Labor is the single largest operating cost in most DCs, and travel often accounts for half of all labor time, especially in order picking. In some DCs, pickers can travel upwards of 12 miles per shift.
To help address the warehouse travel challenge, we have put together this guide that walks through the most proven and popular warehouse travel reduction strategies distribution leaders have taken to reduce travel and improve productivity in their order picking and replenishment processes.
Labor Shortages and the Productivity Challenge
Labor productivity and efficiency is critical to DCs that are struggling to hire and retain workers amid tight labor markets and rising wages. Adding to the challenge, demand for warehouse workers continues to grow as companies build new fulfillment centers and expand existing facilities to satisfy the growing volume of direct-to-consumer shipments.
According to U.S. government labor statistics, the growth in eCommerce sales is driving a 10 percent annual increase in demand for warehouse labor. By 2025 ecommerce is expected to expand by an additional $1.4 trillion and account for 50% of the growth in retail. At that level, the industry will need 20 percent more DC workers than it employs today
In addition, there is an emerging shortage of hourly workers across all industries. Improving productivity is imperative for meeting the ecommerce labor challenge. And reducing warehouse travel is the lynchpin to improving productivity.
Travel in Picking and Replenishment Processes is the Lynchpin
In a conventional, non-automated picking process, travel represents the majority of the time in a DC associate’s day. Warehouse workers following an RF-picking process will typically spend more time walking or driving between pick locations than they do pulling products from bins, slots or racking locations. As illustrated by the diagram (below), hands-free voice technology can significantly reduce time at the pick face, but it does not address the travel time between picks.
For years, DCs have devoted significant efforts to reducing warehouse travel through software solutions like slotting, conventional automation, and lean process initiatives. Those efforts are taking on new urgency as ecommerce growth boosts labor demand.
Tried and True Warehouse Travel Reduction Strategies
1) Process Engineering
The most common process-related solution to the travel challenge is to split order lines by zone (based on product velocity, type/size, etc.), and to optimize picking processes in the various pick zones. (Learn more about process optimizations in our “Ultimate Guide to Improving Warehouse Order Picking Productivity”)
Many DCs place their fastest moving piece pick items in pick modules with flow rack and conveyors. It’s not unusual for workers in a pick module using voice in a bucket brigade process to pick hundreds of lines per hour. Some Lucas customers hit 1000 picks per hour in high-density pick operations.
Slower-moving items may be batch or cluster picked so users can pick multiple orders to a cart or pallet on a single trip through a warehouse zone. Moving from single order picking to batching can dramatically reduce travel by increasing pick density (i.e., the number of picks per unit of travel).
Conventional material handling solutions often complement these process approaches to travel reduction.
2) Traditional Material Handling
Conveyors eliminate the long trips from picking areas to packing or shipping docks, and automated sortation systems improve the eciency and accuracy of assembling orders picked in multiple zones.
Automated guided vehicles (AGVs) provide an alternative to conveyors for transporting products over long distances. First introduced in the 1980s and 1990s, AGVs have evolved from track-guided systems to wire- and light-guided systems that follow set travel paths, eliminating travel from picking to staging areas.
As well will discuss later in this guide, autonomous mobile robots take this travel reduction strategy to a new level with robots that can operate anywhere in a DC.
Slotting software also goes hand-in-hand with process optimization efforts by determining which products should be placed in which warehouse locations in order to improve overall warehouse efficiency. The key objective of slotting programs is to optimize product positioning, but not solely to reduce travel in picking.
Slotting solutions are based on highly engineered models of a facility that take account of product flow and picking processes, product attributes (including physical characteristics and sales volumes/velocity), and storage types, among other factors. Periodic or selective re-slotting (or slot maintenance), focuses on regularly moving products to maintain efficiency and reduce travel, among other goals.
Machine learning technology offers an approach to dynamic slotting that is an alternative to manual warehouse modeling and top-to-bottom re-slotting. In this approach, machine learning algorithms analyze the wealth of data generated in DC operations to recommend slotting changes on an on-going basis.
This is similar to slot maintenance but without the consulting and engineering costs of traditional slotting solutions. (For an introduction to Machine Learning technology, read the recent Lucas white paper: Making Sense of AI and Machine Learning for the DC)
Goods-to-person picking solutions are an increasingly popular approach to reducing travel over the past 20 years. Instead of sending workers out into the DC to fetch products, good-to-person systems eliminate travel from a picking process by bringing the products to people at fixed picking stations. These solutions have evolved from carousels and AS/RS systems to highly automated “robotic” warehouses that are being built throughout the world, largely to support grocery ecommerce.
Carousels have been used since the 1990s to store smaller products densely in bins on shelves that rotate to present hundreds of SKUs to a picker at a single pick or pack location. Workers may need to walk between carousel locations, but travel distances are relatively small considering the number of pick locations they can access at one position. The focus is on maximizing pick density
2) AS/RS Systems
AS/RS systems (automated storage and retrieval systems) follow a similar approach to carousels, but they often rely on rack storage and conveyor systems to transport products to workers at order assembly stations. Goods-to-person AS/RS systems eliminate travel in picking and putting products away, but high capital costs have made them unsuitable for most warehouses.
3) “Robotic” Picking Systems
Over the past 10-15 years, AS/RS systems based on shuttle technologies and dense product storage have evolved into tightly integrated “robotic” picking systems.
UK online grocer Ocado is one of the pioneers in this area. They have built a number of large automated fulfillment centers to serve larger metropolitan areas. Kroger is partnering with Ocado to build up to 20 automated fulfillment centers in the U.S. Traditional AS/RS and material handling automation providers are getting into the act by developing their own turn-key fulfillment solutions for high throughput, high SKU ecommerce operations.
In contrast to Ocado, which is focusing on large facilities, some other grocery retailers are using similar types of automation technologies to build micro-fulfillment centers in urban or suburban locations. A number of North American Grocers are installing micro-fulfillment centers within or adjacent to existing stores. Loblaw, for example, is building an automated micro-fulfillment center within a large store in Toronto in partnership with Takeoff Technologies.
Outside of grocery retail, few other industries can justify the multi-million-dollar investment costs for these advanced, automated goods-to-person solutions – whether in all-new dedicated ecommerce fulfillment centers, or in existing DCs.
Autonomous Mobile Robots
Autonomous mobile robots, or AMRs, are the latest trend in warehouse automation. AMRs use many of the same technologies as self-driving cars and trucks. Scores of start-up robotics companies are applying these technologies in a variety of AMR solutions specifically targeting the warehouse and distribution market. (Still other companies are developing robots that can both pick and transport products, but these systems are early in their commercialization so are not considered here.)
1) Goods-to-Person AMRs
Today’s goods-to-person AMRs take two forms: some that take cartons from shelves to picking stations and others that carry shelves to pickers, similar to Kiva. These solutions are typically paired with dedicated pick-pack stations. They are most applicable to each picking operations with relatively small order sizes.
2) Autonomous Picking Carts
Other AMR systems are intended to function as self-driving picking carts that pickers follow to complete their jobs. The carts may be outfitted with a tablet, scanner and lights to give users their picking instructions and confirm activities. These solutions save users travel time moving carts to and from picking zones, but they do not measurably impact travel within the picking area itself.
3) Conveyance Robots
Several start-up companies are offering robots that are intended to convey products over long travel distances between DC areas. These conveyance robots offer DCs a more flexible and cost-effective alternative to conveyors or traditional AGVs.
4) Autonomous Lift Trucks
There are also a number of companies with a new take on AMRs for full pallet operations. Several start-ups and traditional fork truck manufacturers are adding vision guidance to warehouse vehicles so they can operate autonomously (or, in some cases, semi-autonomously). These self-guided lift trucks show great promise for reducing labor in the most travel-intensive activities in the DC, like pallet put away and replenishment. These tasks, however, don’t typically require as much labor as each or case picking operations.
5) Robots-to-Goods Picking
The latest twist on AMRs for each picking is robot-to-goods AMRs, or swarming robots. In this approach, numerous AMRs travel among workers in pick zones. The robots carry totes to pick locations and meet workers who pick and place the items in the totes based on instructions on a tablet mounted to the deck of the robot. Some ecommerce DCs report doubling or tripling the productivity of their pickers, with a ratio of three or four robots per picker (this is a similar ratio to goods-to-person AMR systems).
The robot-to-goods approach is being used today primarily for each picking in ecommerce fulfillment centers. Unlike goods-to-person AMRs, the robot-to-goods model does not completely eliminate human travel. One drawback is that the pick sequencing is intended to optimize robot activity, not necessarily the travel or activity of the workers. And top potential pick rates for robot-to-goods systems are less than half of the maximum rates for Kiva-style goods-to-person systems.
What Are the Advantages of AMRs?
Compared to traditional fixed automation systems, AMRs offer a number of advantages, including flexibility, scalability, and lower cost.
- Flexibility: Since AMR’s have free reign to travel throughout a facility, they can often be deployed in existing facilities. In addition, DC layouts can be changed after deploying robots without reconfiguring any hard infrastructure. Likewise, you can define new drop off or induction points on the fly through software.
- Scalability: While traditional automation systems are sized to handle maximum throughput (which may only be reached for a few weeks or months per year), robotic solutions can be scaled up or down by adding or subtracting robots and/or pick-pack stations.
- Cost: AMRs cost significantly less than conventional automation systems, but they are by no means a low-cost solution. For DCs that couldn’t justify an investment in conventional automation systems, conveyance AMRs have significant cost and ROI advantages. But goods-to-man or robot-to-goods AMRs offer a less compelling investment case. These AMR picking systems require three or more robots per picker, which represents an initial investment of greater than $1,000,000 for a relatively small operation.
A Software Approach to Reducing Warehouse Travel: AI-Based Optimization
Lucas Systems has taken a different approach to reducing warehouse travel that builds on lean process engineering and applies AI-based optimization technologies that we refer to as Jennifer™ intelligence.
Jennifer increases pick density and reduces travel in picking and other travel-intensive DC activities. Additionally, AI-based optimization does not require any changes to warehouse layouts or storage systems, and the upfront cost is fraction of robotic picking solutions. In addition, Jennifer simultaneously optimizes the travel of robots side by side with people.
Using AI to optimize DC processes has proven to be powerful in ecommerce pick-to-cart operations, where initial customers report doubling productivity without any fundamental changes to their picking processes. Although the biggest productivity gains are seen in each picking, the tool is equally applicable to case picking, replenishment, and other activities where workers are visiting many locations per work assignment.
Jennifer reduces travel through multiple components:
1) Intelligent Batching
Jennifer uses order, inventory, and location information from WMS and other systems and applies real-time optimization algorithms to create batch assignments. Unlike simple rule-based batching used in a WMS (such as FIFO or product and location overlap), Jennifer considers order priority, pick location, travel cost, product attributes, and other factors to create optimal batches of work.
Jennifer evaluates millions of potential combinations to determine the “best batch” or grouping of work from among the available orders. The math behind this is daunting. If 1000 orders are available for batching, and you are trying to create batches of four orders, there are more than 41 million possible combinations. Jennifer runs through the combinations in less than a second as users on the floor request work.
2) Path Optimization
Traditional picking systems use simple pick sequences that direct workers up one aisle and down the next in a snaking pattern. Jennifer’s path optimization algorithms use a virtual map of a facility to compute an ideal travel path that does not follow a strict location sequence. The algorithms take account of a user’s starting and end points, aisle travel restrictions (one-way aisles, for example) and other factors.
Path optimization is applicable to picking, replenishment and other activities where individual work assignments span widely dispersed locations.
3) Waveless Picking/Order Streaming
Jennifer supports waveless order picking (sometimes called “order streaming”) which enables DCs to continuously prioritize, batch and release work as new orders are received. Jennifer makes incoming orders available for picking as they are received. The order pool is continually updated, so that an order received at 2 pm can be released for picking almost immediately. (For more detail on waveless picking, read our recent blog, “Waveless Picking is a Key Capability for E-commerce Fulfillment”)
4) Dynamic Order Prioritization
To increase pick density and create better batches, Jennifer™ prioritizes all available work assignments based on rules that supervisors, in the management console, can configure and adjust throughout the day or week. Rules can be based on ship dates, routes, carriers, order types and more.
5) Robot and Human Work Orchestration
For DCs that have or are implementing AMRs, Jennifer coordinates and optimize the movements of people working with AMRs, driving further improvements in labor productivity and overall efficiency. For example, while within picking zones, a DC can add robots for conveyance between pick zones or other staging locations. This can further reduce worker travel to and from fixed drop off or induction points. Likewise, Jennifer can be used to optimize the travel of robots for full pallet moves, both in inbound and outbound operations.
What Are the Benefits of AI-Based Optimization?
- Increased productivity and throughput: Based on dozens of trials and successful implementations, AI-based optimization has reduced warehouse travel up to 70%. A number of DCs have more than doubled pick rates in cart pick operations. Baptist Health South Florida saw a 100% increase in picking productivity in their main-piece picking area of the DC. Grocery and food DCs have also seen 15-30 percent savings in travel time in case pick applications, which translates to 8-15% productivity gains.
- Labor savings: Customers utilizing AI-based optimization have also seen major reductions in labor hours. For example, Apex Tool Group, a leading manufacturer of hand and power tools, now operates with 10% fewer person-hours. The most we’ve seen is a 50% reduction in labor hours.
- Increased throughput and shorter turnaround times: Improved productivity and a steady work release (through waveless picking) increases throughput and results in more efficient allocation of workers and equipment. Next-day and two-day orders are out the door on time and in the hands of the consumers on promised delivery dates. For example, Rotary Corp, the worlds largest supplier of power equipment parts, was able to achieve 99.99% same-day shipping.
What Strategy Is Best For My Warehouse?
To choose the best travel reduction/productivity strategy, warehouses and distribution centers need to look at their budget, current processes (case picking vs piece picking), facility size/SKU count, current infrastructure, etc.
However, we’ve learned that many top-performing DCs have maxed out their productivity gains and travel savings through traditional travel reduction strategies. For example, slotting solutions are stagnant and happen once. Without implementing solutions that utilize AI and machine learning algorithms to sot dynamically, most DCs won’t see impressive and continuous results in an ever-changing market.
Given these changing market factors, most facilities are looking for other ways to eliminate travel that don’t require fixed, inflexible automation systems. Robotic systems offer a more cost-effective, flexible, and scalable alternative to traditional material handling systems.
For DCs looking for a low-risk, high-return solution that can be implemented quickly AI-based optimization may be the answer. Lucas AI-based optimization has more than doubled productivity without requiring any changes to warehouse layouts or storage systems. In addition, the technology can be used to optimize the work of people alongside robots, allowing DCs to add more AMRS where they make the most sense today.