There is little doubt about it, investing in warehouse automation is a risky business. Committing substantial capital sums to a project requires complete confidence in the forward strategy of the company, a full understanding of present and future supply chain requirements, and a detailed knowledge of the product profiles being handled.
Get it wrong and the consequences can be serious, both for the financial well being of the company and for your own career prospects. But, get it right and fortune will smile upon you, brining far lower logistics costs per unit sold, fewer headaches over labour resources, greater efficiency in controlling the flow of orders, increased capability, and a far higher utilisation of expensive land. The result: higher margins and greater customer satisfaction.
The stakes may be high, but for those enterprises prepared to embrace automation and intelligently apply the technology available in an appropriate, measured and well-planned manner, the rewards can be substantial. It’s a question of risk and reward – the answer lies in a balanced approach.
The big questions
So, what drives companies to consider automation? Where is it appropriate? How do you plan for it? And what of the old reoccurring question of flexibility?
Many successful companies that have been in business for a great number of years find themselves faced with an eclectic network of disparate warehouses arrived at, almost by accident, through a history of mergers and acquisitions. Typically, due to their incidental provenance, such operations are inherently inefficient and difficult to manage. Rationalisation of such a structure can often justify the building of a new automated centralised facility, funded by the sale of older warehouses and the resulting consolidation and reduction of inventory holdings. Here greater stock visibility and product velocity can be very welcome gains from such an exercise.
The supply chain is a powerful weapon on the high street and in the fiercely competitive grocery sector market share can be easily won or lost through the availability or not of goods on the shelf, condition of produce and price.
A highly visible example of where automation is being wholeheartedly embraced to gain that vital edge is in the UK with grocery retailing giant, Sainsbury’s bold $1.5bn supply chain initiative where a series of 60,400 sq m highly automated ‘fulfilment factories’ are being built over a three-year period. Now nearly two years into the project, the first fulfilment factory in Hams Hall – operated under a threeyear open-book agreement by Tibbett & Britten – has come on line.
The level of automation in the fulfilment factories has raised some eyebrows elsewhere in the industry but, as Andy Banks, Sainsbury’s supply chain development director, told Logistics Europe early last summer, ‘I think in most cases, we’re taking proven elements and pulling them together to produce a unique solution – we plan to be leading-edge, not bleeding edge.’
Hams Hall will operate at ambient and 0-5 degrees Centigrade and will serve 73 super stores and up to 100 smaller ‘local’ stores in the Midlands area.
Siemens Dematic has been awarded two major contracts for automated materials handling systems from Sainsbury’s for fulfilment factories in the Midlands, at Hams Hall, and at Waltham Point in north London. The two systems are virtually identical and are designed to handle 2.4 million cases every week. The system includes a sophisticated highspeed cross-belt sorter, 2.25 miles of belt conveyor and twenty automated storage and retrieval cranes. Witron was commissioned for the systems integration for the slow moving facilities in north London and Stoke.
Cost per pallet
Illustrating how automation can deliver that competitive edge, John Harvey CBE, chairman of Tibbett & Britten, told Logistics Europe in a recent interview: ‘If Sainsbury’s get cost per pallet down to the level they want they will set new levels for the industry.’
Of course, the level of automation and the particular technology applied to an application are of vital importance. ‘There’s a huge spectrum of choice, which depends on the nature of the task’, explains Alan Braithwaite, executive chairman of LCP Consulting. ‘You need to divide the warehouse operation up into: technology around receiving, technology around the bulk store, and technology around the picking operation’.
Automated storage and retrieval systems (ASRS) are applied to bulk pallet stores but ‘its justification is less about labour saving as it is about getting maximum use of height and floor space’, says Braithwaite. This is particularly pertinent to the UK market where land is so expensive. In the US cheaper land prices make the logic a little different. According to Steve Knights, business development manager at Siemens Dematic, ‘In the US you can have land for free, given by local authorities, as long as you create jobs.Consequently, the US market is far less sophisticated than in Europe. They have vast projects, but it’s not “how high do I go?” it’s, “how much conveyor do I need to cover this acreage?” In this environment you see high utilisation of conveyors and pick-to-light systems to enhance order-picking processes.
Knights goes on to point out that, ‘The market for stacker cranes in the US is about the same as in the UK’.
Generally warehouse automation can be justified given one or more of about a dozen circumstances: the warehouse is large; stock volumes are high; many lines are carried; throughputs are fast and high volume; flow rates reach extreme peaks; many outlets require frequent consignments; picking schedules are intricate; maintenance of stock levels is critical; dwell times are short; extreme storage environments are required; interest rates are low and stable; there is three shift working.
But, one very important point when considering automation, says Simon Tomlinson, director of supply chain consultancy The Logistics Business, is that ‘you should take care that you do not automate a bad process.’ This is why a thorough analysis of the processes involved and careful planning are essential. According to Braithwaite, ‘If an automated system fails early it’s usually due to the business specification not being right, say the stock profile was wrong. And problems later on can be caused by the stock profile changing’. To a great extent the success of a project depends on the vision of those giving you the input data.
This may account for the frequent accusations of ‘inflexibility’ levelled against automated systems. Could it be that not enough consideration is given to the planning process? Or perhaps, a project is starved of the oxygen of reliable and accurate information?
Robin Vega, director at simulation and modelling specialists, Cirrus Logistics, explains the problem: ‘All too often a spec is sent out to the equipment suppliers as to exactly what the automated system is expected to achieve and under what criteria it will be expected to operate – for example re-stocking at night and what per cent completion of orders will be acceptable for dispatch. Frequently the customer’s expectations and requirements change during the build process, but unfortunately the equipment has been specified and bought to achieve the original criteria most effectively. The customers are then left with an inflexible system that cannot meet their changed requirements and therefore the throughput envisaged is not achieved.’
It’s a sorry tale, but there are tools available to help smooth the planning process and calm nerves – that is, of course, depending the data used is accurate and, as far as possible, future proof.
Simulation allows the actual operating scenario of the automation equipment and the vital operating systems to be built up into a simulation model, so as to mimic the real situation. This allows customers to test output, flexibility, and the software and control set-up before a financial commitment to the project takes place.
Actual staff numbers can be predicted and the level of flexibility to deal with change can be tested – such as increased number of skus or type of order mix. Vega says, ‘Through the use of a simulation to carry out sensitivity analysis the operating boundaries of each piece of kit can be defined and therefore the effect in meeting changing KPIs can be determined.’
According to Vega, algorithms and software control systems are ‘actually more important to efficient automation than the hardware itself – if not configured properly, however good the equipment is, the throughput will not be achieved.’ But importantly, he adds: ‘In the commissioning stage, which is normally carried out under great time pressure, simulation allows up to 80 per cent of the software to be configured in advance enabling the equipment to achieve the agreed throughputs and thus reducing commissioning time significantly.’
‘You should build in flexibility at a cost at the beginning,’ believes Braithwaite. ‘If you’ve got an aisle with 3000 pallet locations in it and an ASRS system, the one thing you can’t change is the speed of the crane. So for a given throughput you will only be able to increase accesses, and you run out of time to do the work in the aisle – thus reducing the utilisation of the storage cube. It’s a double jeopardy. So we look at where the design flexibility needs to be put in and this is all balanced against risk on the business forecast.’
One of the biggest reasons for an expected growth in the use of automation is the cost and availability of labour – this is a major reason for the prevalence of automated warehouses in Germany. Companies will invest to take people out, but the key to success will be appropriate automation which may involve pockets of automation combined with the flexibility of conventional full pallet systems using trucks.
Increasing picking productivity
Order picking is a particularly labour intensive process and as Knights says, ‘Nobody has come up with a better picker than the human being,’ so the trick is to automate processes that increase, as far as possible, the productivity of the picker. Invariably, that means bringing goods to the man. Here automated mini load cranes, intelligent conveyor systems and carousels are preferred methods of achieving this aim. Other technologies such as pick-to-light and wearable voice computer picking systems, such as Vocollect’s, are increasingly being employed to enhance efficiency.
Steve Richmond, general manager, projects division Jungheinrich UK, sees a rosy future for mini load crane systems: ‘We are experiencing an increased demand in the area of mini load cranes for handling cartons and totes. In the right application, high performance systems can deliver products to picking and packing stations at exceptionally high speeds, out performing many conventional, alternative systems. I expect to see the demand for this type of system to increase significantly over the next few years.’
The big integration challenge
But as Knights, Braithwaite and others concur; it’s the integration issue which is the big challenge. ‘An ERP system will have taken an order and will typically feed it to a warehouse management system, but is it doing it in real-time or batching it up? Is the WMS driving the automation directly or through a sub set system? There are some big choices here. It really depends on the type of warehouse management system you have – some run automation really well others not,’ points out Braithwaite.
‘Warehouse software packages do not necessarily operate in real-time and for automation this is essential as the system has to react quickly,’ says Knights.
So what about pay back? How long a period of time are we talking about? Of course, it really depends on the dynamics of the particular application and indeed, the scale. Projects in Germany can reach the giddy heights of 45m and are invariably rack clad structures, whereas in the UK the maximum that has been permitted is 35m. Anything over 30m tends to be rack clad and the build process is much faster, but in the UK there is a reluctance to adopt this approach as it is perceived to be inflexible – the pension funds really don’t like it. However, there are advantages. For one, rack clad structures have a rateable value of zero as they are classed as plant and equipment.
According to Knights, ‘Normal payback is 3-5 years on most of the projects we do, some go beyond that.’ But John Harvey Chairman of contract logistics service provider, Tibbett & Britten looks to 7-10 year contracts for automation to be considered feasible.
In general, contract lengths for third party operators tend to be 3-5 years in Europe which would preclude investment in automated warehouse facilities. The dominance of contracted out logistics in the UK may go some way to explaining the differential between adoption rates of automated warehousing in the UK and Germany.
However, if UK businesses wish to trim logistics costs to gain a competitive edge then automation may well be a significant tool to achieving this. It may require a broader view on the lengths of 3pl contracts.
The ‘Ten Commandments’ of automation
- Despite what is said, automation can be flexible if flexibility is designed in from the start. This begins with plenty of ‘what-if’ analysis or scenario planning. Very few companies do enough.
- Early operational buy-in; failure to gain early involvement of the operations team is one of the biggest causes of problems.
- Own the solution – too many customers rely on the supplier to do all the development and never really take ownership of the solution.
- Remember that a project does not finish at the end of the implementation phase. After implementation comes exploitation, which is when the real benefits of the investment are realised.
- Undertake incremental cost benefit analysis – don’t just look at the benefits of the whole system, but of each element as well. You may find that some elements do not offer a benefit.
- Take a long view. If you are seriously looking for just a two or three year payback then automation is unlikely to work.
- Remember that automation is not just about cost saving its also about building capability. There are things that can be done with automation that are impractical conventionally.
- Many automation projects these days involve high-speed order picking. There are two critical processes which you have to get right if the project is to be a success – the order launch process and pick face and product profiling.
- Try to give quality time to the design, planning and testing phases. Most projects run to tight timescales and these most important phases are usually the first to be squeezed when programmes are compressed.
- Recognise that managing an automated warehouse is completely different to a manual one. Many companies fail to recognise this and have major difficulties in the early months or even years.
Source: Simon Tomlinson, Director of supply chain consultancy The Logistics Business