The Risks of Constrained Target Stock Levels
Inventory managers often believe that setting a minimum stocking levels for all materials will guarantee material availability and improve service levels but that may not always be the case. Stocking one unit everywhere using overrides or minimum service level constraints is generally not an optimal strategy. For example, setting a blanket service level of 95% will apply equally to all parts across all locations. That means that very expensive parts with a low failure probability will require stock at every location since a stockout would yield a 0% service level. However, taking a total cost optimization approach means that a planner may rather accept the risk of stockout since the cost to expedite once every 12 – 18 months may be lower than proliferation of expensive or bulky material. Conversely, parts with low cost and high volume, for which it would be unthinkable to have a stockout, may have their target stocking levels capped at the blanket 95% service level in locations where it would make more sense to stock 98% or higher. In other cases, customers set safety stock rules to keep one unit at all locations supporting install base, but this discounts any differentiation based on customer, SLA, or location attributes. We may be able to achieve acceptable service level over a large group of install base with less inventory if the install base population is dense and the replenishment lead-time is short. Conversely, we may need to stock more parts for the same size of install base (resulting in a higher service part to install base ratio) in sparsely populated or geographically hard to reach areas.
Consider a situation where limited inventory is needed at several locations. For example, there is a shortage at two different sites but there is only one unit available for replenishment. A planning environment relying solely on minimum stocking level rules would assign that material arbitrarily. On the other hand, a planning environment taking a total cost optimization approach (one considering inventory and stockout costs) would assign the stock to whichever site provides the greater business benefit; it would assign the stock where it has the potential to offset the highest stockout cost.
Allowing Prophet to calculate the optimized target stocking level rather than applying broad constraints and overrides is beneficial for many reasons. The Prophet algorithm calculates the point at which adding additional inventory cost does not provide any incremental value (as it does not significantly decrease the likelihood of a stockout). Ultimately, correctly configuring inventory cost and stockout cost parameters to determine optimal stock level enables your service supply chain to achieve the highest service level at the lowest total cost.