Manual versus Information Technology Enabled Lean Manufacturing

Manual versus Information Technology

It is easy enough to grasp the potential benefits of lean manufacturing (see Lean Manufacturing: A Primer, Lean Tools and Practices that Eliminate Manufacturing Waste, and How to Achieve Lean Manufacturing), but selecting the most appropriate lean techniques or tools and the accompanying packaged enterprise software for an individual enterprise has never been that simple. In fact, it is a major exercise for an enterprise to initially identify the most appropriate tools for eliminating different types of waste. For instance, overproduction could be mitigated by improved changeover times and balanced lines, whereas defects and rework could be curbed by improving visual controls, initiating more complete standard operation procedures (SOP) or operation method sheets (OMS), and implementing mistake proofing techniques at the source of error. Furthermore, waste of excessive inventory could be reduced by implementing kanbans and other similar pull systems, while waiting time could be handled by using takt times, and so on.

This is Part Four of a multipart note.

The trouble is further compounded by the army of software providers (including enterprise resource planning [ERP], supply chain management [SCM], manufacturing execution systems [MES], and product lifecycle management [PLM] providers, as well as best-of-breed, bolt-on lean specialists) that have been hyping their lean capabilities, despite the fact that most of them still support mere nuggets of pseudo-just-in-time (JIT) ways of accommodating mass customization. Providing only support for kanbans, order-less repetitive scheduling, or vendor managed inventory (VMI) or supermarkets, so as to push inventory elsewhere (e.g., onto suppliers) rather than to reduce it across the entire supply chain, is a far cry from true support for lean or demand-driven manufacturing. Where most of these flow manufacturing, lean ERP, or repetitive manufacturing systems fall short is that they have simply automated the most basic of tasks within a lean environment, without addressing larger issues of how to implement lean and pull practices in environments that are not easily amenable to these.

Then again, some people question whether computer systems are even needed for achieving lean manufacturing. After all, some lean tools entail merely physical processes and best practices on the shop floor, where transactional enterprise systems have little to offer. Also, given that computers were not widely available when lean manufacturing and kanbans first emerged, many enterprises have stuck with manually-driven lean methods. For such methods, an evolutionary step forward entails the use of custom spreadsheets and reports to support lean functions such as kanban management and heijunka calculations (see Lean and World Class Manufacturing and the Information Technology Dilemma—The Loss of Corporate Consciousness). It is interesting to note, however, that even in such cases, material requirements planning (MRP) systems still can be used to hold core master data on items and bills of material (BOM), though these records have to be tweaked with an eye toward lead time-oriented information.

Some lean purists go even further, and believe that lean manufacturing does not mesh well with information technology (IT) systems. For some, the only appropriate technology is Microsoft Excel spreadsheets. Others claim that the best scheduling method is "no schedule at all", giving the lean enterprise the utmost agility to react to any unpredictable event. On the other extreme, many people have become so accustomed to the use of enterprise systems, that they believe we can no longer return to manual procedures (see Run your Business with No Software!).

As usual, the truth might be somewhere in a middle—lean manufacturing and IT are not in opposition, and all good lean systems have both physical systems in the plant and near real time IT backbones that centralize data, especially if there is an automatic data entry and capture function. In fact, some people say that the whole point of the lean philosophy is to simplify the physical processes so that one does not need to manage overly complex data systems, though it is still necessary to manage the relevant data at the points where corrections are needed. To that end, many IT systems are designed to bring from the field only the data that management or decision-makers can do something about.

The reality is that most companies operate in a hybrid, mixed-mode environment where flow or lean and traditional batch or push manufacturing models coexist within the same facility, and where production and demand requirements can change throughout the different stages of a product's life cycle. Manufacturers can produce both high-volume goods with steady demand and low-volume goods with fluctuating demand, and their product mix may include engineer-to-order (ETO), make-to-order (MTO), and make-to-stock (MTS) items.

To successfully operate in this mixed-model environment, one has to take advantage of the strengths of each model and apply them where best suited. Thus, one should use traditional ERP systems for handling long lead-time items, one-of-a-kind production, and products with long production cycles, and for long-term budgeting and planning. On the other hand, lean manufacturing is often more easily applied to manufacturing operations with low-mix, high-volume, make-to-demand products. Moreover, one should not necessarily preclude pull-based execution processes from being implemented in to-order or highly configured operations, where it has also occasionally been done with great success.

Also, as lean spreads beyond the relatively stable manufacturing environment it was originally designed to support, companies realize that IT can play a vital role in streamlining the supply chain (see Moving Beyond Lean Manufacturing to a Lean Supply Chain). Namely, while the lean factory may use kanban pull signals to move product more efficiently through the manufacturing process and out of the door, it is missing the feedback loop from the factory to other functional departments within the organization or to the entire supply chain. That information is primarily transmitted and received via enterprise systems.

How Can IT Support Lean Manufacturing?

So, how can IT support lean manufacturing? For one, while complex packaged enterprise (ERP, SCM, etc.) systems may seem inconsistent with the simplicity of visual control, they actually work well together. In fact, although visual signals, such as kanbans and status indicator lights, are an effective way to trigger factory floor activities and the movement of materials, their inherent weakness is their lack of memory—visual signals cannot be recorded or tracked to determine historical performance or provide real time status for anyone that is not in direct view.

Yet, by coupling visual controls with real time collection of data from the factory floor, manufacturing enterprises should be able to capture the critical information behind the visual control signals for management oversight, planning, and accounting purposes. This information can be used for statistical analysis, to measure historical performance, and to monitor status—all of which are essential elements of the continuous improvement that lean manufacturing emphasizes. Lean aspiring manufacturers can also use enterprise systems to replace some visual controls, such as physical kanban card signals, with electronic ones, as a way to improve efficiency further and eliminate non-value adding activities.

Furthermore, these systems can play a critical role in establishing and ensuring standardized work. This is because they can serve as the central repository for critical engineering or product data management (PDM) information for standardized work, including BOMs, process routings or operations, valid product configurations, work instructions or SOPs, engineering change notices (ECN), schedule information, and costs. More robust solutions can even track as-designed, as-built, and historical actual product information, which can be analyzed to determine the impact that product changes have on efficiency and productivity.

Lean teams operate visually within the lean factory and move products as they determine necessary by visual signals on the shop floor. On the other hand, enterprise systems only send production information after such data has been entered into the software, which then activates triggers that move the information to the downstream recipient, letting them know it is their turn to work on the part. Even if this delay is not exactly in tune with lean principles, these lean teams still need data stored in enterprise systems, which contains information needed to perform their job (e.g., what to do with the part when they get it), understand the requirements of their customers (e.g., size, color, etc.), and understand the specifications of the job (e.g., quantities needed).

Enterprise systems also allow for this information to be organized, and, in some solutions with built-in workflow management capabilities, make this information easily accessible for employees to support engineering, production, regulatory, and customer needs. Some enterprise systems with constraint-based planning can help manufacturers reduce setup times, while those with strong enterprise asset management (EAM) capabilities can help implement total productive maintenance (TPM). These systems also allow for the near real time monitoring of factory floor activities, as they provide manufacturers with critical status information required to prepare for and execute changeovers. This status capability can be used to monitor machinery and equipment and communicate the completion of jobs or critical events such as breakdowns instantly.

Enterprise application systems, such as ERP, can also be used successfully to support lean enterprise transformations, especially for manufacturers that have highly variable demand for a large number of products and who operate in mixed-mode manufacturing environments. To apply lean principles to these new environments presents manufacturers with special challenges that the right ERP system can help overcome, such as the increased difficulty of calculating heijunka schedules, more frequent adjustment of kanban sizes, and increasingly smaller leveling periods. In these instances, the solution must have a planning system that can smooth demand for items with highly variable demand, and act as a shock absorber to maintain continuous flow and leveled production. The solution could also use a real time monitoring and feedback system to synchronize operations and trigger the movement of materials, as well as have automatic backflushing capabilities for demand-based inventory management and replenishment.

In fact, enterprise systems can even be used to support mistake proofing, thereby helping to prevent manufacturing defects from occurring in the first place and minimizing the impact that defects have on downstream activities. Computerized systems can prevent product defects by making standardized processes, critical documentation, and other quality information available to production personnel on an as-needed basis. Monitoring systems can also be used to flag defect-related issues instantly, alert downstream workers and activities, and record information for later analysis. On the other hand, rate-based scheduling applications can be used to stop production within manufacturing cells, allow workers to identify and correct defects, and then reschedule and restart production quickly to limit the impact on downstream processes. ERP systems can also allow manufacturers to backflush selectively for items and components affected by defects.

Advantages of Electronic Kanbans

Although it is often useful to start by using physical kanbans to give operators the feel for how the system works, once the system is understood, electronic kanbans have many advantages over physical kanbans. Most obviously, it is much faster and easier to resize electronic than physical kanbans without stopping and changing the whole system, so as to combine a satisfactory service level and maintain a low stock level. Moreover, although the number of kanban cards can always be calculated manually according to the formula below, fast number-crunching is certainly the prerogative of enterprise systems.

y = the number of kanban cards
D = daily demand
L = replenishment lead-time
α = safety factor
a = order quantity (container)

Other advantages of electronic kanbans include the following.

• The possibility of audit trailing electronic kanbans to avoid problems such as the loss or sabotage of physical token kanbans
• The ability to use electronic kanbans to quickly transfer information between production units or supply chain partners, which contrasts with the time and effort required actually to move physical kanban cards
• The fact that it is easier to phase in new products when using electronic kanbans as compared to physical kanbans, since it is important to collect and remove all the old kanbans from the system first
• The fact that supply chain networks are becoming more complex due to globalization, specialization, and outsourcing. If organizations want to use lean concepts with suppliers and customers, electronic kanbans are essential for effective communication to minimize inventory and ensure on-time delivery.
• Technology that also supports supplier and customer portals to enable JIT call-offs or pull-based control of material supplies

In addition, while workflow within a production cell or line is managed through physical signals, the design of the cell and its scheduling (loading) can be supported with IT. In particular, as mentioned earlier on, enterprise systems can be used to improve materials management and consumption within a production cell, and to provide real time monitoring and visibility.

Also, since takt time depends on variable customer demand, manufacturers require an effective way to account for that demand and smooth it when calculating takt time. Enterprise systems can assist since they allow manufacturers to account for end-product customer demand and then explode it into item-level demand for related feeder production lines to help establish takt time. These systems should also be able to account for use problems associated with routing operations when calculating takt time.

Enterprise systems can also help with calculations related to other fundamental concepts of lean manufacturing—for instance, every-product-every-interval (EPEI), which is the time it takes to run through every regular part produced in a process. Knowing the EPEI helps determine the manufacturing lot size and supermarket quantities for each part produced in a manufacturing process, as well as the number of kanban cards in the replenishment loop. One should also be able to calculate pitch—the time needed in a production area to make one container of products (e.g., if the takt time equals thirty seconds and pack size is twenty pieces, the pitch is then ten minutes).

Last but not least, software can help with modeling lean environments (e.g., value stream mapping [VSM], perform simulations, etc.); calibrating them (calculating and dynamically adjusting the appropriate lean controls, such as takt time, kanban sizes, complex algorithms for the management of constraints, etc.); operating them (i.e., automatic data collection [ADC] and process monitoring); and continuously improving them (by identifying the points of high variability, poor performance, and their root causes in order to adjust and improve them). After all, lean manufacturing at its core is a philosophy for continuous improvement. To that end, some enterprise systems have an ability to help manufacturers streamline operations by providing the best information needed to control processes, coordinate enterprise-wide activities, monitor quality, and measure improvements as effectively as possible. In particular, these systems' databases and built-in powerful analytics tools may identify customer value and support continuous improvement (see Contemporary Business Intelligence Tools).

In conclusion, regardless of the manufacturing environment, extended enterprise systems are required to deal with long, complex, and global supply chains; better demand visibility; long-term demand planning; trading partner management; sophisticated buy and sell contracts; and data for regulatory compliance and financial management. They also have the potential to eliminate waste in administration (e.g., order entry, order management, invoicing, etc.) through proper deployment of back-office systems. When it comes to forecasting, it may still be done by lean organizations, as seen earlier on. However, rather than trading partners setting fixed-number forecasts, which are set in stone, they will forecast within a reasonable range, tightening the range until a shorter time horizon is reached, at which point kanbans would take over the procurement execution.

Clock versus Calendar

ERP systems schedule activities using the calendar, and not necessarily a clock, as they typically plan for demand requirements stated in terms of days or weeks. However, when integrated with execution processing capabilities, such as those found in JIT scheduling, the combination can be very powerful. In a sense, it can provide the best of both worlds—long-term forecasting of monthly, weekly, and daily requirements is combined with an execution plan that is broken down to the minute.

In practice, long-term forecasting, based on electronic data interchange (EDI) supplier forecasts and ten day production, provides demand information, which is then exploded into material replenishment, production, and assembly requirements. From this, suppliers and planners can plan for items with long lead times, communicate longer ranger intention to their suppliers (allowing them to optimize their own procurement and production processes), and analyze the impact on potentially constrained production capabilities. JIT sequencing can then use this information in concert with its demand signals (measured in hours and minutes) to generate a minute-by-minute execution plan for the sequential and JIT delivery of configured goods to the source of the demand signal.

When it comes to planning, while lean manufacturing moves product through the manufacturing operations, it cannot perform any planning of activities beyond the immediate shop floor work. For this it requires the informational access that is inherent within an enterprise planning solution. The planning process is exclusively handled by the planning module inside the ERP software, which links to other ERP modules that provide data on inventory levels, supplier status, customer orders in the pipeline, expected parts receipts, etc.

In general, enterprise systems can handle considerably more information than any lean team could possibly accommodate manually. As a manufacturer's lean program develops, software becomes imperative for successful program execution and management. Achieving the lean metric goals of the factory requires access to the right information at the right time, and enterprise systems typically provide much of this access. Additionally, it is not always practical to have teams in such close proximity that the visual handoff of information is convenient, since information is often shared among physically and geographically separated teams. Hence, different buildings, suppliers, customers, and plants require information from the ERP or supply chain event management (SCEM) software.

This concludes Part Four of a multipart note. The Lean Manufacturing: A Primer series will continue February 20, 2006.