Modern companies set inventory targets in various different ways, from the entirely ad hoc to the formally optimal. Sometimes there are no targets at all, or companies will employ a “rule of thumb” system based on experience. Others will use single stage calculations for their inventory management, while the most advanced (and usually most profitable) make use of multi-echelon software tools.
Having no targets is a cardinal sin against the process of inventory optimisation but is woefully commonplace. Inventory will be managed by padding out schedules, manufacturing or ordering items earlier than they will be needed. This is highly inadequate for all but the simplest products and poses many risks.
More widespread is the use of “rule-of-thumb” inventory targeting, normally involving setting a Days of Supply, or DOS, coverage target for every item. DOS rules of thumb are very easy to understand but have two critical limitations which can make them of dubious reliability. Firstly, any DOS estimate is based only on average demand and has no way to account for seasonal or market variability. Secondly, DOS relies on an entirely forward-thinking approach, whereas inventory management analysis should ideally be backward-looking as it focuses on the outcome of decisions which have been made successfully in earlier time periods.
Single stage inventory target calculations are a general improvement over any rules of thumb, but still present several key limitations. These models need to be thought of as calculating versus optimising. If management only looks at a single item in a single location, the resulting inventory calculation is straightforward, but does not consider the item’s effect on the rest of the supply chain network. While the average inventory textbook will display this simple, single stage calculation it is a critical mistake to think that doing this counts as true inventory optimisation – these solutions actually simplify problems down in order to perform the single calculation. While this is an immense improvement over DOS, it still falls significantly short of managing a fully optimised inventory system.
The approach used by inventory optimisation consultants is state of the art and involves multi-echelon software which models inventory levels end to end across the entire supply chain. Unlike single stage models, multi-echelon inventory models do not assume that a specific isolated location must hold inventory, which gives considerably more freedom and accuracy to calculations. Modelling multiple stages in this way also allows the proper modelling of different types of inventories, such as cycled stock, prebuild and safety stock with time phased demands, to be properly taken into account. This leads in turn to a level of accuracy far closer to real world scenarios, and allows for proper management and optimisation of inventory with the inevitable benefits to the business such a situation provides.