So how good is your planning?

Was the time when IT vendors that develop software for factory planning would argue the toss over which was the best technology for the job. Should we use APS (advanced planning and scheduling) systems and their earlier finite capacity scheduling (FCS) counterparts, or good old MRPII? It always was an odd argument. MRP's functionality was, after all, directed mainly at material planning and ordering, assuming infinite capacity. Subsequently, many MRPII systems got some capacity planning additions and, more recently, a few also moved up to lean, TOC (Theory of Constraints) and similar 'pull' and 'flow' based improvements to help bring planning more in line with demand-driven manufacturing initiatives. APS/FCS, on the other hand, was about decision support for short time-horizon shop floor scheduling and optimisation, with the primary goal of ensuring cost-efficient, on-time, in-full (OTIF) delivery of customer orders. In recent years, that too has moved up in sophistication - for example, taking on the management of complex combinations of BoMs (bills of materials) and operations with multiple, floating resources and the rest. And APSs' scope for reflecting supply chain interaction has also increased, enabling longer range detailed 'what if' planning, and expanding its goals to include supply chain responsiveness improvement, with customer promising on-the-fly, but also inventory reduction and capex optimisation. So, MRP and APS are a bit 'chalk and cheese', and in hindsight, arguing about which was best may have served mostly to complicate the issue. But it happened because at the time - and still today - too many manufacturers seemed to have their heads in the sand over MRP's limitations. And the fact remains that most older implementations (and that's the vast majority) still suffer from infrequent MRP runs and the inaccuracies of batch updates and assumptions, which have little place in today's short-term, very demanding world. They're also not equipped to respond to the issues of highly variable product mixes and quantities that are also subject to late change. Most MRP systems are just not very good at taking adequate account of details like pegged materials, subassemblies, subcontract work, bottleneck resource loadings, labour and skill sets. APS engines can, MRP mostly can't. Hence in MRP-run shops, safety buffers and WIP (work in progress) mushroom, confusion reigns, customers are disappointed and costs, efficiencies and the potential to develop the business suffer. Which is why, perhaps the APS/MRP debate still rumbles on. We're missing out The bottom line though, is that many would-be manufacturing users have missed out on the benefits that APS can bring, way beyond improving schedule adherence and OTIF metrics - like transforming the efficiency, effectiveness, flexibility and cost of their operations in terms of WIP, materials and finished goods inventory, lead times and workflows right down to work centres, machines and people. They've also missed out on invaluable IT support on the drive for lean, so much of which depends on getting good data and a better understanding of how the factory and its processes and routes actually work. And that's a shame, not least because ironically, many happy users of factory schedulers argue that it needn't be a case of either/or. They demonstrate that APS systems can complement existing MRPII systems - and often for an incredibly small investment in IT and skills development (although the factory and business culture gets a bit of a kicking, which, looking at some operations, is no bad thing!). For example, when used in tandem with modern (and low cost) shop floor data capture (SFDC) systems, it's possible not only to get excellent visibility into the 'black hole' of the factory and its WIP, cycle times, variances and the rest - but to cement in and capitalise on improvements. Real-world factory information and 'what-if' APS simulation can be a powerful combination for good proactive scheduling in the future. Unsurpassed visibility They can also deal with the problems of product mixes. As Tony Anthony, professional services manager at manufacturing ERP conglomerate Infor, says: "The beauty of APS systems is the visibility they give you behind the scenes, for example, of your bottlenecks. You don't have to be super sophisticated, but with the model in place, as orders come in you can go into detail and see why a projected delivery date may be later than the customer wants. Then you take a view on what you want to do. But the point is you can do something about it, rather than just accepting the order and then fire fighting." Paul Williams, sales director at ERP firm K3, makes another point, citing process industry and craft manufacturers like Dartington Crystal, which will have issues like by-products, co-products and off-spec production. "They may not be very predictable but when they come those products may need some different operations and processes - and there's still a requirement for the original production order." The role for a real-time APS system there is to schedule execution for the existing unfinished products, and to raise another works order for the original products and to schedule that, consolidating both where optimal. One excellent example of an APS-led transformation is at £100m turnover, Hull-based holiday homes builder Willerby. Its top line benefits: five years after starting a phased Preactor APS implementation, WIP has been cut by a massive 80%, production lead times are down from six weeks to five days, subcontract work has ceased and three shifts have been compressed into two. Looking around now shows a picture of efficiency and lean thinking accomplished across disparate shops throughout the business. It certainly hasn't always been like that. With holiday homes leaving three assembly lines every 30 minutes, 150,000 machined parts, and different assemblies and products having different cycle times, bottlenecks used to be hard to identify. So rework and missing parts were common and eleventh hour subcontracting of components was also frequent. Missing items and damaged stock due to huge inventory levels caused constant rework and fire fighting, leaving management little time to think about solving problems. Craig Dunn, lean operations manager, explains: "The problem was that the business was very disjointed, with areas all managing their output and capacity independently. There was no regard for previous and subsequent processes, and material was bought in by keying dates into MRP from the assembly line schedule generated by Microsoft Access. Each week a new schedule was created allowing for slippage, and each department would re-plan their area [but] areas achieving planned targets would simply carry on." So the cycle of overproduction was perpetuated, with material and subassemblies eventually consuming all available space. Clearly, batch and queue methods combined with disorganised and inflexible working that couldn't or wouldn't take account of fast cycle time products biting at the heels of slower ones, and vice versa, were no recipe for success. So in 2000, production director Colin Jeffrey made the decision to go for a lean initiative - but driven b y APS. "We were unable to start a lean implementation programme without first understanding and gaining control of our processes, and Preactor offered the solution," he explains. Late in 2000, Willerby made a start by focusing on improving the accuracy of the assembly line schedules and, following development by system integrator RMS around changeover anomalies, improved overall schedule accuracy from 80% to 95%. RMS also linked the Preactor plant model with Willerby's MRP, eliminating the need for weekly data re-keying and also stabilising the supply chain. With the assembly lines done, the company moved onto its supply areas, beginning with the machine shop. With little formal data, model building there depended on extracting routing information and operation times from shop floor personnel, but with that done, Preactor simulations quickly revealed the hitherto hidden bottlenecks. Says Dunn: "Building 'what if' scenarios in Preactor allowed us to create flow by re-routing products and moving machinery... WIP was reduced to sensible levels and efficiency was maximised. Subcontracting ceased and the three-tier shift pattern in this area was reduced to two. Lead times were also slashed and the MRP system was brought in line, reducing the amount of bought-in stock." It snowballed And one thing then led to another: with better schedule accuracy and control of the machining processes, issues on the shop floor became clearer. Visibility improved and the number of missing and damaged items dropped; then future constraint processes could be more easily identified by Preactor and contingencies discussed months in advance. There remained, however, the problem of the jig shop. This section hand-made the sidewalls and bulkheads for the main construction of each holiday home, and was constantly starving the assembly processes due to a lack of sensible scheduling, balanced workflow and even communication. But again, once Preactor was implemented and schedules and work-to lists became a regular feature, the improvement was huge. Adding SFDC with touch screens for logging start and finish of jobs, and linking that to Preactor's visual planning board completed the feedback for management and planners alike. Says Dunn: "Preactor has provided us with the means to start applying lean philosophies to the business. Although still providing scheduled work to some areas, it's now being re-configured to produce capacity planning for the management team in some supply areas." What a difference.