone and the next operator will place the necessary products.

This type of procedure is more used when there are differences in productivity between workers or differences in equipment/technology used in the picking area. With this, the picking zones are determined in such a way until we obtain a balance of the workload between the zones. There is only one period for scheduling the picking activity.

Going back to our example, each operator would be assigned to a certain zone. The first operator would be responsible for collecting products 3 and 4. Operator 2 for product 2, while the last operator would be responsible for product 1.

When the first order arrives on the picking line, operator 3 would collect 10 units of product 1. Then operator 2 would collect 20 units of product 2. Finally, the first operator would collect 5 units of product 3. The first order would then be complete and would be dispatched to the next activity. We note that in this case, the 3 operators worked to complete an order.

Furthermore, after having collected the 10 units of product 1, the third operator would already start working on the second order, while in parallel the other two operators would be completing order 1.

As previously mentioned, it is an ideal strategy when we have different technologies or when the productivity of operators is not homogeneous. In our example, operator 1 is the most productive, having 2 products in his picking area. With this, we have an increase in productivity compared to the previous strategy, but the operation is a little more complex.

4.3. Picking by Batch

In the previous procedure, different products are collected to complete one order at a time. In batch picking, the procedure is different: the operator waits for a certain number of orders to accumulate. Next, the products common to several orders are observed.

When the operator makes the collection, he takes the sum of the quantities of each product, necessary to fulfill all orders. It then distributes the collected quantities to each order.

By working with several orders per collection, this type of procedure has a productivity gain in relation to the others. However, it is indicated only when the products are mostly collected in fractional quantities (not in boxes), and when the orders have few different products (1 to 4) and small volumes. The productivity gain is due to the reduction of operators' transit time. A negative point of this procedure is its greater complexity and its need to use severe measurements to minimize the risk of errors. Such measurements can be made using current technological solutions. Again, we only have one period for scheduling the picking activity.

In our example, orders would be batched together. For example, orders 1 and 3 would be grouped together in one batch. The first operator would exclusively handle these two requests. He would then collect 10 units of product 1, 40 units of product 2 and 20 units of product 3, that is, the sum of the units of products in orders 1 and 3. The other two operators would be responsible for other batches of orders.

In batch picking, an order is processed by only one operator, and different products are picked at each pick. This speeds up productivity, but as we mentioned, it is only suitable for setups with few products.

4.4. Picking by Wave

This method is similar to discrete picking. That is, each operator is responsible for one type of product at a time. The difference is in scheduling a certain number of orders throughout the shift. Generally, this type of procedure is used to coordinate order picking and dispatch functions.

In addition to the strategies presented above, we have combinations between pure strategies.

The zone-batch picking strategy, for example, is the zone strategy, where each operator is responsible for a certain number of products, and where orders are grouped in batches.

In the following matrix, we can summarize the different picking activity strategies, those considered pure and those considered mixed.

 

A more recent and innovative picking organization strategy is called bucket brigades. Developed by professors at Georgia Tech, it differs from the previous ones by being a strategy that automatically adjusts the system

5. BUCKET BRIGADES – A NEW SELF-BALANCEABLE STRATEGY

The "Bucket Brigades" strategy has been used in production lines for its self-balancing functionality. Some companies that have used this strategy:

• McGraw-Hill: picking activity in its Distribution Centers
• The MusicLand Group: picking activity
• Time Warner Trade Publishing/Little, Brown: picking activity
• Bantam-Doubleday-Dell Distribution: picking activity
• Harcourt-Brace: picking activity
• Blockbuster Music: picking activity
• Subway: applied the strategy to assembling sandwiches
• Mitsubishi Consumer Electronics America: television assembly and cell phone packaging
• Revco Drug Stores, Inc.: Achieved a 34% productivity increase in picking activity with the implementation
• Readers Digest: achieved an 8% increase in picking productivity and a 35% reduction in picking and picking errors.

A major difficulty after choosing the picking strategy to be adopted is the need to balance the line, so that no operator or equipment is overloaded and so that the production line has its maximum capacity. These adjustments should be made periodically and using the most recent and accurate information available.

A great promise of this new strategy is the fact that it is self-balancing. That is, by increasing or decreasing the order rate, the system is organized in such a way that there is a self-adjustment, without increasing or decreasing the occupancy of the operators.

 

We can understand how this strategy works through Figure 6. Suppose the system has 3 operators. The operators work discreetly, that is, there are no zones or batches of orders. The operator at the end of the line (3) is more productive than the second operator (2) who is therefore more productive than the first operator (1). By more productive, we mean that the operator performs the same movement or operation in a shorter amount of time.

The strategy starts with operator 3 processing the first request, operator 2 the second request, and operator 1 processing the third request. There are other requests in the queue waiting to be processed, as shown in frame 1 in Figure 6.

Then operator 3 finishes completing a request (table 2, Figure 6). At that moment there is a reallocation of the work of each operator (charts 3 and 4, Figure 6). Operator 3 takes the order Operator 2 was working on. Operator 2 takes the order that operator 1 was working on and operator 1 takes a new order that was in the queue.

When operator 3 completes the collection of products for this order, the process restarts.

Despite appearing to be an apparently simple strategy, it requires rigorous coordination between the operators, a prior study of each one's productivity and preparation of orders according to the physical configuration of the racks. However, it is mathematically proven that this work organization strategy causes workers to gravitate towards the optimal division of labor, eliminating the activity of balancing and planning.

We can then point out the following as the main benefits of using the Bucket Brigades strategy:

• Reduction in the need for planning and administration, as it makes the line self-balancing
• Process becomes more agile and flexible by self-tuning
• Increase in processed units, in addition to the trend of optimal division of labor
• Reduced side work and increased quality by reducing work-in-process

6. HOW TO CHOOSE BETWEEN DIFFERENT STRATEGIES AND EQUIPMENT?

In addition to picking strategies, there are a large number of technological solutions that can be used in storage functions. The subject is so extensive that it could not be fully covered in this introductory text. In general terms, administrators must choose between different solutions for each of the dimensions below:

• Picking strategies (presented in this material)
• storage equipment
• handling equipment
• Control systems (WMS)

Choosing from a range of options within each dimension is not an easy task. Solutions often combine traditional systems with state-of-the-art automated systems. It is necessary that project teams be created with all the areas involved, since the picking activity is central to the order processing and storage activities. Together, the needs of the system to serve the consumers and the objectives must be defined. Next, the alternatives must be evaluated.

So how do you choose between different strategies and equipment? A very useful tool for this type of problem is the use of simulation to evaluate different alternatives for the planned picking system. The following figures show some simulation models of the picking activity.

After designing the alternatives to be adopted, computational models can be created taking into account different levels of detail, such as:

• order profile
• number of products
• work shifts
• number of operators
• number of resources (forklifts, stacker cranes, conveyors, etc.)
• picking strategies
• activity times
• etc.

A computational model is a statistically valid representation of a real system. In this way, we can vary several parameters and observe how the alternative behaves in terms of:

• capacity utilization of operators
• resource utilization
• order picking and picking time
• number of orders collected per day
• etc.

Through the computational model, it is possible to financially measure the cost of each alternative, observing whether the cost is compatible with the desired performance of the system.

One of the advantages of its use is to avoid the advance purchase of equipment or hiring of resources, since all alternatives can be modeled and tested "virtually".

 

 

 

7. CONCLUSION

We can see that the picking activity is a critical activity in storage. This is due to the need for manual work and intensive movement of materials and the growing demands from the market for the reduction of cycle time.

For the planning of a picking system, in addition to the equipment and systems involved in storage, it is necessary to define a strategy for the collection and separation of products in order to meet the requirements of productivity and flexibility of the line.

There are 4 basic strategies that can be combined. In addition, we observe one of the most recent and innovative strategies that eliminates the need to balance the resources involved in the activities.

Finally, we highlight the difficulty in choosing the best strategy and the best mix between storage and system solutions. We present the simulation as a useful tool in testing different alternative solutions for picking.

8. BIBLIOGRAPHY

• “Warehouse Distribution & Operations Handbook”, David E. Mulcahy, McGraw-Hill
• “Warehouse & Distribution Automation Handbook”, Adams, Firth, Brown & Misenheimer, McGraw-Hill
• Bucket Brigades: http://homer.isye.gatech.edu/people/faculty/John_Bartholdi/bucket-brigades.html