Getting the best out of what you’ve got

Advanced planning and scheduling (APS) systems – originally defined as capable of planning materials and capacity simultaneously and fast – have been around for years. Yet, to date, they’re still poorly understood and little used compared with MRP II (manufacturing resources planning) or integrated ERP (enterprise resource planning) systems that operate in the same functional space. Take-up has been increasing more recently, but among most would-be users there remains a combination of the fear factor and the usual resistance to change. For smart manufacturers, however, which recognise that the pressures they’re under have much to do with the fact that manufacturing business itself has changed – and that the IT they’re using isn’t up to it – that might be about to change. Because APS has the potential to deliver a great deal to manufacturing – and not just large organisations, but also SMEs (small to medium size enterprises) and indeed right out along the supply chain. So let’s demystify it: just as ERP can be thought of as a collection of co-operating business applications (sales order processing, scheduling, financials and the rest) delivering business benefit, so can APS. Until recently, however, it has been presented to manufacturers as high tech, sophisticated software, sporting novel and complex algorithms. Early systems were seen as tool sets requiring considerable work from expensive IT consultants and simulation specialists to produce a manufacturing business model that could then be used to seek improvements. In a sense it had to be: APS turned what had become ingrained standard practice (MRP) on its head, and factory mangers, it was thought, needed technical reassurance. But this approach may actually have served to slow its adoption. Had APS been sold as a way to get sites away from the old MRP ‘bucket’-based master scheduling (with its drawbacks of inflexibility and long range commitment), to high speed, agile materials re-planning and work rescheduling – with dynamic reallocation and the rest – history might have been different. Now, however, many of the main enterprise application vendors are offering, or are about to launch standard configurable APS suites. And in some cases their reach has spread beyond single site MRP replacement to include what looks like quite impressive supply chain planning and scheduling. APS has moved into package territory, the sales messages are louder and clearer, and it’s feeling more comfortable and looking good. Steve Banker, director of supply chain research at analyst ARC, is enthusiastic. “MRP was built many years ago and had to make some simplistic assumptions for it to work. Today we build more sophisticated models, containing advanced constraint information relating to how long processes take to set up and how work moves along the supply chain. There are a lot of different supply chain models available, but generally there’s a big payback, reduced investment and lower inventory carrying costs. In an ARC study about 18 months ago, we found that over 60% of users were achieving payback in less than one year.” There are three main types of supply chain, each with different constraint characteristics. Where the constraint is in manufacture, the biggest problem is throughput: semiconductors and metals production come to mind. Where the constraint is distribution, users will look at managing demand, distribution and transport planning. Examples include consumer packaged goods (CPG) and suppliers to supermarkets. Finally, where the constraint is sourcing, manufacturers will want collaborative forecasting and planning across different plants. “Implement the module that gives you the biggest return first,” says Banker. “We are seeing big pay-offs, and the main drivers at the moment are measurable ROI (return on investment) improvements, and improved customer service, though it’s hard to put a price tag on that sometimes.” But he adds: “With buying over the Internet, people want accurate promises.” Many businesses work on an ‘available-to-promise’ (ATP) basis. You have a plan of what you intend to make, plus your stock, so you know on a particular day what you have available. Put this against customer demand and you can consume stock, but customer orders might not be a part of your forecast. “‘Capable-to-promise’ (CTP) takes this one stage further,” says David Lofthouse, business support consultant at process enterprise software vendor Aspentech. “If stock for an order is not available, CTP has the ability to dynamically simulate additions to production demand. ERP gives you visibility of utilisation and capacity, but APS here shows impacts on raw materials and so on. For example, you may not be able to source some for manufacture – and APS provides the number crunching tools to simulate the options. It lets you to take decision-making to a higher level and make better decisions faster.” ‘Profit-to-promise’ (PTP) takes CTP and also weighs up the cost implications, looking at options for rescheduling onto different production lines or into different plants. It optimises the impact of varying workloads, again by simulation, and if it’s not clearly in the company’s interest to make something because of business implications, it can either say ‘no’ to the customer or suggest charging a premium price. But although, arguably, modern faster MRP done well and helped by finite capacity scheduling, real time factory integration and, for example, decent feasible flow manufacturing, can deliver most of this, APS is better at it. Because the fundamental attributes unique to APS are their high speed and sophistication (helped, of course, by today’s much higher compute power), which together allow manufacturers to deal on the one hand with the high velocity and changeable world of business, and on the other, the real world of manufacturing. As John Layden, charismatic vice president of APS development at US mid-market ERP software vendor Symix, puts it, they are designed to handle orders and potential orders “that constantly parachute into your manufacturing plan”. And not only do they perform the PTP simulations and decision support quickly; they carry that through to new manufacturing schedules and material purchasing that stay fluid and uncommitted right up to the wire. And today that includes web-based intimate contact with external supply networks. It’s scary, but it works. In a very real sense, says Layden, APS allows you to relax the old MRP rules – and let commitment slide much closer to days, rather than weeks hence. “There is a range of products across the industry working at MRP level and for forecasting,” says Andrew Dalziel, supply chain planning manager for ERP software vendor Intentia. “In Movex, the APS modules sit in the supply and execution segment of the product. An advanced production planner is used to refine production schedules, working at the operational and execution level daily or even several times a day. It creates execution schedules and work-to lists for different work centres.” He continues: “At a higher level, a supply chain planner works on optimising the external supply chain, modelling suppliers and customers. Using forecast demand, costs can be minimised or profits maximised. You can also optimise flow through the supply chain to get dynamic supply rules and a modified master supply schedule. A multi-site planner looks at the way orders are distributed and lets you see how different loadings affect the internal supply chain.” Sounds fair enough? Layden has a different view. He insists that getting a manufacturing site’s own APS right is the first essential step to getting the whole supply chain functioning. Beyond that, he says, you can’t hope directly to effect control; you have instead to think of a ‘supply network’ from which you need information like ATP and manufacturing commitment, fast. Hence, he says, Symix’s APS is an asynchronous model, not the more common ‘period planning’ APS (fixed period, variable quantity). And that, he says, makes its approach to supply chain scheduling so simple, pragmatic and, by implication, better than the others. Layden describes it as being about “fixed quantity, variable period”. He explains: “We get supply chain nodes to tell us what’s the minimum tolerable batch size they’ll make, taking into account pallet sizes, logistics, set-up times, whatever. Companies then specify what we call their ‘minQ’, the minimum batch size they’re comfortable with. Then at the ‘pull’ end of the process, if demand is for one item and the supplier’s minQ is 10, the supplier launches 10 and the system knows there are nine left. You get to 11 and he launches another 10.” And he explains that as demand grows you move up through two more levels – ‘multipleQ’ “where you might bring on another or several machines, cells or whatever,” and ‘maximumQ’. How does it work? “Our messaging workflow architecture supports it,” says Layden. “It all happens at the customer order level on the fly.” He says software resident at each of the supply (component manufacturer) nodes ($5,000 worth per node in the US) reacts to RFQs as described, and maintains ‘threading’ across the supply network, thus automating the APS flow of materials and build. In fact, it looks less like manufacturing planning than ad hoc, but robust and self-organising supply chain collaboration. “That’s what makes it make it so massively simple,” he says. Referring to MRP, ERP and what he considers the less advanced forms of APS, Layden goes back to first principles. “You can’t manage a business by looking at inventory, nor by load on machines. These are states – resultant outcome variables, dependent variables – they’re only useful after the fact.” He’s adamant: “You try and it backfires; you can only manage an independent variable that you have control over – and that’s time.” One manufacturer in the US that seems to prove him right is Rapid-Line, a sheet metal fabricator serving the industrial and general markets, manufacturing office furniture. The firm had been operating using an MRP system for ordering materials, but scheduling production based on supervisors’ experience. “We found that customers were starting to want their suppliers to hold inventory,” says Mark Lindquist, Rapid-Line president. “Our competitors went and bought warehouse space, but because of the nature of our product, which can rust and become obsolete, holding stock on inventory didn’t seem to be a good idea.” Instead, Rapid-Line decide it wanted to be able to react in 10 days – so it went to Symix for an APS, the aim being to optimise production and free-up capacity driven right from the front end of the selling process. “It’s quite easy to install APS, but it took six to nine months for the plant as a whole to get used to it,” says Lindquist. “Now everybody has to think in terms of time being critical for everything. We use modelling to find the best scenario and can hold orders back until they can be made in a given timeframe for a particular due date. “MRP was not time-sensitive, but now we optimise manufacturing resources as opposed to optimising time.” And he emphasises the point: “The thing you’ve got [control of] is time and the schedule shows how to use it. We can also give a reliable delivery date when the order is taken.” It’s been a huge success: “We’ve started to achieve our ROI (return on investment) targets, and the time for turning round work has gone from 25 days to eight. On-time delivery is up from around 89% to 99%; the order queue is down; and customer lead time has halved. Plant management has shortened its planning horizons to one, two or three days; resources have been optimised; and production problems can be dealt with in advance.” And he concludes: “It’s a question of loading the plant as close to 100% as possible.” Symix UK’s new managing director Adrian McNay sums it up: “APS looks complex, but that’s because it’s been targeted at the techies.” He accepts that Symix, like others in the APS frame, has got to get the market to understand that it’s just another, albeit powerful, manufacturing business tool – and that actually, it’s easier than MRP II. The point is, good APS not only lets you see the consequences of scenarios fast through simulation – it also lets you make better decisions and enact them to stay on top. In short, in today’s world of speed, change and complex supply networks, APS software gives you a better chance to satisfy customer demand, make a profit and improve your, and your partners’, business. Don’t let the salesmen at CIM 2000 blind you with science and maths. Insist on understanding what the business benefits to you would be and how they would help you implement to get them.