The rise and rise of voice-based systems for picking has been one of the most remarkable facets of logistics operations in recent years. Voice technology has gone from being clunky, slow and error-prone to top of the list in less than a decade.
But, that does not mean it is appropriate for every situation. There is a strong case for automated systems and pick-to-light for some applications – and there might even be times where a manual system will do a good enough job for minimal investment.
Darrel Williams of Psion Teklogix says: “Like all technology, voice recognition is evolving, but it can no longer be perceived as the ‘new kid on the block’ or an untried, untested or cost unjustified tool.”
Understanding what is truly important to the operation is key to making the right investment decisions. Speed, accuracy and cost are obvious considerations. But, says Brian Turner, development controller at Gist: “I would add scale, robustness, flexibility and seasonality but I would also add questions about the stability and predictability of the operation we are installing it into. For example, is the business growing or reducing? Is it acquisitive? And lastly, I would add the softer factors of ‘how will the workforce react’? ‘Which would they prefer’?
“Companies introducing voice picking, for example, are seeing unplanned improvements in productivity because pickers are spending longer actually doing the pick job rather than talking to their fellow operatives, and the pickers report that they actually prefer using the voice-pick technology. That’s a win-win that wasn’t foreseen in the pre-implementation modelling.”
TDG has been operating a 330,000 sq ft distribution centre for Argos at Mossend for more than four years. It has 29,000 pallet spaces, 7,500 pick points, and staffing non-peak of around 260 rising to more than 500 in peak. Argos took a strategic decision to go down the voice picking route. Gerry Friel, general manager of the RDC, says: “The client decided that voice picking was best and balanced high productivity levels with accurate picking, and this is borne out in our daily measurements, across the Argos network.
“The voice system within Mossend is relatively fast. In the main pick area, speeds of some 95-100 items per hour are possible (inclusive of two-man picking in some products). In the small item area speeds of 225 items per hour are possible allowing the accuracy to be optimal. The site is continually looking at warehouse layouts as products come and go in each catalogue to optimise warehouse efficiency. In general terms, accuracy must take precedence over speed but the balance must be right for any warehouse operation to work to its best efficiency,” says Friel.
For Paul Evans, IT and infrastructure manager for Oakland International, cost is probably the most significant factor initially when deciding, as it is difficult to quantify the benefits of speed and accuracy.
And Philippe Bentz, senior business consultant at Kewill, points out that, “if you are a typical 3PL, offering pick and pack operations to multiple customers, speed, accuracy and cost are obviously key elements. However, the complexity resides in how to handle your customers’ specific needs. Usually, a 3PL will not only organise the picking and packing of goods on behalf of their customers, but very often will also organise the final dispatch and delivery. In this scenario, it is essential to consider how your transport planning can assist you in optimising your picking process.
“If you have a fixed schedule of deliveries through a delivery network, you can use this to your benefit, as your trip planning will allow you to work back when best to start picking to meet the departure of scheduled vehicles on delivery runs.
“This is commonly known as ‘backward scheduling’ and ensures that, by taking care of the resources available for picking, the length of time required to pick the complete order, and the other orders to be prepared, the picking instructions are released in time to the picking operators to meet the scheduled departures.”
Speed and accuracy are essential for Steve Porter of Linde. “I would place equal emphasis on these two factors as orders must go out correctly and quickly. If one were asked to say which is more important, I might tend to think that speed will wind up as the main driver, since it is likely that the percentage of inaccuracy might be less of a problem than (for example) a poor pick rate. This is because the picker, who can do (as an example) 200 cases an hour with a ten per cent inaccuracy rate is still picking 180 accurately, and that is better than a picker in the same operation only doing 150 cases. But again, it is difficult to offer a precise answer.”
Roger Peart of Vanderlande highlights the fact that the choice of picking concept is critical. “For example, man to goods or goods to man. Order pick, batch pick, zone pick and so on.”
“Solutions such as goods to man are more flexible in terms of accommodating increases in the number of different product lines (SKUs) handled, as they are not subject to the limitations of a conventional pick face,” points out Simon Barnwell of Dematic.
“Some solutions are also more efficient in handling orders from a diverse range of customers, such as retail outlets including large supermarkets and convenience stores. Others are more suitable for handling just one type of customer. Finally, some solutions enable store-friendly orders, where items are packed in the order required for easy store shelf replenishment, improving in-store logistics.”
Terran Churcher, managing director of Codegate, homes in on the fact that voice directed and light activated tend to be proprietary systems while a WiFi-connected HHT-based system is likely to be open architecture.
“Over a three year period, cost of ownership of each will vary but the variance is unlikely to be considerable. Light activated will have higher installation and continued maintenance cost than voice activated, while voice is best suited to the lower speed picking environments and demands personalised operator equipment. Taken as a cost per pick over the life of the system, light activated will probably prove the most expensive, followed by the HHT’s and lastly voice activated.”
Peart says: “Voice picking and RF give similar pick rates in a zone picking system with a similar investment. Pick-to-light in this application may be justified for a proportion of the goods to be picked but is expensive if installed across the board. Pick-to-light comes into its own for a goods to man solution where product totes are transported to a static pick station satisfying several orders.
“A critical consideration in goods to man systems is the ergonomic design of the work station to give the picker a sustainable work rate without repetitive strain injury. Moving product to/from, and around the picking functions in an automated/mechanised environment will generally be a combination of conveyors linked to possibly a mini-load ASRS. Depending on volumes packing and sorting to dispatch will be evaluated,” says Peart.
David Hibbett of SSI Schaefer is a strong advocate of automation in the picking area, arguing that if you want faster reaction/cycle times, on-time delivery, same-day delivery, high picking accuracy, order fulfilment, manpower savings, store friendly picking, RFID, tracking, tracing, information transparency and visibility with a significant reduction in damaged stock then automation is the only way forward.
“Time and time again SSI Schaefer is faced with the same responses ‘it’s not economical’, ‘we’ve no resources or time’, or quite simply ‘we don’t need it’. SSI Schaefer is urging companies to think ‘outside the box’ and to look at the range of proven concepts that are available and search for opportunities to handle today’s supply chain challenges.”
SSI Schaefer has just launched a new automated picking system: Schaefer Case Picking (SCP). The SCP system is a complex modular, scalable solution with expansion possibilities for fully automated order picking and warehousing processes.
In today’s consumer driven world the customer demands that a product be available to him/her when he/she wants and not when the retailer chooses, says Peart. “To meet smaller order quantities quicker delivery schedules and lower stock holding then automation becomes the viable option.”
Flexibility is a key consideration and Russell Davies, managing director UK of Savoye Logistics points out that as companies grow so does the flexibility requirement of the warehouse solution, be it automated or manual.
“Automation should only be used in areas of the warehouse where it adds to the overall fulfilment process by either reducing headcount, responding to increased demand, increasing service levels or a mixture of all three. It is seldom used to increase flexibility, in this area there is nothing as flexible as a human. In a move to provide an answer to the flexibility question, Savoye introduced the Picking Tray System (PTS) – a flexible goods to man solution,” says Davies.
“Goods to man picking is not a new concept and has been around in many different forms for a number of years, but has normally only been used for a small section of the product profile, generally slow moving products where the capital investment is offset against substantial increase in pick rate.
“The range of SKUs in a conventional goods to man system is generally linked to the throughput of the machine, so if a system could deliver 250 different products to the picker per hour then that generally defined the products to be stored in it as there is no point in putting fast moving products where those 250 products are picked frequently. The only additional way to then increase the rate is to pick multiple orders of the same item at the same time. In a flexible system this may not always be possible, as this often means the pick is forced to be the first pick,” says Davies.
The decision whether to go for PTL or voice is dependent on the number of SKUs, says Barnwell. “As the cost of PTL is per SKU, they are suitable for picking a limited number of SKUs, typically fast movers. The cost of voice, on the other hand, is per operator, so they are suitable for systems handling a large number of SKUs, typically slow and medium movers. There is no conclusive evidence that voice picking or PTL differ in terms of accuracy. Although productivity is highly dependent on the design of a system and how SKUs are laid out in terms of fast versus slow movers, PTL is approximately ten per cent faster. Some operations, such as the clothing retailer American Eagle, are opting for dual-technology systems; PTL is used for the limited number of fast movers and voice for the rest of the range. This approach ensures the most effective technology is used for different parts of their range.
Tom Kozenski of RedPrairie says that he has come across cases of PTL systems being taken out because they could not cope with large changes in the number of SKUs being picked. PTL offers high levels of accuracy but can’t accommodate an increase in the number of pickers easily.
RedPrairie has seen a dramatic growth in the use of voice, he says. Compared to a standard RF system, voice can give an eight per cent improvement in productivity. The old weakness of voice was throughput, but the technology has improved significantly – particularly the performance in noisy environments which could confuse the system.
Williams says: “The introduction of voice recognition systems that enable warehouse employees to have their eyes and hands free during their daily routine, can have a dramatic impact on productivity levels. But it is not only beneficial in an order selection scenario. The system can also be applied to achieve significant efficiencies throughout the entire supply chain operation – from goods-in to put-away, replenishment and even dispatch. In fact, one of our clients runs their entire operation on voice technology and the gains they have made have been phenomenal.”
Aldata has been supporting Vocollect-based systems for several years but now, says managing director Mark Croxton, it is also offering a voice application via a PDA. This, he says, is cheaper than proprietary systems opening up opportunities with companies for whom traditional voice systems have been too costly. The Aldata system uses voice recognition technology from Nuance.
He points out that in a seasonal business, there will be proprietary kit doing nothing in the troughs but PDAs can be used for other activities. “Once you have got voice working on a PDA, you can also put other applications onto it, such as scanning, printing and RFID.”
Views differ on the trade-off between accuracy and speed. Gist’s Brian Turner says: “By implication, if I just go for speed and don’t check anything I get fast pick rate but low accuracy and conversely if I check every item I get a slow pick rate but high accuracy. Again, the solution is usually somewhere in between. Some customers need accuracy and are prepared to pay for it to reduce waste, improve the customer experience and maximise sales, whereas others can accept a reduced accuracy for the benefit of a lower price. The real key is to understand the impact on the business of reduced or reducing accuracy.”
Bentz argues that there can never be a trade-off between the two. “At the end of the day, if a 3PL puts speed ahead of accuracy, the stock will end up being wrong…and the speed will disappear when trying to find out why the stock quantities are incorrect and the orders cannot be fulfilled. On the other hand, if accuracy is put ahead of speed, customers may be dissatisfied by a poor quality of service and stop re-ordering.”
Ultimately, says Turner: “The ideal picking solution is often a hybrid of several different picking technologies and as the technologies develop, the opportunity for linking them and deriving benefit increases. To a large extent the opportunities depend on the trust we have in the technology, that it will perform, and the accuracy of the information that is passed into and between the component parts.” Guide for initial consideration of options – the final solution is likely to be a compromise of several factors.