Time was that the main worry in warehouse management was to ensure that one had enough stocks to fulfil orders efficiently without running out but not too much so that slow-moving stocks became a serious financial burden. Inventory holding costs, after all, could surpass all other warehouse running costs combined. In the past such a problem was partly addressed by stock (demand) forecasting programs that could react in real time to big weather forecast changes, like pending heat waves, particularly valuable in the drinks industry where demand could soar four-fold over 24 hours. These programs, of course, are still widely used and can reduce total stocks by one third without affecting customer service quality.

The problem today, however, is that new retailing formats like those generated by e-commerce are forcing a revolution in how order picking is executed, where the pointers are clearly emphasising a diminished role for human pickers as they are replaced by automation. This has meant a new uncertainty plagues the planners, namely how to prevent early redundancy in a costly new automated warehouse system. Such a system that a buyer forecast to last three years could end up lasting only three months, such are the uncertainties generated by e-commerce order fulfilment. The problem has been aggravated by the more frequent peak demands created by special, nationwide promotion days like Black Fridays. In such businesses the difference between the lowest volume day and the highest could be tenfold. The MHE industry has tried to respond to this by making their equipment more adaptable and scaleable but the jury could still be out about their effectiveness in a more frenetic, order-picking world.

While it may be true that if you don’t automate you don’t stay in the marketplace it is important to ensure the best mix between manual and automated picking, made more urgent owing to the extremes placed by online shipping. It does not make sense to invest in top-end technology for peaks only six times a year, for example. Those demand patterns should be fulfilled by labourintensive, low investment equipment because costly automated equipment should be sweated throughout the year. One solution, therefore, could be to use top-end automation for, say, 35 weeks of the year for your core volumes and at peak times when you need to employ more people to complement that with other techniques like voice picking systems that could take 25% off the wage bill or double productivity.

Given that it is becoming much harder to handle large volumes of single-piece orders via traditional picking methods in ever-shorter delivery times, it is important that whenever introducing robotic picking to make the employee training as quick and simple as possible. A good example is the goods-to-person robotic picking from Locus Robotics used by Geodis at its Indianapolis distribution centre. Geodis began looking for a safe, scaleable, productive solution that was easy to implement and required minimal training at a time when it was plagued by a dwindling labour market. These remarkable robotic pickers, resembling multilingual AGVs carrying two tote boxes, allowed team members to be taught how to pick to the robots within minutes. By picking SKUs to the robots, personnel are saved the chore of pulling pick carts. The result of this investment is that 80% of SKUs are picked to the robots daily, employee productivity has doubled, training time cut in half and pick accuracy much improved.