Extending into lean

When George Onderdonk joined Oregon City-based Benchmade Knife Co. as business system analyst in March 2005, the company had made its move to Lean manufacturing. But the rate of progress in achieving leaner production processes wasn't reflected within the company's production control

If the idea was to eliminate nonvalue-add activities, manual data entry from paper-based work order travelers was incongruous to company goals. Worse, a manual kanban system meant that inventory and costing information was lost from sight until kanban completion, at which time it was manually keyed in, and its bill of material back-flushed at the end of the production process.

A much better solution, thought Onderdonk, would be to scan bar-coded work orders and kanban at each point in the production process to gain visibility and eliminate manual data entry. But how? The SYSPRO enterprise system in use by Benchmade offered limited lean functionality, and any major investment in alternative solutions was out of the question.

Sunken costs and commitments

It's an all-too-common dilemma. For most businesses, enterprise platforms represent multiyear commitments to specific vendors based on a match of features and requirements at a point in time—in Benchmade's case, 1995. As those requirements change, wholesale replacement isn't an option. To support lean, manufacturers must find a way of leveraging existing platforms with innovative, cost-effective approaches. Only rarely is a fresh start a genuine option—and when it is, the rationale had better be good.

In Benchmade's case, Onderdonk knew a potential solution lay in SYSPRO's e.net solutions product, which delivers software functionality as discrete objects of code—each object being a Microsoft COM model that users access via scripting languages such as Visual Basic for Applications.

“Think of them as objects containing business logic in component form, allowing users to integrate with third-party products, or build up a custom solution,” says Rene Inzana, product marketing manager for SYSPRO. “Because the business object takes care of things like data relationships, it makes development pretty simple.”

Toward full power

It's a simplicity that appeals to Onderdonk, who sees it as a major step forward from previous generations of component building blocks. “In the past, companies deployed a lot of Microsoft Access-type ODBC [open database connectivity]—which works just fine, until a field name or other parameter changes,” he says.

Working part-time on the code from last January until the summer, Benchmade management was so encouraged with the progress that Onderdonk was assigned to the project virtually full-time. “It's very much still a work-in-progress, but we can see that the prospect of eliminating significant amounts of data-entry time is a very real one,” says Onderdonk. “When it's complete, we'll be operating far more efficiently, and have far more visibility into what's going on in the production process.”

Extending the scope of the enterprise platform is another way of delivering much-needed lean functionality. Such was the case at Stockholm-based Husqvarna Group, which in 2005 saw difficulty with a lean implementation at its Beatrice, Neb.-based manufacturing facility, where a recently acquired product line—Bluebird garden machinery—is made. The problem: a combination of space constraints and process constraints, exacerbated by a spreadsheet-driven attempt at pull-based scheduling.

Complexity in maintaining so many spreadsheets prompted Husqvarna to model production resources around lean principles, using Pelion Systems software to analyze demand and calculate kanban. The software uses value-stream and process-flow analysis to determine where changes in a facility could improve material flow. Furthermore, using electronic kanban replenishment to signal suppliers would allow the company to pull work through the line at the exact pace dictated by demand.

“We had no effective means of analyzing the peaks and valleys in our demand,” recalls Steve Habrich, materials manager at the Beatrice plant. “Pelion could help us do that analysis and determine the right kanban sizes based on those peaks and valleys.”

Measurable results came within 90 days of implementing the software, with on-time delivery performance climbing to more than 95 percent. And with optimized material flow, Benchmade makes better use of factory space, thereby easing the process constraints that dogged the company.

“A manufacturing plant is about activities, connections, and flows. That's what delivers product, and that's what ERP doesn't do,” says Kevin Fallon, CEO of Pelion. “Clean up the plant—then model it, and then pull demand through that model. Our customers have taken millions of dollars out of their inventories doing just that.”

Fast work of it

But “clean up” how, precisely? For Northampton, U.K.-based Cosworth Racing , a wholly owned performance-engines subsidiary of Ford Motor Co., it meant augmenting the enterprise platform with a finite capacity scheduling solution from Preactor International —a decision prompted by the desire to “lean out” the plant's high-precision, capital-intensive machining operations, which could not be accomplished with Cosworth's UNIX-based enterprise platform: ManMan, acquired by SSA Global in 2002.

Beginning in 2000 with the Preactor P300 system, the solution has been enhanced with additional Preactor modules and extended throughout the enterprise.

For example, says Darren Dowding, Cosworth's production scheduling manager, machine shop operations are greatly simplified by a “what-if?” facility that can be run on-demand. Electronics assembly, previously scheduled on an infinite-capacity basis using spreadsheets, is now finite capacity-scheduled with Preactor.

Component due dates also improved, given the solution's ability to recognize work orders that combine with others to form wholly new work orders. Throughout the company, this type of waste is slowly being whittled down.

But if a finite scheduling extension to an enterprise platform can help a business schedule its operations within a lean context, says Graham Hackwell, technical director at Preactor, it also can help create that lean environment in the first place—”cleaning up” the plant by eliminating fat changeover times and constrained processes that generate waste. Thanks to underlying plant data that must underpin a finite capacity schedule, he notes, “The system acts as an agent for change: It helps you decide where to focus improvement efforts.”

For example, it's possible to locate bottlenecks within the plant by looking at the time that orders spend queuing at resources. By differentiating between an operation's run time and its set-up time—a fundamental aspect of a finite capacity scheduling application—the plant can identify those operations with the longest set-up times, allocating engineering and improvement resources toward reducing them.

“Classic lean treats all kanban equally, and is indifferent as to the order in which they are filled,” explains Hackwell. “But the real world isn't like that: In practice, there might be rising demand for one kanban, falling demand for another, and static demand for still another.” In which case, he argues, the best order to manufacture the kanban would be the one with rising demand first, then the one with static demand—and finally the one with falling demand.

Can't kanban alone

Yet kanban on its own doesn't provide the information to enable manufacturers to do this. What's needed, Hackwell contends, is a way of supplementing the operation of the basic kanban with insight to the demand history for the item in question, or a trend indicator of the frequency of replenishment. “You need visibility into the entire process,” he says.

In some situations, bolting on lean functionality to the enterprise platform won't work—especially when that platform is old, or underpowered, or no longer supported. In such situations, rather than extend the functionality of their enterprise platform to achieve smoother coexistence with lean, companies might choose to replace it altogether.

That's the path taken by Kansas City, Kan.-based Accessible Technologies (ATI), manufacturer of ProCharger brand systems for automotive, marine, and industrial applications. Arguably something of a lean manufacturer to begin with—the company has a Dell-style build-to-order business model with minimal inventories—ATI chose the EnterpriseIQ enterprise platform from IQMS to replace an assortment of niche solutions that handled accounting, shipping, and sales.

One feature of the IQMS solution, explains IQMS President Randy Flamm, is its constant recalculation of the forward production schedule as new orders come in and existing ones are completed. For ATI, this meant an opportunity to schedule the machine shop more precisely and frequently. The schedule is reissued twice a day, while a drag-and-drop user interface enables production control personnel to fine-tune it if required, changing the sequence of jobs to take advantage of opportunities to reduce setups.

The benefits of the new system are clear. “We've nearly doubled our production capacity with existing staff, and cut customer lead time to just three business days from order receipt to shipment,” says Ken Jones, CEO, ATI. “There's much greater visibility of information, and much better discipline in the shop. Compared to what we've achieved in the past, we're making better decisions.”

Better still, adds Director of Engineering Ray Dicks, ATI's in-house lean guru, a lean-aware enterprise platform is giving the business the confidence to extend its still-young lean program with more formal activities such as value-stream mapping.

“We're still picking the low-hanging fruit, and haven't got into full-blown lean yet,” he concludes. “The IQMS system is about creating the infrastructure to give us the confidence to take the next steps. When you eliminate inventory, you eliminate the comfort factor that the inventory provides.”