Visit Our Sponsors |
A lot of big companies have been taking it on the chin in financial terms recently as introductions of new product lines are hit with schedule delays, cost over-runs, and quality challenges. The plight of several prominent A&D companies has been front-page news, especially following recent GAO reports citing the large percentage of DOD programs suffering from these maladies. But commercial product companies across the board have recently stumbled over similar problems.
Two of the most frequently cited underlying causes are late deliveries and poor quality from suppliers. The first point at which OEMs know there is trouble is when a shipment of critical parts misses a delivery date, or a supplier component fails in end use. The result is always costly, and sometimes catastrophic. Recently, some of the more progressive OEMs have been engaged in the development and implementation of predictive supply chain metrics, designed to act as an "early warning system" that enables the OEM to get out in front of trouble, and be more proactive to head off problems while they are still embryonic and surmountable.
The types of metrics being developed by these companies are appearing around critical path activities through the upstream production processes within suppliers of critical components. It is a bit like the inverse of Quality Function Deployment (QFD). Rather than looking at which characteristics are critical to the OEM's customers and assuring that those characteristics are met by internal processes, the OEMs are identifying critical characteristics for themselves, identifying those processes in upstream supplier operations that produce those characteristics, and instilling real-time or near-real-time monitoring mechanisms at those processes.
These companies are approaching the problem in different ways, with the most frequent approach centered on those components that have historically proven to be the most problematic. This approach has met with some success, but is a bit like driving with the windshield painted out, using only the rear-view mirror. Even if the metrics used to look upstream into supplier operations are sound, the components and suppliers under surveillance will not always be the ones from which new problems flow. New suppliers, suppliers undergoing significant changes in equipment or manufacturing processes, and other issues frequently provide nasty surprises. A more comprehensive approach, based on an evaluation of the complexity and difficulty of producing all of the components provided by suppliers is proving to be more successful. Companies performing these evaluations focus on the complexity of the design, the tightness of tolerances, the uniqueness of manufacturing process steps, the maturity of both product and process design, and the position of the component in terms of final product critical path to identify which purchased components should be included in the purview of predictive supply chain metrics surveillance. Purchased components that reflect a mature design in both product and processes, sporting solid process capability numbers are least likely to be problematic.
For example, rotary wing aircraft (helicopter) manufacturers have determined that the main rotor shafts, gearboxes and engines are typically critical path purchased components. They are not only among the most complex and expensive components, but they are elements of the overall product design that directly and immediately impact end-to-end cycle times when they are not available on their need dates. In the case of the main rotor shaft, unless the manufacturer is dealing with a completely new product, the design of both the component itself and the manufacturing processes involved are both usually quite mature. As long as a few critical predictive metrics such as availability of the raw material at the supplier, supplier machine capacity, and supplier process capability are monitored with fidelity the likelihood of a problem from late receipts or defects is low. Engines, however, are often more challenging. Even when the product design is mature, complex components like engines are more frequently the target of design revisions, and an engine is naturally a more complex element than a shaft with hundreds of subcomponents and often involving multiple tiers in the supply chain. Gear boxes (transmissions) lie somewhere between the shafts and engines in complexity and frequency of design revisions traffic. When multiple tiers are involved (think multiple levels in an indentured bill of materials), the likelihood of a failure somewhere in the chain is multiplied, and a deeper understanding of a broader set of manufacturing processes must be understood in order to effectively establish predictive metrics. In the case of the gearbox, machining operations associated with each gear and shaft are involved, and bearing diameters, lubricant performance, and other factors come into play. As multiple tiers of suppliers become involved, the linkages and interdependencies of the supply chain make the process of identifying predictive measures and instantiating them becomes harder. More assumptions enter the equation, and fidelity can suffer as a result. It is far less difficult for the Tier 1 supplier to introduce and maintain predictive metrics with their Tier 2 supplier than it is for the OEM to maintain predictive measurement mechanisms that reach through Tier 1 and into Tier 2 operations.
The challenges faced by companies on the path of developing and instantiating predictive supply chain metrics are three-fold: First of all, there is no small amount of work involved in identifying the critical path elements in the OEM product, and establishing the appropriate predictive measures involved in producing those components. While a list of the critical path components often contains many of the same components that a dollar-descending list contains, they are not usually identical. Some of the critical path elements are not necessarily among the most costly parts of the end product. Identifying critical path elements involves an analysis of the way in which the product comes together, along with the buffers and interdependencies. Then, once the critical path and components associated with it have been identified, that information must be maintained over time. Secondly, getting the suppliers involved to participate in providing the metrics data can be a daunting task. Those who are most successful are typically the OEMs whose business represents a substantial part of the revenue of the supplier, so that financial leverage can be brought to bear to induce the suppliers' participation. Some of the techniques being deployed effectively here include building the requirement for such reporting into the overall contract language, or a related services clause (sometimes called a master services agreement), and utilizing an independent third party to gather and report the required data so that appropriate confidentiality levels can be maintained between OEM and supplier, utilizing strong nondisclosure agreements and similar protections. The third challenge involved is the accurate, consistent, and effective reporting of the conditions where predictive metrics indicate that potential problems loom on the horizon. In this area, the most effective implementers are utilizing business intelligence tools to generate management "dashboards", turning those metrics from "Green" to "Yellow" when a potential for problems is identified, and from "Yellow" to "Red" when the problem is actually experienced at the supplier, and will likely result in trouble for the OEM in future weeks. Management dashboards like these that enable executives and Supplier Managers to "drill down" to the underlying supplier, component, and manufacturing process involved are becoming a popular choice.
The benefits from instantiating predictive supply chain metrics are primarily rooted in cost avoidance. There has been a great deal of recent press related to schedule slips in new product introduction, schedule slips and cost over-runs cited by the U.S. Government General Accounting Office in defense programs, and the increasing dependency on multi-tier supply chains born of deliberate vertical disintegration of operations by major manufacturers. Many companies over the last decade have elected to divest themselves of fabrication and subassembly in an attempt to "move up the value chain", endeavoring to widen profit margins by eschewing themselves of the "messier" end of the business. They have been working hard to adopt a model as far afield from Henry Ford's original vertically integrated automobile manufacturing model as possible. As a result, the management of complex product manufacturing has mutated into a far more supply chain management focused activity. Essentially, in order to be as effective as they once were at the integration and introduction of complex products into the marketplace, they have traded the hard work of managing internal operations into what they initially perceived to be the easier task of keeping supplier performance on the rails. The development and instantiation of predictive supply chain metrics is one tool that shows real promise as a way to regain management visibility over the design and production domains that were previously under their direct control. While it still doesn't make direct intervention possible, at least the introduction of these tools gives OEM manufacturers an opportunity to see trouble coming, and intervene by spooling up other sources of supply or adjusting their own product mix and volume to compensate for the problems headed their way.
Contact William Duncan at wduncan2@csc.com.
Source: Computer Sciences Corporation
RELATED CONTENT
RELATED VIDEOS
Timely, incisive articles delivered directly to your inbox.