High-tech Darwinism
Some thoughts on—and examples of—adapting to circumstances
By Frank O. Smith, contributing editor -- Manufacturing Business Technology, 11/1/2003 12:00:00 AM
Stu Reed helped engineer IBM's integrated supply chain (ISC) initiative, bringing procurement, manufacturing, distribution, and logistics into a single global division—which, with 19,000 employees, makes it the fourth-largest division in the company. Now the VP for IBM worldwide systems group manufacturing, Reed must deliver on a sizable chunk of that vision.
There is just one problem.
"As long as we live, we'll never be able to actually read the minds of our customers and accurately forecast what they want," says Reed. This being the case, "It's imperative we become more adaptive to meet their demands."
The crux of the matter is that term "adaptive." It often seems there are two basic approaches to managing complex operations—the brute force of a Goliath or the adaptability of a David.
Chess computers use brute force, evaluating every possible move on a crowded chessboard as many as a dozen moves ahead, to compete against flesh-and-blood players who bank on pattern recognition, tactical incisiveness, or other uniquely human skills. Often, the humans still win.
Chaos theory talks about systems based on self-directing agents, as opposed to building an all-encompassing model programmed to meet every contingency—an approach doomed to failure. The evolution of supply chain planning applications can be looked at this way, too. There, the ideal of having in-memory models so complex that they account for every SKU in a supply chain is leavened by recognition that the best approach may be a blend of planning and execution functionality. That way, events are reacted to free of the central bureaucracy.
Adaptability is a concept ripe for the times, says Navi Radjou, principal analyst at Forrester Research, Cambridge, Mass., which began looking at the notion of adaptability following 9/11.
"We need to look at supply chain strategies differently," says Radjou. "Companies need a portfolio of coping strategies to be adaptive. It will be different [depending on] companies and industries."
At IBM, says Reed, "We apply the concept of sense and respond—the ability of the value chain to sense change and respond immediately. And we want to do it end to end, across the entire supply chain, pegging the customer's order all the way back through the chain so that when one end changes, it pulls the other."
The reorganization for ISC, one of the most sweeping in IBM's history, focuses all decisions and activities on ensuring greater on-demand responsiveness, as opposed to stovepipe objectives. The company says it's cut average cycle time—from requisition request to procurement—from two weeks to two days. The company also says overall cost savings total $5.6 billion.
Sample solution set
At least one simple solution for increasing adaptability seems counterintuitive to mainstream supply chain thinking. Instead of getting the product as close to the customer as possible, postponement strategies involve stationing inventory in a central location. The benefit comes in having the flexibility to better meet fluctuating demand at points of sale by postponing final product configuration until after an order is in hand.
Postponement strategies are nothing new, especially in electronics and semiconductor. Several small software vendors, reputedly just emerging from academia, offer software for postponement. But Oracle Corp. says it's the first major enterprise vendor to feature postponement, adding that major benefits derive from its full integration into the Oracle enterprise suite.
Says Nadeem Syed, VP, advanced planning products, Oracle, "We did an aerospace and defense industry benchmark. By applying the 80/20 rule, [whereby it's assumed 80 percent of the business is focused around 20 percent of the parts] we identified a handful of parts, which, handled through postponement strategies, led to a 26-percent reduction in inventory and raised service levels from 68 percent to 94 percent."
(For more on postponement, see story on page 28.)
As supply chains get more complex, setting sourcing and staging policies that allow adaptive response becomes harder. Vendor Optiant is all about intractable supply chain problems, including supply chain design and configuration; low-cost, high-powered inventory policies; and risk analysis.
"Most folks think narrowly in terms of their own stovepipes," says Optiant CEO Dan Ross. "They don't always consider how changes impact the entire supply chain."
Optiant PowerChain helps companies make decisions in the best interest of the entire supply chain with "actionable intelligence," which sits on top of advanced planning & scheduling systems (APS), "just as APS sits on top of ERP," says Ross.
"It assesses the impact of strategies regarding inventory levels and locations, looking at how each affects lead time, costs, service levels, and cash-to-cash cycles, showing in half an hour—not the months it takes with spreadsheets—how they compare."
One customer Ross cites, a digital camera manufacturer, uses PowerChain to plot its Christmas season strategy. "They want to minimize obsolescence, but provide the best channel mix with the highest service levels," says Ross.
Collaborative and intermittent
Forecasting and demand management vendors are rushing to use the Internet to allow real-time collaboration with sales teams and others stationed in remote locations.
One example is the announcement at APICS in October by Demand Management of its DS enterprise (DSe) solution, which supports "100-percent native" Web collaboration.
"DSe's open architecture and platform independence offer users flexibility unlike any other industry application," says Demand Management President and CEO Mike Campbell. "It uses technology that no other software vendor has implemented for demand planning and forecasting."
Based on a distributed, thin-client architecture that is scalable to the needs of any size organization, DSe is written entirely in Sun Microsystems' Java programming language. DSe can be administrated from a single location, simplifying work for IT departments and improving connectivity to ERP systems. Users in different locations can view and adjust forecasts on different sets of items at the same time.
In similar fashion, Smart Software announced at APICS it would be introducing in the near future a version of its demand-planning business application that would allow collaborative feedback from remote personnel in pursuit of consensus forecasts. Smart Software is perhaps best known for its work solving the forecasting problem many companies face when the demand for some of their products is slow moving or intermittent. This problem is particularly costly for capital goods manufacturers, retailers tracking demand at the SKU level, and service-parts inventory managers.
Finally, both PeopleSoft and J.D. Edwards recently announced enhancements to their demand-planning offerings. To improve the ability for customers to generate statistical forecasts, PeopleSoft has licensed John Galt Solutions' ForecastX software and will integrate it with PeopleSoft's demand-planning applications. ForecastX lets manufacturers combine business knowledge, such as events and promotions, with historical data to build a forecasting process with a low range of errors.
"Our focus was to not reinvent the math around statistical forecasting," says John Webb, PeopleSoft VP for supply chain. The demand-planning applications sit on top of PeopleSoft's data warehouse, which helps in the analysis and collection of data.
Customer decides
Adaptability entails many things, but in manufacturing and supply chain there is one point of clarity around which everything else revolves: meeting customer needs.
"There are moments of truth with a consumer product, the first being when the customer decides to buy it at the store," says Larry Kellam, director of B2B network innovation with Proctor & Gamble, the Cincinnati-based consumer packaged goods (CPG) manufacturer.
Conventional wisdom in the CPG industry holds that out-of-stock situations typically run between three and five percent of total customer requests for products. When P&G recently tested this convention with its own research, it was disheartened to discover the reality was three times higher.
"We failed the customer at the first moment of truth more than we thought," says Kellam. Despite years of supply chain effort, "that's a very unsatisfactory outcome."
Efforts were too focused, Kellam asserts, on reducing costs and "not on servicing the customer." This realization prompted reordering of the supply chain to be adaptive to demand variability. "When you're talking about a $43-billion company, capturing even three or four percent of that missed opportunity results in a huge ROI," he says.
The company wants to be able to produce every one of its 37,000 SKUs daily. "Not necessarily that we would, but that we could," Kellam says. Over the last 18 months, the company says it has reduced cycle time from 30 days to seven.
Kellam speaks of a supply network of interdependent nodes of activity. Each node is a "decision sphere" with authority to take adaptive action to ensure product availability. For example, past policy mandated trucks be shipped full, often resulting in delays. Managers now act autonomously to meet delivery dates.
The company intends to put RFID tags on every pallet and case by 2004. (Wal-Mart, a big P&G customer, will require its top 100 suppliers to do so by 2005.) This will improve real-time visibility of product status across the network.
P&G works closely with the Auto-ID Center at MIT. In fact, the center director is a P&G employee on loan. SAP, P&G's worldwide ERP vendor, is a center member, and is working with P&G to develop agent technology tied to RFID tags for real-time alerting. P&G's goal is to have a sizeable percentage of goods-in-transit uncommitted, with point-of-sale information transmitted electronically to "rolling warehouses" for delivery-point determination.
As can be seen, P&G applies the principles of adaptive systems, leveraging awareness of best management principles with forward-looking use of technology. Its initiative, and others, contribute to the continuing evolution of supply chains. The goal is supplying the world's peoples with both basic goods and sophisticated consumer products, a worthy activity.
Solutions Spotlight
01/31/2004B2B with a twist
06/30/2002In the driver's seat with demand management
05/31/2002
