Warehouse Optimization Strategies: Order Picking

Warehouse costs are dominated by labor. Picking is estimated to account for 70% of warehouse activities. A picker spends nearly half of his or her time looking for an item and the other half walking around the warehouse. Reduced travel time can save a warehouse up to 5% of its operating costs.

Christophe Van Huele
4 May 2022

Labor costs account for the majority of warehouse costs. Picking activities can account for roughly 70% of all warehouse activities when these costs are looked at. Almost half of the time the picker is working is spent looking for and grabbing the right item, with the other half spent walking around the warehouse and being unproductive. This can be seen as the lowest hanging fruit for a warehouse manager to solve the challenge of improving warehousing operations. If travel time could be reduced, the impact on the bottom line of a warehouse could be as much as 5% of total costs.

Warehouse Operational Costs

The Path to Warehouse Optimization

Previously, keeping a clean, well-organized facility and scheduling a large enough number of pickers was sufficient. However, managing and optimizing a warehouse in the 21st century is an entirely different ballgame.

Warehouses are now staying competitive by optimizing every aspect of their operations, from picking to packing to shipping. With new warehouses opening on a daily basis and fluctuating customer demand for quick deliveries, order fulfillment centers don't have the option of phoning it in. They must thoroughly examine their warehouses from top to bottom in order to identify and eliminate inefficiencies.

The travel time of pickers is one warehouse process that is ripe for warehouse optimization - picking path optimization. Picking, according to some estimates, consumes more than half of a warehouse's labor efforts. This isn't surprising given that, despite advances in warehouse technology such as automated storage and retrieval systems, picking is still largely a human-led process.

Nonetheless, picking and, more specifically, the pick path, can be improved. Motion waste – the unnecessary movement that causes a given task to take longer than it should is frequently the most significant cause of inefficiency during the picking process.

Warehousing processes

Warehouses are part of a logistics supply chain, and their main activities include receiving goods, storing and putting away items, picking, order consolidation and checking, packing, and shipping:

  1. The first operation in the warehouse is receiving. This process begins with notification of the arrival of goods. Following that, unloading staff begins the process of unloading, counting, identifying, quality control, and goods acceptance related to a type and quantity in accordance with company rules. The receipt is issued once the goods have been accepted. Acceptance is determined by the delivery status, the delivery date, the quality of delivery, and the planned schedule, which should minimize truck waiting time.
  2. Storing operations include distributing goods to storage areas, identifying, assigning storage, and put-away, which is a simple determination of a storage bin based on physical dimensions and weight of goods.
  3. Put-away is a process that necessitates a specific storage location. This is critical because the information system must always know what storage locations are available, where a specific type of good is located, and where each specific pallet is stored. This data is also used to create an efficient pick-list design. This process necessitates approximately 15% [6] of the operating costs because it involves numerous transfers from the gate to the storage location.
  4. Picking refers to the process of warehouse pickers receiving picklists and then traveling, searching for, and extracting items on pick-runs that begin and end at a depot. Picking can be classified into two types: homogeneous and het-erogeneous. Homogeneous picking is quite simple; the picker simply works with a full pallet. In heterogeneous picking, the picker is told where and what to pick, as well as how much and in what units to pick. Because of customer requirements, heterogeneous picking is logically more common. The disadvantage of heterogeneous picking is that smaller units cost more.
  5. Order consolidation is the process of completing a customer's order if it was picked by more than one picker.
  6. Order checking is the process of determining whether or not an order is complete and accurate.
  7. Packing ensures that the picked and consolidated goods, as well as the order's completeness, are packed for transportation and delivered to the shipping department.
  8. Shipping ensures that the packed consignment is delivered to the transport destination, assigned to the truck, and optimally loaded onto the truck. The shipping department, which can also secure three preceding jobs, ensures the shipping process.

An overview of warehouse optimization methods

The basic technical structure includes, for example, the layout design of the logistic warehouse or entire distribution center, the voice and dimensioning of conveyors and warehouse equipment, the design of physical interfaces to other systems or attributes related to the warehouse's technical structure.

Warehouse Layout Optimization

The warehouse layout design is an important component of subsequent optimization tasks and has a significant impact on order-picking and travel distances in the warehouse. The layout is typically rectangular in shape and relies on pallet manipulation. The number of blocks; the length, width, and number of picking aisles; the number and shape of cross aisles if present; the number of rack levels; and the position of input and output gates in the warehouse are all factors to consider in the layout design.

Storage Strategies

The two most common slotting strategies are random and dedicated. While a random strategy allows for the same probability of storing a pallet at any arbitrary empty location, a dedicated strategy allows for the storage of a pallet at a dedicated predetermined location.

Order Picking Strategies

If the order is small and does not exceed the picking capacity, more orders can be picked in a single order-picking tour. In the literature, this is referred to as order batching or simply batching.

Goods from the same product group are kept in the same zone. When compared to batching, zoning has no discernible effect on order picking system performance. The benefit of zoning is that it reduces aisle congestion, and when the goods are concentrated in a small area, it also reduces travel time. The main disadvantage is order consolidation when multiple pickers from different zones complete it.

Heuristics for warehouse routing

A routing strategy is a method for determining the best path through a warehouse. A route is a path that passes through all of the items in an order. If you want to keep order picking costs as low as possible, the route should be as short as possible.

  • The Return heuristic combines the Largest Gap and the S-Shape heuristic, which means that an aisle is either completely traversed or entered and exited from the same side of the aisle. The best option (= shortest route) from these two options is selected, and the next aisle is entered. These steps are repeated until the last item is chosen and the best route from the two options is determined.
  • The S-Shape heuristic requires that every aisle containing the pick item be traversed entirely in the shape of a S. In comparison to the other heuristics, the S-Shape heuristic (also known as the Traversal strategy) will result in shorter routes if the number of items per aisle is high. This heuristic is simple to apply; if an aisle contains a pick item, it is traversed completely; otherwise, it is skipped. This is most likely why this heuristic is so widely used in practice.
  • The Largest Gap heuristic states that a picker should enter an aisle as far as the largest gap. A gap is defined as the distance between two adjacent picks, or between the first and front aisles, or between the last and back aisles. The Largest Gap heuristic, unlike the S-Shape heuristic, will result in a short travel time if the number of items per aisle is low. The most significant gap is the section of the aisle that is not visited by the order picker.
  • Regardless of the layout or location of the items, the Optimal Method can calculate the shortest route. Optimal routes are typically a combination of S-Shape and Largest Gap. Optimal methods are typically algorithms that can solve TSP (traveling salesman problems) with a large number of locations.
Warehouse Routing Heuristics

The case of warehouse congestion

Many factors can contribute to warehouse congestion. Typically, it is the result of increased business while the warehouse layout remains unchanged.

The following are the most common causes of traffic congestion:

  • Too much inventory: even the largest warehouse facilities face bandwidth and storage space constraints. Congestion can occur if the warehouse is overcrowded with inventory.
  • A disorganized layout: The layout of your warehouse can also cause congestion. Most of the time, this has an impact on the operational efficiency of your forklifts and equipment, but it can also cause delays in picking and packaging, as well as shipping products to customers.
  • Lapse in communication: issues with relaying updated information through a fulfillment warehouse can also cause a clog in operations. Whether it's customers updating their orders online or requesting a change of address, communication gaps never help to relieve congestion – in fact, they make it worse.

Congestion can also be caused by picking activities. If one picker has a picking assignment in the same (narrow) aisle as another, depending on the width of that aisle, a waiting time is incurred.

As the business expands, more pickers will perform more activities in the same aisle at the same time, resulting in increased congestion.

What if pickers could use technology to automatically guide them through congested areas? What if warehouse pickers had something similar to Waze? Waze is widely regarded as the gold standard in real-time traffic route navigation. To implement real-time navigation in a warehouse, three things are required:

  • A positioning system that can track pickers down in near real time and with high accuracy (50cm). RTLS and Bluetooth LE are two examples.
  • A guidance system that guides and leads pickers through the aisles while also reducing search time when guiding to the correct location/shelf in the rack.
  • A dynamic routing system that recommends the best picking sequence across all picklists and dynamically reroutes, and reduces congestion due to lookahead planning.

Currently, a Warehouse Management System (WMS) is similar to a stand-alone routing engine that knows the shortest route from A to B without knowing the current situation. This route is based on a fixed location sequence and will not work if the warehouse layout changes.

A dynamic routing system, on the other hand, that is integrated with a positioning system (e.g., Bluetooth) and a guidance system (AR goggles) could function as a Waze-like solution that calculates the fastest route from A to B while taking current and forecasted traffic into account.

An algorithmic approach to reducing travel time

The goal, regardless of warehouse layout, storage strategies, or order picking strategies, will always be to reduce operational costs. The workforce is the largest cost driver, and traveling in a warehouse is the most time consuming activity.

The case of picklist optimization in a warehouse where picks are distributed amongst a team of pickers and the goal is to create picklists with the least amount of travel time while taking capacity, batching, and sequence constraints into account. The warehouse layout should be converted to a weighted graph in order for the algorithm to be aware of different positions in the warehouse and especially their relationships to other positions. A graph can represent the network structure between any position in the warehouse, with nodes representing positions and edges representing distances between two positions. By introducing a graph, one can use shortest path algorithms (such as Dijkstra, A*,...) to calculate the distance matrix from one position to another. This information is required in order to solve routing requests for a set of picks and a set of pickers.

Warehouse Layout to Graph


Picking and labor optimization give online grocery a competitive edge


How Does a Manufacturer Achieve Optimal Warehouse Demand Planning?

Get more insights