Today’s global marketplace requires manufacturers to profitably respond to the dynamics of market demand in a timely manner. Due to the convoluted and often far-reaching nature of today’s supply chain, this is frequently more difficult than it seems. Pressure to cut costs, decrease inventory and raise profit margins make the process even more challenging. The end result is often a misguided effort to implement “more effective” demand-driven supply chain methods.

However, this approach is unlikely to maximize savings. Demand-driven supply chain guidance is a broken system, and no amount of precision will create greater dividends.

Manufacturers often operate on a forecast or plan for a “bucket” of time. These forecasts lead to lot sizes or economic order quantities that amplify an outcome known as the “bullwhip effect.” When actual demand varies from the forecasted plan over the course of the allotted time, the bullwhip effect creates large swings in inventory. In other words, small changes at one end of the supply chain have multiplying effects further upstream.

The bullwhip effect is primarily due to two factors. First, daily demand can be highly variable close to the customer’s point of consumption. And second, forecasts rarely consider the cost, capacity or time to produce and deliver the product (not to mention changeovers and sourcing cost, time and capacity).

Here’s a hypothetical example of this kind of demand-driven supply chain model: Based on input from sales, a customer sends a forecast to the manufacturer.

The manufacturer then consolidates this forecast with other forecasts from several customers, preparing its own composite forecast and sharing that with its contractors, suppliers and other players in the supply chain. In this scenario, the manufacturer is forced to carry substantial inventory (either owned by them or their suppliers) at each transaction point in the chain. These safety buffers ensure that stock-outs do not occur if the supplier cannot deliver on time or if there is a spike in demand.

In theory, with an accurate forecast, this system should work out well. However, reality reveals otherwise. Unforeseen variables divert real transactions from the
projected path, leading to stock-outs as well as significant excess stock throughout the supply chain.

Becoming “Demand Responsive”

It’s worth rewinding a bit here and returning to a more basic question of the objective of supply chain management. The goal is to source, make and deliver the product from the point of origin to the point of consumption in the least amount of time at the lowest cost. Given that goal, the two most important attributes of supply chain management are responsiveness to the velocity of product flow, and the ability to move products quickly and with agility. These attributes enable the transition from push-based replenishment to pull-based replenishment. To focus on these attributes, it’s vital to look to the channel toward the customer, the customer’s customer, or the end-user of the product.

Since few companies can source material and produce all their products in the volume required to meet a day’s consumption, the solution lies in consumption-based replenishment. Consumption-based replenishment has its roots in lean manufacturing and “kanban” replenishment—a Japanese term for a card used to signal the need for inventory replenishment. Kanban has come to describe the “pull” method of keeping production lines optimally stocked with parts at the exact time and quantity needed. The simplest way of describing this is to think of a carton of milk in the supermarket—as one carton is pulled off the shelf, another slides into its place, immediately restocking the inventory.

In today’s connected world, this kanban principle can be effectively implemented with the use of technology, as consumption-based replenishment has evolved from its simple card-based roots into highly sophisticated software applications. These applications can help manufacturers determine the most optimal inventory levels for their operations, and they can rapidly recalculate efficient replenishment trigger points as demand varies over time. Better still, they can be implemented without going through the journey of a business transformation.

By taking a customer-centric approach to daily operations, manufacturers focus investments on enabling operations to build only what is needed to replenish what the customer has ordered. This paradigm shift results in companies embracing a goal of delivering the products customers want, when they want them, and in the quantities they want.

Figure 1 shows the differences between replenishment driven by a material requirements planning system (MRP) and replenishment governed by a consumption-driven system for the same demand pattern. The blue line shows the actual usage on a day-by-day basis for 365 days. The red line shows the on-hand inventory using an MRP system. The green line shows the simulated on-hand inventory using kanban replenishment for the same usage pattern. The most glaring disparity is the gap between actual usage and on-hand through MRP replenishment. While the maximum usage on any single day does not exceed 150 pieces, actual on-hand inventory ranges from 525 to 1,100 pieces. Clearly, consumption-driven replenishment aligns on-hand inventory more closely with usage.

Consumption-Driven Replenishment

For those who are used to MRP replenishment, this concept might take a little getting used to. An MRP system spits out replenishment orders based on several factors, including forecasts, planned demand, actual customer orders, open supplier orders, allocated inventory and available inventory. The focus for buyers is to monitor MRP runs and take care of exceptions, such as expediting, deferring or canceling purchase orders.

The consumption-driven model responds to demand changes, but always triggers replenishment on actual usage or consumption. By definition, it also deals with standardized lot size and lead times. So if the demand increases, the consumption velocity increases and vice versa.

The simplest form of consumption signals occur when a raw material warehouse triggers replenishment after material has been picked or moved. Even though it may not have been actually consumed, technically it has moved to a work-in-progress state (or an “allocated state” in MRP terms). The triggered lot size could potentially be set to match the standard supplier lot size, making the units of measure easily transferable.

In this instance, this could be triggered either from the supermarket or from the manufacturing line side. This tends to be more accurate, as the material in question is physically consumed prior to releasing a replenishment signal. However, it can pose some challenges, when:

The same part is consumed from multiple supermarkets or manufacturing lines.

The replenishment lot size of the supplier is much larger than the bin sizes being consumed (pallet vs. boxes).

It is difficult to deliver material directly to the line when it arrives.

In a consumption-driven environment, it is critical to keep track of inventory health. This really deals with the risk of goods and items stocking out. A mechanism is needed for a system to alert or identify high-risk parts.

One such mechanism is an electronic dashboard that highlights to the buyer just those parts that are at risk of running out. Stock-outs are typically caused by one of two things—either consumption has gone up dramatically or the supplier is not shipping within the lead time.

An electronic dashboard will highlight whether it is projecting stock-outs based on when the next shipment is due and the average usage of the last 90 days. It is then up to the buyer to expedite or inform the planner to change schedules or other variables to prevent shortages.

Establishing this kind of closed-loop process dramatically improves material replenishment in several ways. First, it helps companies reduce inventory by over 45 percent, since they only carry “just-in-time” inventory at the plant floor. Second, sending clear, electronic replenishment orders to suppliers based on actual consumption eliminates the bullwhip effect and streamlines delivery. This will improve the supplier’s on-time delivery performance.

Let’s return to our original example scenario. Collaboration based on timely communication of actual order and inventory quantities simultaneously across the supply network improves the performance of both individual organizations and the network as a whole.

In a consumption-based replenishment model, ordering activity becomes a series of well-timed, interconnected, synchronized loops between the preceding and succeeding processes. Buffer inventory that is usually maintained at tier transaction points can be substantially reduced. Material replenishment levels and continuous information flow provides visibility for suppliers and customers. More predictable replenishment eliminates intermediate tiers and buffer stocks in both the customer and manufacturer supply chains. Replenishment signals flow continuously through the supply chain and are delayed at each tier only as long as it takes to consume or ship material.

To learn more about how to create a demand-responsive supply chain using electronic kanban, download a free whitepaper here: