Multi-enterprise Responsiveness-Can It Ever Be Achieved?

To keep pace with the volatile nature of today’s global market, many global manufacturers have done away with the traditional vertical supply chain. Horizontal and virtual supply chains are the latest trends in supply chain management (SCM), as these systems are more flexible and more responsive to the changes that inevitably occur when doing business today. But with this greater flexibility comes a price: manufacturers now have less visibility into and less control over their supply chain processes. For more background on trends in SCM, please see part one of this series, Who Could Object to Faster, More Responsive Supply Chains?

To know what tools manufacturers need to regain control over their supply chains, it may help to look at the applications and systems they are currently using.

Reassessing Existing Tools and Practices

Enterprises need to re-examine and redesign their supply chain processes and supporting IT tools to accommodate more responsive collaboration within a multi-enterprise, multi-echelon context. Most current enterprise resource planning (ERP) systems in use (as technical backbones) not only suffer from the vertical integration mind-set (i.e., they have a single-enterprise or manufacturing in-house orientation), they also suffer from being forecast-driven rather than demand-driven (see Demand-driven Versus Traditional Materials Requirement Planning) and from dealing with extended time brackets (i.e., weeks, months, or quarterly cycles). However, these systems merely record transactional history, and they require many complementary processes to address operations. In other words, ERP systems have to trigger too many additional external (often manual) transactions for more granular scheduling to occur.

To illustrate, sometimes users must perform manual steps on the ERP data to make it fit for use. Such steps may include creating production dispatches and schedules for production lines, which are often presented in a post-processed spreadsheet instead of coming directly from the ERP system in a useful format. Also, ERP systems typically cannot perform the following: map individual items to product lines; recognize the most appropriate order-scheduling rules; split days into shifts; and present input, such as adding finished goods replenishment needs to the production scheduler in an out-of-the-box manner.

As for the order-promising that is needed when moving toward a make-to-order (MTO) environment, ERP systems can typically show available inventory. But what MTO manufacturers need to know is the exact product line’s (work center) capacity for a particular routing operation by seeing the next open slot in real time from a single view. These companies also need to know raw material availability to ensure that the capacity can be used.

A manufacturing execution system (MES) can help to resolve these issues to a degree—see What Are Manufacturing Execution Systems? However, in addition to the well-known issues of integrating two systems that “live in different worlds and think in different terms” (see The Challenges of Integrating Enterprise Resource Planning and Manufacturing Execution Systems), the question remains of how large a step forward an MES is for responding to unplanned events versus doing more of the same (i.e., recording history, albeit in more granular, plant level).

Further, advanced planning and scheduling (APS)—see Remember APS?—and supply chain planning (SCP) systems came as improvements to ERP in the late 1990s, but only in terms of strategic- and tactical-level optimization (and again, mainly in the realm of long-to-mid-range planning), with hardly any help in terms of real-time operational advice to provide a solution or action in the nick of time.

APS uses linear programming, which imposes limitations on how it arrives at optimal solutions, since linear programming does not deal well with uncertainty. The APS system assumes that the input parameters are fixed and certain, that relationships are clear-cut, and that a single action results in a single result. However, in a sophisticated supply chain, actions may have nonlinear results that these systems cannot predict. In other words, planning-oriented applications do not allow for a fast enough response when changes in demand, inventory or supply, capacity, product mix, or orders occur. At best, these systems will offer another replanning exercise, and analysts then have to pore over mountains of irrelevant data to find the cause of a problem.

While this does not mean that APS calculations are useless and cannot be trusted, it does mean that the calculations should be compared to real results, and some processes may need to be modeled or simulated separately. One possible solution for managers suspecting that some of the APS’s inputs are highly variable would be to run a Monte Carlo simulation, which uses random variations to simulate chance. However, even if such commercially available solutions exist (similar to ERP and APS products), these too would typically be confined to a limited number of trained users and would not lend themselves well for the collaborative real-time environment.

Some organizations will then turn to business intelligence (BI) and analytical solutions, since if the ERP and APS systems have weak analytics, they will probably arrive at merely feasible rather than optimal solutions. However, while investing in management decision support systems (DSSs) should become a priority in terms of time and spending once transactional systems are complete, BI DSSs mainly score and magnify history. They are unable to provide a useful answer to the “now what?” situation of a customer canceling a major order (or increasing an order quantity) or an engineering department introducing a new product. Predictive analysis of demand and customer behavior can help in such situations (see Predictive Analytics—The Future of Business Intelligence), but to our knowledge, such commercially available solutions for manufacturing and distribution processes do not currently exist.

Sales and operations planning (S&OP) also comes to mind as a helping tool. APICS Dictionary defines S&OP as

a process to develop tactical plans that provide management the ability to strategically direct its businesses to achieve competitive advantage on a continuous basis by integrating customer-focused marketing plans for new and existing products with the management of the supply chain. The process brings together all the plans for the business (sales, marketing, development, manufacturing, sourcing, and financial) into one integrated set of plans.

Still, while S&OP is a huge step toward establishing and instilling effective and efficient collaboration—one by which all parties can explore options, wrestle with trade-offs, and develop a shared understanding and mutual commitment to a resolution—the problem is in S&OP’s focusing mainly within the single enterprise and on the level of tactical plans (versus operational ones).

APICS Dictionary continues to define the process as being

performed at least once a month and is reviewed by management at an aggregate (product family) level. The process must reconcile all supply, demand, and new-product plans at both the detail and aggregate levels and tie to the business plan. It is the definitive statement of the company’s plans for the near to intermediate term, covering a horizon sufficient to plan for resources and to support the annual business planning process. Executed properly, the S&OP process links the strategic plans for the business with its execution and reviews performance measurements for continuous improvement.

In summary, traditional applications were designed for the longer-term planning that takes place in a single enterprise. Hence, their approach and architecture are based on a batch sequential-processing (i.e., plan and measure execution feedback), and with cycle times taking weeks. The mass of complex analytics requires several hours to regenerate, so ERP updates are often run overnight or on weekends. Such a time delay is no longer acceptable or appropriate, as supply chain participants are looking for immediate information and answers.

Resorting to Spreadsheets Won’t Help Supply Chain Responsiveness Either

Therefore, despite the fact that many manufacturers have invested in ERP and APS systems, most continue to use ill-timed batch reports and wretched spreadsheets to manage their operations’ performance. These tools have proved to be inefficient and error-prone methods of supporting decision making, resulting in reliance on educated guesswork rather than on accurate, dynamic analysis to align operational decisions with strategic objectives. Thus, despite the use of spreadsheets and other desktop tools (see Vendors Harness Excel (and Office) to Win the Lower-end of Business Intelligence Market), this “reality-gap, problem-solving process” is still mostly manual and hunch-based, and as such, hardly ever quick or effective.

To add salt to the wound, traditional applications do not easily allow for information-sharing between enterprise partners, as they have a proprietary electronic data interchange (EDI) or a flat files-transfer setup at best, which means they cannot create and remove partnership links dynamically. The shortcomings of such solutions are only magnified when it comes to outsourcing; they offer very little in terms of visibility across multiple partner sites and data systems.

Understanding the overall business context is impossible with so many disparate and disconnected systems, because no single application can access all the data. For example, when a transportation management system (TMS) is tightly linked with a warehouse management system (WMS), the TMS might alert the transportation manager of a shipment problem, for example, and the WMS would in turn notify the warehouse manager. The manager would then ensure that the rest of the order is still delivered to the customer (with the customer being notified of the back order). Conversely, in the case of a disconnect, if the transportation manager found and shipped an alternate, neither the warehouse manager nor the customer would know this, and the customer would likely receive duplicate shipments (albeit who knows when).

In addition to minimal automation and insufficient speed and optimization, possibly the most troublesome reality-gap problem is the “multiple versions of the truth” situation. With several people working on the same problem (both within the company and across the supply chain), differing data sources will be found among each person’s individual spreadsheets. Even if each party has the same data source in theory, each one will likely have acquired the data at a different time and, as data inevitably changes over time, these participants will again have different data, which in turn yields different versions of the truth. As a result, different people will likely arrive at different “optimal” solutions.

Within the intense time pressures of the reality gap, a slow, manual decision-making process is further complicated by the “which version do you believe?” problem. These factors only add to the complexity and the lag. Namely, when issues arise (a major change in demand, a rush order, supply disruptions, and so on)—as they always will—there is no time to wait for ERP reports, dig for data among multiple sources, or perform ad hoc analysis using spreadsheets. Thus, to coordinate responses across the virtual enterprise, brand owners often have to manage demand response across their fulfillment networks by brute force, and relying on ever-changing spreadsheets to cope inevitably leads to misallocated goods and needless partial orders.

This is part two of the five-part series Who Could Object to Faster, More Responsive Supply Chains? Part three explores the features and functionality that today’s manufacturers need from SCM to regain control over their supply chains and maintain their competitive edge.

comments powered by Disqus