After seven years of unbridled growth, it was perhaps inevitable that Internet shoe retailer Zappos.com Inc. began experiencing growing pains last year. Early on, the company had made a strategic decision to handle its own warehousing and order fulfillment. But as sales surged toward the $600 million mark last year, the fulfillment operation started displaying unmistakable signs of strain. The error rate began to climb and throughput began to lag. Many companies facing that situation would conclude that they needed more people. Zappos decided it needed something else: more intelligence.
To boost the operation's productivity, the e-tailer installed a warehouse control system (WCS) in its 400,000-square-foot distribution center in Shepherdsville, Ky., last fall. The system, FKI Logistex's Warehouse Optimizer software, essentially serves as the brains of the operation, overseeing the facility's high-speed conveyors, sorters, merges, and other material handling equipment. The WCS takes order information from Zappos.com's homegrown warehouse management system, then manages product flow to and from various portions of the warehouse, including receiving, presorting, put-away, picking, packing, and shipping.
And that's not all it can do. Along with choreographing the material handling equipment's movements, the intelligent control system automatically balances the workload at the various packing stations. The WCS monitors the actual workload for the number of orders and products per order per lane, and then makes decisions on routing product to optimize the workload across multiple packing stations.
"Previously, all the movement was handled by people and conveyors without intelligence," says Jonathan Field, director of development at Zappos."The error rate was high and throughput wasn't high. That's why we decided to switch over." The decision has paid off, he says. "This year, on our busiest days, we shipped much more product than last year and we got a lot more volume through the distribution center."
Growing capabilities
Zappos took advantage of the fact that warehouse control systems have gotten smarter over the years. In the early days, the WCS served mainly as a messenger boy for the warehouse management system (WMS). The WMS would gather up data on the orders to be filled from the company's main computer system and devise a work plan. Then it was up to the WCS to relay directions to the material handling equipment—the conveyors, sortation systems, and the like—to see that the orders were carried out. It did that by communicating with sensors to direct the opening and closing of gates to direct product onto a specific conveyor arm or chute. "In the traditional world of warehousing, the WMS functions as a thought process for a batch of work and passes that information to the machine control system (WCS) that does the task management," says Daniel Ahrens, a product manager at Fortna Inc. in West Reading, Pa.
But as material handling systems have grown more sophisticated over the years, the WCS has evolved to keep pace. Today's systems are capable of controlling not just conveyors and sorters, but pick-to-light systems, radio-frequency identification systems, and voice-directed picking systems as well. "The mechanical systems are getting more and more advanced," says Michael Hahn, U.S. chief sales officer for Knapp Logistics and Automation Inc. in Kennesaw, Ga., "and you need more intelligence on the software side to control them in the right way."
Perhaps more importantly,WCS have also begun to think for themselves. Many of today's WCS are imbued with extra "intelligence," algorithms written into the software that can react to feedback from equipment sensors and adjust material flows—something that a WMS, which generally doesn't have a direct connection to photocells and scanners, is not in a position to do. "The WCS has the touch point with the WMS to know what orders are coming into the warehouse operation," says Ahrens, "and it has the touch points with machine controls to know what it's doing."
Among other things, that capability allows companies to maximize throughput and inventory velocity in their warehouses. "Typically, you add more intelligence to achieve better product flow," says Jerry Koch, product director of software and controls for warehousing and distribution at FKI Logistex, which has its U.S. headquarters in St. Louis, Mo. "That allows you to move product with the least amount of touches."
Managing the workload
What has made this possible is a technological breakthrough in the way the WCS collects and uses feedback from the sensors. In the past, the WCS could only take data from one sensor at a time and then act on the information by, say, shutting down a lane if it detected a jam on the conveyor line. Now, however, the WCS is able to collect data from all the sensors to form a big-picture view of the warehouse operation. "In the past, sensor data was only used to stop flow down a lane," says Ahrens. "Now they use that info to control upstream processes. Instead of having islands of automation in a warehouse, you can take a holistic approach to the entire operation."
Once the WCS has collected information from the various sensors, it aggregates the data and applies special algorithms to balance the workflow. "Advanced algorithms balance out the flow automatically," explains Koch of FKI Logistex. "As products move into the central merge area, the more intelligent software releases products such that there's no backing up on a conveyor line so you get maximum throughput of the material handling equipment."
Abandoning the wave
The advent of intelligent WCS has also opened the door for distribution centers to switch from wave picking to a continuous flow method. Under the wave picking approach, a group of pick orders is released at one time. Workers pick orders in a batch from the racks and convey cartons or cases to a high-speed sortation device, which in turn diverts the product to a chute or spur dedicated to each order. At the end of the spur, another worker loads the product onto a pallet or into a truck at a loading dock.
Although wave picking is widely considered more efficient than single-order or batch picking, it also has its drawbacks. For instance, if workers loading cartons into a trailer at a loading dock can't keep pace with the flow from the sortation device, the result is a backup. In addition, even when working properly, there's a natural start and stop flow to the warehouse operation as each wave retriggers the work process.
"With wave picking, you'll see shipping docks that are jammed up with product because of the emphasis on picking," says Ahrens. "The new approach is a pull-based system. Instead of releasing 2,000 orders at once, you release the orders piecemeal based on the needs of the shipping dock."
The intelligent WCS accomplishes this by controlling the release of orders to balance workloads between machines and pick lanes, or even between shipping lanes being shared for an outbound shipment. "You can look at the current workload and you control the release of orders to match the capability of absorbing them," says Larry Kuhn, president and founder of Glen Road Systems Inc. in Conshohocken, Pa., which provides warehouse control systems.
The more intelligent WCS can also perform "dynamic balancing," adjusting workflow in real time to changing circumstances. "It monitors the progress of each piece of equipment and the person, and adjusts staffers' work based on their neighbors," says Ahrens. "It can even send a signal to the supervisor to tell him that the worker in zone 4 is twiddling his thumbs," adds Sam Flanders, president of 2wmc.com, a material handling consulting firm in Portsmouth, N.H. "It can manage labor very well."
Advanced warehouse control systems give management more visibility into the warehouse operation as well. In fact, some systems even offer so-called dashboard controls, graphical displays that show the progress of work for managers to make adjustments. "Graphical interfaces are becoming more common with WCS," says Bob Harris, president of Cirrus Tech in Raleigh, N.C. "They can notify you via audio signals or visuals about problems, and they can provide statistical information on equipment so you can see what's happening in terms of equipment failures or lane backups."
Stuck on the wave
As for the results, Ahrens says the continuous flow approach has been shown to increase warehouse throughput by 40 percent without any changes in equipment. But it does require an investment in software. To take advantage of the benefits of dynamic flow, companies must either buy new warehouse control software or upgrade their existing WCS. "Legacy systems are built around batch management as opposed to balancing work flow based on information feedback," Ahrens says. "In order to realize the benefits, there has to be some upfront investment in changing the software."
Despite all the talk about the benefits of continuous flow picking, many companies have been hesitant to abandon wavebased picking. "Warehouse management likes waves because they can reallocate people at the end of each wave. It's this idea 'I can get everyone back on the starting block together,'" says Ahrens. "It's also a security blanket. I need waves to manage productivity."
Ray Becker, a vice president at the consulting firm Tom Zosel Associates Ltd. in Long Grove, Ill., agrees with Ahrens that companies still question the validity of the continuous flow approach. "From an overall perspective, it's adjusting a plan on the fly. The back end is unknown. It might optimize my wave but what happens to the work behind the wave?"
Ahrens says in his experience, winning the skeptics over is mostly a matter of time. "A client's first reaction is that this won't work," he says. "But after thinking about it, they say, 'Wow, this is revolutionary.'"
