HOW TO THREAD THE DEMAND SIGNAL
Two demand signals are broadcast to each trading partner. The first demand signal originates at the point of use and represents an actual customer order. The first demand signal initiates action at the customer end of the supply chain and at the system constraint (see Figure 1). The second demand signal originates at the system constraint, and is broadcast to all the trading partners as a synchronization signal. Every trading partner receives both demand signals because the system constraint may shift among trading partners as the SKU mix varies with demand. In this way, the entire supply chain quickly adjusts by being synchronized from a different constraint.
A shipping buffer of finished goods inventory (FGI) is maintained close to the customer end of the supply chain. The shipping buffer should hold about three days' supply of FGI at the maximum throughput rate. This quantity should be enough to give the system constraint time to catch up with unusually high demand that exceeds its daily capacity. The end trading partner uses the shipping buffer to satisfy the daily quantity of customer orders from the first demand signal. The system constraint trading partner compares the quantity of the customer order from the first demand signal against its maximum daily capacity. The system constraint then
broadcasts the synchronizing demand as the second demand signal. If the customer order quantity exceeds the daily capacity of the constraint, then the system constraint trading partner manages a backlog for the supply chain until it can catch up with daily customer orders.
For each SKU:
• RULE: The downstream trading partner (retailer) uses its shipping buffer to ship the full customer order quantity from demand signal one.
• RULE: If the system constraint can meet the full quantity of demand signal one, then it broadcasts that same quantity as demand signal two.
- If the system constraint cannot meet the full quantity of demand signal one, then it broadcasts its constrained quantity as demand signal two. The system constraint adds the backlog quantity to the demand for the next period, and compares this
total quantity to the constraint. The next day the backlog quantity will ship first.
- For each SKU demand signal two will be exploded into an equivalent number of assemblies or component required daily at each node.
• RULE: All other trading partners produce the daily quantity of products, assemblies, or components as broadcast over demand signal two.
• RULE: The upstream trading partner (parts supplier) orders raw material to match the rate of incoming customer orders on demand signal one.
If for some reason a trading partner falls momentarily behind, then that trading partner will manage its own backlog for one or two days until it is caught up. If a trading partner falls steadily behind for a predetermined period, such as five working days, then that trading partner becomes the system constraint.
COLLABORATIVE FORECASTING, PLANNING, AND REPLENISHMENT (CFPR)
Customers are unwilling to wait until their order can be processed, materials can be ordered, the product can be built, and the finished goods can be shipped. They want it now! As a result, forecasting, sales and operations planning, distribution requirements planning (DRP), master production scheduling (MPS), material requirements planning (MRP), and purchase order launching for long lead time part replenishment is just as necessary in a synchronized setting as it is traditionally. Since every trading partner now has access to the actual daily customer orders, collaborative forecasting, planning, and replenishment among all trading partners becomes practical.
There are some differences in the way traditional push planning interconnects with non-traditional synchronized operations. Synchronized operations depend more on probabilistic than deterministic planning and analysis. And it is important to separate planning the rate of material consumption from planning the mix of material consumption. On the one hand, increasing the rate of supply chain throughput can only be done when both the rate of incoming materials is increased and the capacity of the system constraint is increased. This can only occur beyond the time frame of the planning horizon as defined by the longest path through the ordering cycle time, plus material lead time, plus manufacturing cycle time, plus transportation transit time. A lead time map should be created for all the parts in the bills of
material. Keeping some inventory on hand for the 10 longest lead time parts can provide some degree of upside response time. On the other hand, altering the mix of supply chain throughput can only be done by keeping an inventory of statistical safety stock on hand for each unique part number in the product family mix. A list of unique parts can be drawn up by dividing the complete bill of materials into the set of common parts and the set of unique parts. An analysis of actual order history can then be used to compute the statistical range of unique parts inventory required for a given level of customer service. In a synchronized operation, the material and capacity to support the rate of supply chain throughput is forecast, while material and capacity to support the mix of supply chain throughput is based on history.
Use the following steps to analyze statistical safety stock:
• Compare the individual bills of material from each end item to determine the list of purchased parts that are unique to an end item SKU.
• For each unique part number, compile the actual order quantities by end item, by month over the past six months, then multiply by the quantity per.
• Compute the mean of the six monthly quantities for each SKU.
• Compute the standard deviation of the six monthly quantities for each SKU.
• An inventory level equal to the mean plus two times the standard deviation will provide a 95 percent service level to buffer uncertainty in the demand mix; a greater or lessor percentage service level can be set by
using a different multiple of the standard deviation.