Zone Picking Systems: Beautiful and Complicated

Zone picking systems (often called zone routing systems) have been around for decades. And yet, it seems they are being underappreciated these days. Part of the reason might be that few people understand how they work and how their potential can be tapped. And just to get this straight: everybody understands – in principle – how these systems function, and there are dozens of companies on the market that can build simple zone picking systems that will work. Very few individuals understand, however, how zone picking systems can be turned into highly productive picking systems. Media attention is clearly directed towards goods-to-person picking concepts, and increasingly often to those based on autonomous mobile robots. These systems can be highly productive; they do come at high initial capital expenditure, however. Zone picking systems, on the other hand, can be built at fairly low cost. A conveyor-based zone picking system requires shelves, conveyors, and IT. Conveyors are mature technology, and no expensive AS/RS are needed. They can be installed in just a couple of weeks, are cheap to maintain, and, if designed properly, can almost match productivity of goods-to-person systems at a small fraction of the cost. 

As with all automation-supported picking solutions, economic applications of zone picking systems requires that the balance of static and dynamic performance requirements be in the right range („sweet spot“ would be the a good description if that balance were to be found at only one balance point; however, there can be many). The static performance requirements on zone picking systems are derived from the number of SKUs that are to be included in the system. So, static performance requirements effectively means: number of pick-from locations. Dynamic performance requirements are derived from the number of picks to be performed in the system. As with other automated systems, either variable can be the driver of system size and setup.  

The two variables are co-dependent, however. Under the assumption that each SKU added to the system is subject to some additional customer demand, the greater the share of SKUs to be included in the zone picking system, the greater the number of picks to be performed in the system will be. Because most, if not all, stations (zones) in a zone picking system need to be manned, you don’t want to have a large number of stations to represent all the SKUs you want to include in the system if they don’t generate sufficiently many picks to keep them busy. Similarly, you don’t want to have a very low number of zones that would be overloaded with picking orders and can’t keep up with demand. Identifying the right number of SKUs in the systems to be picked in the right number of zones is not rocket science, but it does require some careful analysis. To make things more complicated, you can vary zone size such that you fit more or fewer SKUs per zone. Generally, the shorter a zone, the higher picking performance can be. Also, the shorter a zone, the fewer buffer locations are available, making load balancing in the system more complicated (eventually resulting in lower picking performance). Picking performance in very long zones can approximate picking performance in conventional person-to-goods systems, which would render installation of a zone picking system unnecessary.  

Apart from designing the hardware of the system in the right way, defining the order release logic can turn out to be rather complicated. Orders need to be released into the system such that all manned stations receive enough order totes to stay busy with picking (else productivity numbers will drop) while not overloading any individual station, else order totes will begin to loop on the conveyor belt, jamming the system. The fact that order totes – depending on order size and structure – may have to visit several zones for completion makes the order release logic even more complicated. And with every tote looping in the system when a station was overloaded, the behavior of the system becomes less predictable, and the order release logic will have it increasingly difficult to understand which orders should be released next and to prevent the system from spiraling out of control. Add time constraints for individual orders („instant delivery“ or similar) and the mess is complete.  

So, zone picking systems are great. They can be highly productive at very low initial cost with even lower running cost. The hardware requirements are low (conveyors and shelves), and the technology is mature and can be provided by almost any logistics automation company. But while they look really simple, they are, in fact, complicated to design and to control. The result is that most zone picking systems do not enable outstanding productivity; their potential and their challenges tend to be insufficiently understood by sales engineers and IT staff of logistics automation companies. The few companies that are lucky enough to have well-designed and highly productive zone picking systems in place consider them a true competitive advantage and are secretive about them – and rightfully so!  

Talk to us if you want to understand better if a zone picking solution could be right for you.