Demand-driven Planning in Manufacturing

Push versus Pull

In traditional manufacturing, the time and cost of changeover to produce different products is high, as are the costs of inventory, planning, and expediting. Thus, goods are pushed through production at levels determined by often inaccurate scheduling and forecasting tools common in material resource planning (MRP II) and enterprise resource planning (ERP) systems. These levels often exceed demand, resulting in building excess finished inventory, while in a flow/lean/just in time (JIT) environment, orders are conversely "pulled" through the process, based on actual demand.

Part two of the Demand-driven versus Traditional Manufacturing Requirements Planning series.

In contrast, rather than emphasize planning such as ERP, lean manufacturing and JIT concepts, emphasize the continuous improvement of processes. For example, it emphasizes processes that lead to reduced inventory throughout the supply chain, shorter lead times, and faster cycle times, all enabling improved response to customer demands. Consequently, many vendors have bolstered their commitment to lean manufacturing practices with functionality designed to promote rapid response to customer orders based on demand "pull". This includes kanban and mixed-mode manufacturing, which makes several different parts or products in varying lot sizes, so a factory will produce close to the same mix of products that will be sold that day, with the goal to build every model every day, according to daily demand. It also has the flexibility to schedule and manage flow orders for products within product families, with or without using MRP. Again, the idea is for products to arrive exactly when they are needed, in the mix and quantities that are required. Production only begins immediately after a customer order is confirmed, without having to run MRP, or create and release a work order.

Materials are consumed from point-of-use locations or raw-in-process (RIP) locations, whereby both internal and external material suppliers receive replenishment signals at point-of-use locations and RIP, and build or supply material when the location needs to be replenished. These features are attractive, at least to existing ERP users, and in particular, to automotive suppliers, who have been pinched by a tightening economy and are under pressure to speed up operations and adopt JIT and lean manufacturing practices. The requirements for the automotive industry is particularly conducive to the deployment of lean concepts and can even extend into the accounting arena. Evaluated receipt settlement, release accounting, self-billing, and retro-billing, all speed up processes while helping suppliers optimize their limited resources. For more information, see Pull versus Push: A Discussion of Lean, JIT, Flow, and Traditional MRP.

This is Part Two of a two-part tutorial.

Part One covered traditional MRP planning.

Demand-Driven Planning

Professor Anders Segerstedt from the Swedish Lule University of Technology has developed an alternative to MRP that makes demand-driven manufacturing a reality. International Business Systems (IBS) (XSSE IBS B) is one of the few companies to incorporate his ideas of cover-time planning (CTP), into its business software applications—CTP (not to be confused with capable-to-promise) has been a part of the IBS demand-driven manufacturing solution since the mid-1990s.

CTP is also a "pull" system, but with the intent of planning and not merely waiting for the execution time to act. It does not take into account every minute detail and parameter, but rather, nothing gets produced unless it is going to be ordered or will be consumed. Using CTP alongside MRP, IBS software offers an alternative way of working by prioritizing time instead of quantity as the basis for production planning.

Taking the amount of parts and components that are on hand and ready to use, and dividing that number by the average daily consumption, yields a number corresponding to the number of days the inventory will last. Then, by subtracting the expected lead-time for the part (e.g., five days, ten days, etc.), the result will provide production planners with a clear indication of how far they are from a reorder point. This prioritization by time is the root of CTP, since by subtracting the lead time in a manufacturing process from the cover time (i.e., the amount of time existing stock can be used), creates a priority number. Thereafter, the manufacturing orders for the different items can be sorted and handled in a priority sequence to focus on what is most important and to avoid shortages.

In summary, MRP concerns itself primarily with the quantities of parts and components that a factory has in stock, and the quantities and dates for future requirements, such as "How much buffer stock does a factory need to meet demand before re-ordering?". Conversely, with CTP, the question becomes "Given the present and future consumption of parts and components, how much time do we need in order to cover anticipated consumption in the near future?".

Don't Forget Warehouse Management

Demand-driven manufacturing is also related to warehousing management, since the warehouse is no longer merely a static storage facility, and often has to use near real time data to closely match supply to demand, eliminate the need to hold excess inventory, and increase the flow of goods throughout the supply chain. Therefore, because of the capability of supply chain execution (SCE) software to handle these complex requirements, the trend has been to postpone many light manufacturing operations, such as final assembly, customized packing, labeling, engraving, etc., from shop floors to warehouses and distribution centers (DC).

A warehouse management system (WMS) package plays a key role in any company's manufacturing postponement strategy to delay the customization of products until the products, or a set of common components, have left the manufacturing plant. To that end, introducing the value-added services capability such as kitting assembly and disassembly, multiple or multilevel bills of material (BOMs), special instructions, and labeling, within the WMS module is an apt answer to helping customers reduce the costs associated with their supply chains. Here one of the most significant facilitators is postponement, and that is where kitting helps because it allows enterprises to keep their products in a more generic state for as late as possible.

Two different WMS solutions, one for traditional and the other for highly automated environments versions, are sometimes needed to cover all the necessary functionality for efficient operations and the management of warehouses. They provide the tools and functions for optimizing the use of space, while automating reception; put-away; location replenishment; picking; warehouse maintenance; packaging, shipping; radio frequency (RF); storage and product definition; and warehouse administration and monitoring. Some users have optimized their WMS processes to pick orders faster and with increased accuracy (based on product movement frequency); and to optimize resource activity, through precise and accurate instruction handling. The best way to describe their WMS is to work backwards. Just like in the movie Memento, where the main protagonist suffers a serious short-term memory loss, and the movie is told backwards in many discrete short time brackets, it is best to understand WMS by looking at the process in reverse order, from truck dispatch, to warehouse pickup.

The last action is often a truck dispatched to a specific location country—or worldwide. In these shipments, a consignment for a specific customer could comprise anything from several pallets to individual packages, whereby each batch within the consignment contains single products in consumer packs, each identified by its bar code. Prior to this, customer orders are automatically received by dispatch and the shipment is scheduled when orders are entered into the system, since the warehouse module is populated with parameters, allowing it to divide and suggest how each order should be packaged (e.g. whole pallet, half pallet, etc.).

As customers are able to use the Internet to place orders whenever they wish, the application must also automatically determine and update, in near real-time, all quantities for the particular customer at the moment of picking. The system then makes up work orders in sequence, and wirelessly relays the message to the screen of the relevant forklift operator in the warehouse. The first item in the work list is the shipping box or package, which the operator picks up and labels. As the operator scans in the bar code to confirm the item has been picked up, the next consumer package to pick, with its location, appears on the screen. The software then checks the scanned bar code against the work order to minimize errors. Once the shipping box is full, it is sent to the packaging station where cushioning material and the outer packing material seal the package, along with its shipping documents. From there, the packages are driven either directly onto a trailer or to a staging area awaiting the arrival of the shipping truck. In this way, trucks and drivers are continuously in motion, synchronised by the business logic of the software.

On their display all employees and users, such as truck drivers, forklift drivers, warehousing shift manager, etc., can see only the information they need, but they can always find more information if necessary. As items on the picking shelves are removed, a replenishment work list is then generated, which is again queued and sent to the relevant forklift truck display screen (with visual representation of the pallet movements), so that operators can move items from the storage area onto the picking shelves.

The analogous efficient processes (in reverse) can be seen for goods' reception, whereby the "interleaving" principle optimizes the use of forklifts (a forklift with a new pallet from production will go to a put-away place, to be then instructed to pick a new pallet for dispatch). Each originating production unit typically enters its upcoming shipments into the system, and generates the label with bar code so that the warehouse can automatically receive the item. This information is processed to generate a work list for a truck to take the goods to the storage area. Fast-moving items are automatically assigned to the more accessible lower levels to allow faster picking, while slower-moving products are assigned to higher levels.

To overview and manage this involving warehousing operation, staff extract daily, weekly, and monthly reports on an as-need basis. IBS particularly excels at traceability capabilities, whereby products can be tracked from purchase order to warehousing bin location, and to customer order with just a simple inquiry. Even batch tracking information can be included within extensible markup language (XML) message to trading partners. For more of pertinent information, see ERP and WMS Co-Existence: When System Worlds Collide.

User Recommendations

Manufacturers and distribution should understand the "part and parcel" of a more complete demand-driven solution, so that they can decide how much functionality they need for their business. Although many ERP and supply chain management (SCM) vendors have been professing lean or demand-driven functionality, most of them still support some nuggets of pseudo-JIT to accommodate mass customization. However, rather than reducing inventory across the entire supply chain, just supporting kanbans or vendor managed inventory (VMI) to push inventory elsewhere (e.g., suppliers) is a far cry from truly supporting lean or demand-driven manufacturing. On the other hand, MRP should not be discounted as useless just like that, since it will often have an important purpose that could and should be used together with lean, or demand-driven practices. MRP will typically handle planning, while lean can deal with the execution. Manufacturers must also be fully aware of whether their system uses actual demand, sales forecasts, or a combination of two in order to populate their MPS.

Also, manufacturers need to do some preliminary work before even thinking about deploying demand-driven manufacturing software, such as adapting their plants to a flow production model. In other words, they will have to operate in work cells that build families of products (rather than functional work centers producing large batches of components or products), and they will have established rules for sending replenishment signals to their internal (i.e., preceding work station) and external suppliers. These changes will not happen overnight, and the process should begin with the conversion of a few appropriate products with relatively simple production processes, and then progressing to other product lines. Thus, many manufacturers happen to be in a hybrid production mode, with part of the plant running according to flow principles, and the rest using traditional MRP methods.

One can never overemphasize the need for complete understanding of internal needs—from operations to IT to board level—and how everyone has to be on the same wavelength. While collaboration throughout the supply chain remains imperative, collaboration within the company is just as important. To that end, supply chain operations reference (SCOR), which breaks down the supply chain into individual processes and applies a consistent set of metrics throughout the supply chain, could often come in handy to measure the relative value of processes.