Just ask Val Zanchuk, president of Graphicast Inc., a 29-employee foundry and machine shop in Jaffrey, N.H. The entrepreneur was able to make this dream a reality by adopting the Lean Manufacturing philosophy.
He admits, though, that he didn’t have a choice. His business is built on the rapid turnaround of machined castings made from a zinc-aluminum alloy injected into graphite molds. Consequently, his people need a reliably fast way of changing over the machinery to make low volumes of a high mix of parts. “From any given customer, we may see orders three to five times a year,” he says. Retaining the flexibility necessary for quick turnarounds as the company grew would have been impossible without a continued commitment to Lean principles.
He began applying these principles in an orderly way as soon as he bought the company back in 2001. “When I walked in the door and sat at the former owner’s desk, I began looking at the paper flow, not only in the office and accounting, but also how information was flowing to the floor and throughout the operations,” he says. His goal was to streamline administration so that nothing impeded the flow of information to manufacturing.
Six months later, Zanchuk’s Lean agenda migrated to the factory, where he began applying various Lean methods and allied practices as they made sense. Among the first of these methods was 5S (for sorting, straightening, shining, standardizing, and sustaining), which strives to boost morale, safety and efficiency through organization and good housekeeping. Because frequent set-up was a key constraint for the business, another of those early projects was set-up reduction.
In fact, set-up reduction was so crucial that he enlisted the help of the New Hampshire Manufacturing Extension Partnership (MEP), an agency in Concord that is part of the federal extension program organized through the National Institute of Standards and Technology (NIST). Patterned after the old agricultural extensions, this network of extensions provides small manufacturers with training in Lean and other manufacturing methods that could help them to be more competitive.
Next on Zanchuk’s agenda was the variability that had been plaguing the casting process. “A part would run fine one day, and on the next, half the part runs would have to be scrapped,” he explains. “Everybody had a theory about the cause, but we really didn’t know what the problem was.” To solve the mystery, he got a research grant from the state to work with the Engineering Dept. at Dartmouth College, in Hanover, N.H. The researchers conducted a series of statistical experiments to determine the process’s key parameters and to create an accurate model.
Based on their findings, they developed a flow chart for the operators to use for troubleshooting problems. “We were able to reduce our scrap rate considerably,” says Zanchuk. “We also found that we could increase our yields, and therefore our overall throughput, by running the process slightly slower.”
Perhaps the most stunning results came from the new Easy Lean scheduling module for the ERP Visual software that the Graphicast had been using from Infor Global Solutions, of Alpharetta, Ga. “As our business grew, lead times were getting longer,” says Zanchuk. “Our schedule was booked solid, two shifts a day, for two months out.” The daily schedules generated by the old module, moreover, did not tolerate disruptions well and required frequent human intervention to tweak assignments that did not quite make sense.
The new module changed all that—quite dramatically—by using throughput accounting and applying the theory of constraints. Consultants at Infor had predicted that Graphicast would be able to cut its 16-week lead time in half, so the installation technicians programmed the software’s lead-time setting at eight weeks. By the end of the first day of the week-long installation process, however, it became apparent that eight weeks was way too long. So, Graphicast decided to go to six weeks on the second day and five on the third. A month later, after the company got comfortable enough with the shorter lead time, it dropped it again to the current four weeks.
“It was incredible to think that we were booked solid for 16 weeks and that we couldn’t expand without adding people and machinery,” notes Zanchuk. “And a month after the installation, we were running a four-week lead time.” Capacity doubled, and the work-in-process was cut in half. The change also saved about $100,000 a year in overtime.
Another benefit was that the module flattened the decision-making structure in the machine shop. Because the schedule automatically creates itself every morning based on the four-week lead time, the skilled machinists in the shop are now responsible for checking it and deciding which machines will work best for which jobs. Not only do the machinists now have more control over the shop, but the people who had previously been doing the scheduling were freed up to work on activities that were more suited to their skills.
Observers and successful practitioners alike agree that such results are possible only when top management either drives the effort personally, as Zanchuk did at Graphicast, or makes the Lean program a priority. The reason is found in basic human nature: The priority of the boss tends to drive the employees’ energy and excitement. In fact, the biggest reason given for the failure of Lean programs is a lack of interest on the part of senior executives.
Ralph Rio, research director at Dedham, Mass.-based ARC Advisory Group Inc., reports that the secret to interesting top management in a Lean program and holding that interest is to wrap the results in an enticing package. “Senior executives are not all that much different from you and me,” he says. “They have lots of things on their plates.” So, if you want to attract and keep their attention, translate your successes to the metrics that these managers use to run the company—that is, the items on the financial statements.
Often, the efficiency gains generated by a kaizen event or other Lean activity amount to a quarter-person’s worth of work. Although these gains are real and permit the business to grow without adding people, they do not appear on anyone’s profit-and-loss statement because no one actually leaves the payroll. “If the savings are a bunch of quarter-persons, then they are not going to be credible to financial folks,” says Rio.
To generate results that are credible, Rio urges the members of Lean teams to read their companies’ annual reports and to find ways to connect their successes to the line items on the financials. Inventory reductions and inventory turns tend to be the easiest metrics to affect. “If you reduce inventory by $100,000, that is going to show up right away on the balance sheet,” he says. So, he recommends looking at inventories first for improvements that will catch the attention of senior executives.
Build corporate memory
Rio also recommends embedding any improvements directly into the business’s automation in order to make them a permanent part of the corporate memory. One way of doing this is to adjust the information technology to enforce the improvement. Another is to design the improvement into the equipment used in production. In stamping presses, for example, a simple dowel pin can automatically align the dies at changeover. This is much faster and more repeatable than the old iterative process of stamping a part or two and moving the die around into the right position.
Automation has been the centerpiece of the Lean-Manufacturing strategy at Plastic Components Inc. (PCI), Germantown, Wis., since Tom Duffey and the other two founders started the plastic injection-molding business back in 1989. “We wanted to build a facility that has as little human involvement in the manufacturing process as possible,” explains Duffey, who serves as the contract manufacturer’s president.
Consequently, each of the 42 injection-molding machines currently on the shop floor is a self-contained manufacturing cell tended by robots and other supporting automation. Only seven people on each shift are needed to handle the materials, change the molds and load part programs into the machines. The technicians put the molds into the same press that they used in the initial qualification stage to develop and optimize the part program. “This ensures the best possible consistency from lot to lot,” says Duffey.
When PCI was about 12 years into its growth, it became evident that it needed another kind of automation, one that would replace the Access database, Excel spreadsheets and yellow legal pads that it was using to manage its resources. This primitive information management system was simply inadequate for scheduling the nearly 130 million parts that the company now makes annually for its customers all over the world.
“We couldn’t get our arms around the information that we needed to run our operation efficiently,” explains Duffey. “As a consequence, we were adding an extra 5 percent or 10 percent into every step of the process so we wouldn’t be caught short. If we thought that we needed 18 hours for a production run, we would schedule 20. If we thought we needed 167 pounds of material, we’d bring out 200 pounds—and run all of it just to use it up.” There were some years in which the resulting extra inventory wound up being a $200,000 write off.
To tighten the discipline in his shop, Duffey automated the flow of information by hardwiring the machines, warehouse and office into a local-area network. Then, he installed EnterpriseIQ software from IQMS, an enterprise resource planning (ERP) system provider based in Paso Robles, Calif. So, when the office releases an order for 10,000 parts to be delivered on Tuesday, the software now schedules the appropriate mold to go to, say, Press 26 at 3:15 this afternoon.
Meanwhile, it also tells the warehouse to deliver the necessary 167 pounds of material. “The material handlers scan exactly that much material from the warehouse and bring it to the press,” says Duffey. “When the machine completes the 10,000 parts on schedule at 4:30 the next day, the parts are scanned from the floor into finished goods at 4:31. And another job can be scheduled into Press 26.”
Not only does the automation instill discipline on the shop floor that keeps machine utilization high, but it also controls inventory. Management always knows exactly where the finished goods are. Because, moreover, PCI has linked itself electronically to its resin supplier, it has established a just-in-time relationship with it. “If we schedule a 10,000-piece job to start tomorrow afternoon, IQMS automatically triggers a release to our resin distributor,” says Duffey. “They put the resin on a truck and have it in our plant at eight o’clock the next morning.” Consequently, PCI is turning its inventory 24 to 26 times a year, which is three to four times faster than the industry average of seven. So, it’s no wonder that Lean is “in” there and elsewhere more than ever.