Could 80% of us be best served by finite scheduling?

It’s astonishing that with so many in manufacturing having potentially so much to gain, only some 22% of sites in the UK say they have some form of electronic finite capacity scheduling (FCS), according to manufacturing market research specialist Benchmark’s latest figures. Even that is probably far higher than reality. Benchmark’s research manager Paul Watts warns that its Computers in Manufacturing (CIM) survey was not geared to further qualifying the responses. Was it really finite scheduling or just ‘computer-based’? Was it switched on and actively used? How much of the site did it cover? Nor did this Benchmark report set out to test success, the specific IT, manufacturing environment, whether it was integrated... But in the US, Production and Inventory Management Journal published APICS research in 2000 (APICS Vol 41 Number 1, ‘Computer-based scheduling in manufacturing firms: some indicators of successful practice’) which did, and its conclusions are really quite revealing. Not only did it find figures strikingly similar to Benchmark’s (although of the 25% with FCSs, only 6% were using it for the entire facility), but it also investigated what benefits users were looking for, and how well their systems scored. Then, most significantly, it filtered against those actually using FCS. And then it re-filtered against scheduling practices – integration of the scheduling systems with other factory and business systems, and frequency of schedule regeneration. It found that by far the best hard benefits went to those companies that used FCS across their entire facilities, that had integrated it with their other key systems and that were regenerating once or more times a day. Accepting that FCS was unlikely to be their only initiative, nevertheless the difference for these was consistently significant. I’m generalising, but on-time delivery was 20% better, schedule reality 30% improved and responsiveness up 45%! Why labour the point? Because with all the hype over big ticket supply chain applications and the like, FCS has been relegated way down by most IT marketers. And there are analysts and consultants who either condemn FCS famously as “the last resort of a company whose processes are tortuous, ill-considered and inflexible – an IT Band-Aid applied to a festering wound”, or simply foster the notion that it’s hard to provide solid ROI (return on investment) and that, at best, users can expect “incremental improvement”. Yet the experience of those that have done it – and done it using what turn out to be fairly obvious best practices – absolutely flies in the face of these assertions. And when you dig a little deeper and find that for up to 80% of sites FCS could make a substantial contribution, it’s clear that manufacturers need to take a new look. Because this needn’t cost you an arm and a leg (perhaps that’s why it’s being down-played) and you can expect great gains in anything from production control and efficiency to customer service. So lets, once and for all, uncover the facts about FCS and its near relation APS, and best practice, and the kinds of organisations that are likely or unlikely to benefit (see panel). Ron Bligh is ERP (enterprise software) vendor Mapics’ resident FCS and APS specialist. He notes that systems fall into three categories – backward schedulers which take the order book and work back to deliver on time in full; forward schedulers (the majority today) which consider only efficient resource utilisation; and APSs which can look both ways, also scheduling materials and capacity concurrently. Ed Stubbs, pre-sales consultant with enterprise software vendor JD Edwards, explains: “If you think of MRP II, which is infinite backward planning, finite backward scheduling is just-in-time backward scheduling. So it’s geared to the due date, and leaves the factory with some idle time which you’re happy with because you’re not committing WIP (work in progress), increasing inventory and so on. Forward scheduling leaves no idle time – it’s geared to machine utilisation. In the real world you need a mix and match of these – backwards to at least meet your orders; then forwards to fill in any gaps, do a bit of demand-pull forward, or whatever.” We’re talking value here And that’s what the more modern FCSs and the APSs do – and fast because, as PC-based FCS/APS/production control software vendor Preactor’s managing director Mike Novels says, “they answer in memory the question ‘when can we ship’ based on current stocks and current capacity using material allocation rules.” Must you replace the production planning (MRP II) part of your ERP? If its straight FCS, certainly not; if it’s true APS, it’s up to you. The likes of Mapics, Profax (Asprova), Lilly and Manugistics (with STG) would certainly encourage you to do so. And if you’re brave, anecdotal evidence suggests strongly that the benefits are excellent, not least because you remove a suite of software that never was geared to today’s complex, tight lead time manufacturing realities. However, according to Novels, you don’t have to. “Our approach is to build on what ERP already has. The first level of that is to do what [supply chain software firm] i2 calls ‘capacitated ERP’ whereby the sales and works orders generated by the ERP BoM explosion (under MRP II) are then scheduled against a finite model… We see this as enabling ERP rather than replacing it, and this obviously has implications for both the software spend as well as services… Rightly or wrongly, most prefer the ‘comfort zone’ of the latter approach, but either way, Novels is onto another key practical point. As Bligh says: “If you’ve got piece parts, subassemblies and assemblies and you try to plan production of all that conventionally using MRP II, you’d end up with due dates for subassemblies coming up earlier than the piece parts you need for them – because there’s no link between them.” And it’s revealing and eliminating these kinds of conflicts that makes FCSs so special. Says Stubbs: “It’s usefulness and contribution to ROI come from improved decision-making; improved demand satisfaction (due date); prevention of inventory violations (raw, WIP and finished goods); and accounting up-front for capacity constraints – labour, skills, machines, routings, optimal batch sizes and sequences, set-up and changeover times, tooling and so on.” And he could also have added order promising, and manufacturing analysis and improvement. John Wroot, logistics and operations manager at £50 million turnover Birkbys Plastics, which makes plastic injection mouldings for the automotive industry, is implementing Profax’s Asprova APS on top of its existing QAD Mfg/Pro ERP, and finding all of this. “We’re looking for £250,000 savings per annum – a 4.7% reduction in inventory value, and a 1% improvement in labour efficiency. On top of that we’re expecting lead times to drop from 17 days to nine.” And that’s on a factory that employs 500 people, with about 70 plastic injection moulders, spraying lines, welding and assembly lines, producing around 780 finished items with 790 resources (580 tools). And it cost £60,000. Asprova provides a sensible and dynamic sequence plan, he says. “Suppose one of the jobs isn’t required till seven days out, but you’re running a similar job involving the same tool on a machine at five days out. It will let you make the decision to run the later job early, trading off set-up time against temporarily higher WIP. And you can blow it back into MRP, so purchasing, for the first time, can see when it will be manufactured, not just the standard lead times. “Also, instead of being reactive as we have been, we’ll be able to be proactive.” And he cites injection moulding: “Scheduling wasn’t bad, but this will do it in seconds and allow us to sit back and understand how we can make improvements. That’s a major benefit.” In this sense Wroot sees his FCS system as an analytical tool. “It’s about learning to become more slick in our manufacturing – having the time and the tools to streamline our processes. For example, you can see when bought-out components are coming in, so you’re scheduling only achievable jobs. So it’s taking cost out, but it’s also going to be an active tool to help us become more lean.” Other benefits he lists include being able to track inventory levels for all SKUs, minimise WIP by seeing when materials need to be released to the shopfloor and when they need to be purchased; being able to switch on material and resource constraints and embed appropriate rules. The result, he insists, is that you always end up with one factory-wide achievable and optimised – commercially and practically – plan. He’s not alone. John Catton, materials and IT manager at printing machine manufacturer Technigraph’s Thetford machine assembly site, who implemented Preactor two and a half years ago, says: “Preactor paid for itself in a matter of hours.” His system cost the princely sum of £2,000 including training! “It’s a very good, very simple, very cheap system,” says Catton. And he adds: “there’s loads of potential for taking it further.” Currently in the process of closure as the business moves to Halifax, the site last year turned over £12 million, and he claims Preactor made an instant and “massive difference” specifically to production agility. And the benefits, although not expressed in financial terms, are very clear. On order promising, for example, Catton says it’s simply a matter of loading the telephoned order parameters straight into Preactor, and the system “generates realistic delivery dates within minutes.” But what about the requirements? Best practice? And here, perhaps, is a bit of a clue to poor FCS uptake? Because top of the list are: very high data accuracy (BoM, routing and inventory), a manufacturing culture that respects the schedule, integration and the generation of a plant model to drive the system. All can be big deals, although model building need not be: the information is almost always on a system somewhere (worst case, in the production managers’ heads), and FCSs are getting better at helping. As Stubbs says, “With the newer systems’ adaptive solvers all you do is define the environment (machines, materials, costs, BoMs, routings, resources, etc) – and they sort out the bottlenecks.” But there’s no escaping the rest. JDE’s Stubbs insists: “BoM and inventory accuracy have got to be 99%-plus, and if you don’t do that right it will kill any scheduling project.” And he adds that FCS has to be tightly integrated with ERP and shopfloor data collection (as borne out by the APICS study). Stubbs also encourages users to ensure that their schedulers are updated as frequently as possible – even to the extent of being event-driven from ERP. “Remember, we’re trying to cut down on cycle times here. The ability to re-sequence production quicly is critical to enabling take-up of that opportunity.” Finally on culture, if you can’t get this right you will fail. Bligh comments: “If operators make off-schedule decisions at one of the work centres, or even fail to move a job on, the next work centre’s finite schedule is instantly rubbish. But if you can break down the old ways you can get the real benefits of on-time deliveries, good resource utilisation and factory flexibility – the very attributes British manufacturing is crying out for now.” FCS - Who needs it? Finite capacity schedulers (FCSs) serve all kinds of plants and factories, across the size range, in most industry sectors, and with most manufacturing operations. Indeed, a straw poll of software vendors, consultants and users rapidly results in the conclusion that it’s easier to define where FCS hasn’t got something to offer than where it has. JD Edwards’ Ed Stubbs believes that FCS has very little to offer only if your plant has lots of capacity, declining market, fairly recent investments in flexible machining plant, few constraints and predictable demand – “anywhere you can be reactive enough by rescheduling on the back of a fag packet”. And he believes the same applies to large companies where plant is dedicated to one or just a group of SKUs (stock keeping units) – “although mass customisation may be catching up with them”. He also notes that plants have to stick with their rules: “if they can’t do that, don’t use finite schedulers.” Mike Novels, managing director of Preactor, attempts the positive. “Businesses that benefit from FCS have one or more of the following attributes: planning is carried out by a single long-standing employee who does it in his or her head’; companies that currently schedule by the ‘he who shouts loudest’ rule; companies with high raw, WIP and finished goods inventory; companies that struggle to give realistic delivery dates; companies who need more agility to respond to short lead time orders; companies where individual departments plan their work independently and without reference to other parts of the company.” Sound familiar? Operationally, he says, these firms tend to have: “complex or alternative process routes for their products; significant impact on productivity due to the sequencing of batches; variable demand, product mix and/or resources available, including staff; variable yields and/or process times; multiple resources required for tasks that are product and/or task specific.” Mapics’ Bligh looks at the extreme of manual PCB assembly work, where there might be no sub-assemblies and no material or resource issues, no bottlenecks and capacity (people) is flexible. “Then you wouldn’t need finite scheduling.” But for the rest – even what might at first seem relatively simple factories – FCS can add value. Bligh: “I visited a company in Germany that turned out to be a prime candidate. They were making spanners and it was just 18 operations on each spanner – no sub-assemblies, no final assembly: just straight from raw material to final product. But they had hundreds of machines of different sizes and they did all the forging – and sent some out for plating. “Finite scheduling allowed them to handle the variables they couldn’t deal with. Different spanners, different machines, forecasts… too many machines and too many jobs waiting. Without finite scheduling, they had problems with economic order quantities, over-running jobs, machine breakdowns, long set-up times, large batches and the rest… And if your forecasting is poor, everything gets done in a rush, and that means tearing down jobs non-optimally – and then where’s your efficiency?” The bottom line: some 80% of manufacturers would benefit from FCS, most of them significantly – presupposing they did it right. Says Stubbs: “You can think of small and complex factories at one extreme, and large, simple ones at the other. At the small complex end it’s about trying to get the best out of a fundamentally constrained environment. In the large, simple operations it’s about helping to formalise scheduling – help them move beyond spreadsheets and planning boards – and provide a real world communication and visualisation tool.”