How to avoid automated warehouse failures

Lest there be any initial thoughts that automated warehousing is only for the brave, there are now many such centres in the developed world running successfully and fulfilling their payback criteria. However, although now much more reliable than automated warehouses of the 1980s, examples of spectacular failure still occur and the consequences so devastating as to risk bankruptcy.

The reasons for failure are many and can be of an incredibly crass nature, like when a British middle ranking retailer using a third party logistics (3PL) provider as its turnkey partner refused to invest in simulation to expose potential bottlenecks. Both parties agreed on the need for simulation but neither would pay for it. The result was plenty of warehouse bottlenecks leading to empty shop shelves, a fired 3PL and a badly mauled share price. Spoiling the ship for a ha'porth of tar in this way is a potent reason why, despite such very costly investments, price should not be the prime consideration in deciding which scheme to buy and should never be at the expense of flexibility. So what are the main reasons why automated warehouses fail to meet expectations and when should they be considered for investment?

Most "automated" warehouses in Britain are a mixture of automation and manual/mechanical handling and so the payback criteria for such centres will obviously depend on the level of each. These hybrids may even involve only the automation of information technology, as with Amazon, who rely on it for their order picking functions only. Their operation is highly labour intensive, typically involving order pickers walking 7-10 miles per shift, a process some might imagine is highly time-wasting. Yet, Amazon conducts round-the-clock, effective e-fulfilment centres where delivery times to consumer are critical. This shows how careful one must be before going down the warehouse automation route.

Generally speaking, a fully automated warehouse dealing only with full pallet loads is unlikely to achieve a payback in less than five years but payback criteria can be misleading or incomplete. In such exercises, for example, accountants are unlikely to factor in the beneficial impact of warehouse automation on customers, like prompt, damage-free, and correctly picked orders. It's true that figures cannot be placed on that but it has significant value nevertheless. There is also an environmental dividend because 'dark' warehouses need much less heating and lighting.

Points favouring automation

Generally, the case for warehouse automation is strengthened as the scale and demand for speed and efficiency increase. Favouring factors include:

•  24-hr, seven-day week operations
• When repetitive tasks are regularly performed
• When long travel distances are involved
• When lift height rises above 12 mts
• When high land values dictate the need for very high bay warehouses to exploit height
• When a hazardous environment precludes human operations
• When tight inventory control is essential along with security and almost negligible damage levels. 
• When the use of very high storage heights and densities could close scattered distribution centres

The factor most likely to drive businesses to invest more in automation is the need for more speed and efficiency to cope with the seismic change in consumer shopping evinced by the remorseless rise of online shopping (e-tailing). This trend, which some estimates suggest will wipe out 40% of Britain's shops in five years, will place far more importance on single item picking, leading to demand for automated goods to pickers, be they mini load cranes, robots, carousels/paternosters, or whatever. But is automation inventiveness keeping up with these momentous changes?

Yes, they are, and good examples are some of SSI Schaefer's unique solutions. They have, for example, developed a picking and sortation system which offers the additional dimension of sequencing of order items so that different elements, or lines, of an order can be picked over a period in different areas of the warehouse, then held and delivered together to one packing station. This has the obvious advantage of being able to consolidate all items into one package as well as being able to prioritise packing of picked order items. It is based on overhead conveyor system technology, called the Fulfilment Factory, and it was devised primarily for e-commerce operations.

Truly fully automated systems are increasingly using a 'teach in' system which captures physical appearance, weights, dimensions and other statistical product data upon goods receipt and before put-away. This enables pallet delayering and put-away on trays, subsequently allowing automatic picking and multiple case selection, sequenced in order to build mixed SKU pallets using a robotic palletiser and wrapper.

Avoiding the tender traps  

The business of avoiding costly project failures starts with the selection process of a suitable project provider. For the more complex projects, taking the turnkey route makes the best sense when starting with a green-field site but even here great care is needed. At one time in Britain a significant number of projects relied on the builder acting as the turnkey main contractor. Fortunately, that has waned following some bad experiences. The fact is, a builder does not have the expertise on the special problems that can arise with warehouse automation. All automated MHE suppliers agree that an automated warehouse must be designed from the inside outwards and so it is the working equipment that is paramount. The building, after all, is only there to keep the rain off.

Another route to take is the appointment of a 3PL provider to build and run the project. This may be reasonably safe provided the 3PL has plenty of experience building, integrating and running automated stores. A third route is to appoint a systems integrator as the project manager but the downside outweighs the upside of this choice. The safest turnkey choice would be to make the main handling equipment supplier as the turnkey operator for they should have a good track record for successful projects, helped by their own software expertise. They will also have the substance and after-sales technical service support as a comfort if things go wrong.

Once satisfied with the choice of turnkey partner, the client must then watch out for the warning signs at the tender stage. These include:

• Exaggerated claims
• Lack of well-defined quotation for equipment, with too many qualifications and exceptions
• Lack of information on commissioning trials and acceptance tests and how the system will achieve the throughputs
• Sensitivity of the system to single point failure in mechanical, electrical and control areas
• Design life of equipment at specified duty
• Lack of established support services for equipment and software
• Enough potential throughput capability to cope with sudden unexpected peaks

Watch that packaging

Packaging may seem of little consequence but automated warehouse operators cannot afford to ignore it. If there are two lessons that can be learned about the hazards of packaging for automation they are: consult your handling designer as early as possible, and that the automation solution should come before cost. Packaging should heavily influence the design process. Compared with manual handling, warehouse automation requires better quality packaging and pallets, more consistency and more care over palletising and labelling. Poor quality timber pallets are probably the greatest problem for warehouse automation. Pallet damage is a big factor, with broken boards and protruding nails jamming conveyors, especially the roller type. Some operators prefer plastic to timber pallets because of their greater dimensional accuracy and other advantages, though good quality timber pallets like those from pallet pool operators Chep and LPR would be acceptable. Belt conveyors pose fewer problems than roller conveyors when meeting with loose stretchwrap or banding owing to less snagging. Loose wrapping materials can also give incorrect signals to sensors and cause inefficient accumulation or machinery stoppages. 

Finally, the client must examine its role in making the venture a success even more rigorously because it is here that the seeds of failure flourish. Most MHE suppliers would agree that most of the many reasons behind disappointment lie at the clients' doors and topping the list is poor planning with not enough attention paid to functional specification. There is too little effort put in at the front end to the cost of the back end. Clients must realise that their own attitudes can seriously impair efficient working. A system supplier must have accurate information on the business issues underlying the new system. When the project is handed over, the client's operators and system managers must have been adequately trained on how to run the operation before start up, not after. If not, the new system may be unjustly criticised for providing lower than expected performance. In a fast-changing world dominated by multi-channel distribution, future growth prospects should allow for expandability and flexibility through modularity to be designed in for both the mechanical layout and control systems.

If all the preparatory homework has been done diligently the automated warehouse should  deliver all that is expected of it, with one potential exception -- the vicissitudes of future demand that could change so much, for whatever reasons, as to reduce throughput levels to uneconomic rates. Perhaps one day Big Data will reduce that risk.  

-----------------------------------------------------------------------------