Sortation Best Practices:

Using Your Head To Reduce Overhead

By Darrell Krasoski, Principal

The most important new development in sortation is not in equipment. It's in thinking.

The most innovative, progressive best practice for sorting products in distribution operations lies in understanding the overall demands of year-round operations and designing sortation systems that handle peaks without breaking the bank or introducing complex machinery that invites crippling, single-point failures. You can find the fastest machines that can sort the highest peak volumes, but is that what you really need? A solid analysis of less-expensive, moderate throughput sortation technologies—coupled with a thorough understanding of order profiles and distribution needs—can lead to a parallel sortation system, one that teams up a network of smaller devices to achieve the same results as high-powered, high-priced hardware.

What's New Out There

New sortation development in recent years has centered on three areas. The first area is new technologies, many of which have been slow to penetrate the market because many system users have found them to be too complex or not able to provide value. The second area is upgrading existing technologies for more reliability and less noise. This has been a welcome trend as end users appreciate vendors who listen and learn. The third trend is the development of mid-range sortation technology that fills a gap in sortation throughput ranges at a moderate price.

Finding the Right Approach for the Majority of the Market

High-end users have always used high-speed systems as their backbones—tilt-tray and sliding—shoe sorters are the norm. Mid-range users have always struggled with the cost of those systems and have typically settled for low-end systems such as pop-up wheel and divert sorters, and even lower end users often opt for manual systems. The lower-cost systems lack the capacity of the higher-technology options, but many users cannot justify the cost of higher-tech equipment. One answer that takes advantage of midrange equipment and can still yield the desired results without high cost, is parallel processing sortation. Parallel processing sortation can be an answer for the users with mid-to-high requirements but who may have trouble funding the highest technology. With this approach, users can achieve high speed and throughput at lower cost. In addition, inventory and hardware control systems such as warehouse management systems have evolved to a point where they can manage parallel processing sortation.

In one design by Tompkins Associates, lower-cost sorters that each had less throughput capacity than higher-tech equipment were networked together, and the system used the parallel approach to achieve the required overall throughput. Designs like this one can grow incrementally as demand increases, but do not require the user to purchase now in order to fulfill a projected peak.

Such an effort requires careful thought on the upstream side of sorting, including a thorough understanding of order profiles and volume patterns, but not extensive, expensive hardware. A parallel design by nature can meet changing needs and changing forecasts by using the control of product to the sorters to manage the required throughput of the sortation system.

What Is the Goal for Sortation?

Consider this: DC pick engines are all sending product to outbound sorters. The traditional approach says that as the total volume handled by the pick engines grows, the DC needs bigger and bigger sorters. The smarter approach, however, is to understand the entire operation in order to see if it is amenable to having "decision points" at which product is sent to parallel lower-volume—but also lower-cost—sorters. If so, a lower-tech system of parallel sortation may offer a more reliable and less-expensive alternative.

The concept is analogous to the thinking that goes into designing the highway system to service a major sports arena. The stadium itself must be built for peak-volume capacity in order to maximize income from ticket sales. However, those fans will be less than completely satisfied customers if they must spend two hours trying to leave the stadium parking lots through one highway. To minimize distress and reduce cost, traffic engineers design stadiums to have multiple exits onto a network of roads. The fans still have to get to their homes, but the load is distributed across a number of routes. Even if half of the fans eventually go east, sorting the traffic load onto several roads before they combine to go east on an interstate highway reduces traffic jams at the stadium and speeds up traffic flow.

A best practice using parallel systems for distribution sortation employs the same logic. Four two-lane roads can clear traffic as fast as one eight-lane highway, will be cheaper to construct, provide flexibility and they may be able to employ some existing infrastructure. When there are no games (which is the vast majority of hours during the year), the smaller roads can serve day-to-day traffic. Multiple sortation pathways using lower-throughput proven technology can match or exceed high-volume single hardware that is designed to meet peak load but is used for lower volumes much more of the time.

And, in the end, any sortation system is designed to separate, not amass, product for delivery to multiple docks.

Making it Happen

Once the thought process and analysis is done and a flexible design evolves, there is one necessary ingredient to pull it all together. The WMS, or order system, has to convey its command to the picking and conveying systems that in turn have their own controllers. Tying them together is what makes it happen. Middleware software and hardware are vital to have the whole operation work smoothly and with the flexibility required. Tompkins has seen with its own middleware product, the Tompkins Control System, that the proper control system eases implementation, daily operation and maintenance.

The Take-Away

The traditional approach in sortation has been to seek higher sorter speeds to do all the work at the end of the order process as volumes increase upstream. If we think earlier in the order process, several lower-cost systems could handle the same peak volumes while offering the ability to scale down during less-than-peak periods, easing the scheduling of maintenance, and eliminating the possibility of single-point failures that can cripple an entire operation. This approach can work for the entire pyramid of sortation users, from the few very-high-volume companies through the larger universe of smaller businesses. All the necessary hardware and technology exist. The information exists. Most warehouse management system today will allow this kind of design. What it takes is thinking differently, truly understanding operations and focusing on the results.

Case Study: Parallel Sortation In Action

A successful online retailer had shipping requirements that were quite unbalanced. At many national holidays such as Christmas and Easter the customer needed to have a flexible high-speed shipping system. The requirements at these peak periods were two to three times volumes during the rest of the year.

The originally proposed solution was for a very long high-speed shoe sorter with approximately 24 divert points. This was a conventional solution, but the cost of the sorter was significant. During the busy seasons the customer could not tolerate any downtime that could arise from a single point of failure system. The customer understood that the maintenance requirements could be quite high and therefore needed a system that would be more fault tolerant.

To address these concerns, Tompkins conceived a system comprised of a pair of narrow-belt sorters that were a about half the investment of the shoe sorter. By tracking and routing the cartons precisely and allocating the proper sort plan to the sorter, the periods of very high shipping were accommodated. During normal shipping periods the units were rotated in service to limit the wear and tear on equipment and allow for preventive maintenance.