Much more to it than a sausage factory

Pork pie and sausages producer George Adams is well on the way to reaping the benefits of a top to bottom ERP implementation Key benefits One-off de-stocking saving of £300,000 to £400,000 on raw materials by improving accuracy of forecasts and enabling better production planning and process control Greatly reduced waste through tighter, more dynamic production planning and accurate WIP tracking Automated electronic audit trails and variance analysis to improve yield and reduce waste further Strong foundations for variance management and continuous improvement Further development across six plants will mean planning, production and financial reporting all share a common database, providing a single integrated view of the whole business No more manual counting, with everything being swiped in and out Better reporting for business management and further efficiencies from real time, instead of shift- or batch-end reporting Max three year payback – but also essential to support rapid business is growth Greatly reduced inventory and waste through tighter, more dynamic production planning and accurate work in progress (WIP) tracking; automated electronic audit trails; and strong foundations for variance management and continuous improvement. Those are among key achievements at £35m Ruskington, Lincolnshire-based pie and sausage producer George Adams Pork Products, from its ERP and wireless RF shop floor data collection (SFDC) implementation. Says Mark Wood, operations director: “Conservatively, we’re looking at a three year payback, but the other way of looking at this is the business is growing rapidly and we needed better systems.” So good is it looking that the firm intends to roll out its system – SSI’s Tropos ERP – as one instance across six plants, so that planning, production and financial reporting all share a common database, providing a single integrated view of the whole business. Pork pie and sausage production have more in common with general manufacturing than you might think – the main differences being fresh materials and products, and some semi-continuous processes. The Ruskington plant – which is 15,000sqm with 600 employees producing around 1m pork pies a week, 500,000 Cornish pasties and 150 tonnes of sausages – doe the whole bit. Pork pie production starts with butchery, then raw pie manufacture, cooking in ovens, blast chilling, packing and despatch. Cornish pasties are similar, while sausage production, although starting at the same front end, involves large scale mechanised mincing and chopping, then packing for fresh sausages, but cooking in 12 metre fryers followed by multi-head machine packing and chilling for pre-cooked sausages. There are controls on modified gas atmospheres, cooking, intermediate and final storage, and chilling throughout. The bottom line: it’s a complex production planning and management problem involving multiple storage locations as semi-finished and finished goods, with important shelf-life issues – further complicated by very short lead times from the big retail chains, certainly less than the make times. “So we’re constantly having to run to forecast and stock,” says Wood, “and with notoriously inaccurate forecasts and daily call-offs from our customers.” And that’s another issue: there are considerable variances between the supermarket chains’ forecasts and their orders. One of George Adams’ top customers can’t achieve better than 30% accuracy with week-ahead forecasts because of the vagaries of consumer demand due to short-term factors like the weather on demand for pies. Looking briefly at the before scenario, George Adams had evolved several spreadsheet-based systems for forecasting, materials and labour planning, variance management, stock control, efficiency management and monitoring asset utilisation, but data collection and representation were inconsistent. “None of the systems were integrated and they were all separately and manually maintained,” says Wood. “It was becoming so complex that we had to keep upgrading PCs to cope. We built a temporary system at Spalding with an Access database, but that hadn’t got the level of mixed production and complexity we have here… Ultimately, we knew we needed to capture data right through from stores to despatch and have it in real time so we could control the stock better.” There were also other issues: it was difficult to obtain accurate yield figures and so on for costing, planning and financial management. “The spreadsheets could do it, but we’re in growth mode and it was becoming more difficult. Also, we couldn’t work back from the spreadsheets into factory production: we couldn’t see the detail of product and stock locations.” Meanwhile, on the ground, he says: “The result was that all the departments were working in ‘silos’. They knew the plan, but each department decided what it was going to run and when, to optimise their local efficiencies. But that wasn’t necessarily what the next department needed at that particular time. We were fulfilling the overall orders for the day, but we were carrying far too much stock to do so. So we decided to re-look the whole thing.” Hence Tropos for overall financial management, forecasting, procurement and integrated stock and also production control, using barcode labels to track batches in detail through the process with touch screen PCs and handheld scanners to gather data from the shop floor via a new wireless RF network. Says Wood: “We had installed Tropos about six years earlier as a purchasing system, but it hadn’t been rolled out any further. This time we looked at other systems but decided that as we had already paid for Tropos, we might as well make it work even harder. Looking at top line benefits, Wood says: “Now we know exactly how much we’ve got of all the materials and ingredients, the products, their status and where they are – and we can plan and re-plan against that. There’s no more second-guessing the stores all the time.” Also, with actual usage via the barcodes, not only are their clear advantages for the essential food industry audit trails from consumer to pig – all automated and with facilities to make that paperless – but valuable analytics. “It means we can look at variances and see what’s happening in detail,” says Wood. There’s more. At a practical level, he says: “Because you can pre-program factory disciplines, the system won’t allow you to use the wrong ingredients. In the past, that’s happened occasionally and it would be picked up later in the process but too late to do anything about it, so the product was rejected. Now we can prevent errors – or at least pick them up a lot sooner.” And he refers to the power of barcode tracking: “All products have a use-by date. If stock is at the back of a chiller with 250 racks it’s difficult to see, but Tropos will know it’s there and it will make sure products that are going out of life first are selected.” Moving on to the front end, Tropos’ forecasting capabilities are used to the full to get away from the old ‘rock and a hard place’ problem of being shot if the firm fails to deliver, but throwing away thousands of pounds’ worth of stock if it over-produces. “We can use the combined forecast and do final planning on the day to meet the changing real demand for the various products as we get real orders,” says Wood. In fact Tropos production scheduling is run twice a day – at 9.00am and 4.00pm – on a rolling 24 hour basis, re-cutting against what’s in the stores at the various stages each time – and automatically taking account of shelf-life for coded material items. Then at the plant level, batch processes are now run effectively to plan: George Adams hasn’t moved to scheduling per machine because the floor is too dynamic and fast. “It’s up to the guys on the shop floor to use the system to make decisions,” says Wood. “They’ll know if a machine is not working or Jim’s not in, and work around that.” Some would argue that machine scheduling and planned recovery would be a good way forward, but he makes the point that the numbers are big and the processing times fast (for some products just 15 minutes). Also, some of the production is semi-continuous, with runs of 400,000 pies, while other production is for just 50-off, with some hand-made. So detailed scheduling would be a classic case of diminishing returns. Returning to the wireless network and scanning, Ruskington is one of the first to go all the way with this in food manufacturing where the advantages are not only cost-cutting and flexibility, but cabling reduction, which is a factory health and safety issue. Wood says it’s reliable and was simple to install. “They came and did a factory survey; then they installed the antennae. It works: there were a couple of blind spots, but they were sorted out… The touch screens are hard wired, but the hand-helds all operate using RF.” So where is the firm now? Beyond production planning and SFDC, the system is now moving on to material variances – for further stock control and de-stocking – to deliver more savings. “At the moment it’s being used to analyse variances in the top 100 products,” says Wood. “We will do more and get to a greater level of detail: we want to completely de-stock the stores, but we’ve got to be very confident of the forecast and the ability of MRP to deliver the materials we need just in time to be able to do that.” And when that happens: “By improving the accuracy of forecasts and enabling better production planning and process control, the system will give us a one-off de-stocking saving of £300,000 to £400,000 on raw materials. We will carry only what we need for the next week or so instead of the current three to four weeks’ stock.” Not bad. And he alludes to other benefits: like no more manual counting, with everything being swiped in and out, and better reporting, which will enable managers to keep a tighter rein on the business by monitoring yields, costs and profitability more closely. He also makes the point that with the system being real time, there are efficiencies that can’t be had when information only comes, say, at shift-end or every hour. “We can get a better handle on variance and do more to improve yield and reduce waste further.” It’s all about facilitating continuous improvement. It’s looking good. For the future, Wood sees the system potentially being expanded to provide web-based visibility for its customers of what’s in the plant and the various stages of production, as well as what’s been logged out onto the lorries. “We could work that back into our key suppliers so they can see the demand on us and synchronise delivery better.” That visibility could also help its retail customers manage their pipelines better, making transportation, regional depots and stores more efficient. For now the system is being extended to George Adams’ new frying plant, also at Ruskington, the other pie and sausage plant at its Spalding head office and its cooked meats division’s two plants at Boston and one at Frampton. And at the main Ruskington site, it’s being gradually extended. “We’ve got two or three years of fundamental development. The challenge will be ourselves, because the system is infinitely expandable. What it ultimately delivers depends on the imagination and drive of the individuals running it on the site.”