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).

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).