The life cycle of items is getting shorter and shorter. Many items nowadays exist for only one year or sometimes only one season. Just think of items in the apparel industry or consumer electronics. Managing these short life cycle items is complicated. If these items are purchased in Asia, only one purchasing decision can be made, often long before the selling season starts. Information regarding consumer behavior is in those cases not, or only limited, available. Hence, the demand forecast is nothing more than a shot in the dark. To cover these uncertainties, the sales department decides to provide a sales forecast that is on the safe side.
During the selling season it is impossible to react to changing circumstances. The long lead times of Asian suppliers ensure that the selling season is long over before a replenishment order arrives. So the purchasing department also decides to play it safe. If an item proves to be more popular than expected, at least the purchaser need not explain why the item is not available and why nothing can be done about it.
In this way safety is put on safety. The consequence are massive leftovers at the end of the selling season, virtually impossible to sell against a decent sales price. High discounts need to be given, leaving no margin at all. At the same time, these discounts disturb the market and cannibalize next season’s regular sales. While higher management blames the inventory manager for the warehouse full of non-moving inventory, the sales department gets the ungrateful task of selling these items, even though the newer, innovated, items are already available.
To solve these kind of problems, the newsvendor model was created. A mathematically optimal solution, since day and age taught at supply chain management educations. The analogy is that of the newsvendor roaming the streets to sell newspapers. Each morning the newsvendor has to decide how many newspapers to purchase. Does he purchase too little, then revenue is lost. Does he buy too many, then at the end of the day he is the proud owner of a pile of old paper. Every morning the newsvendor faces the same challenge of finding the optimal purchasing quantity maximizing his expected returns.
Mathematically speaking the solution to the newsvendor problem is relatively easy. Practically speaking however, the solution is less evident. The first hurdle is expected demand. If the newsvendor exactly knows how many newspapers he will sell, then the number of newspapers to purchase in the morning would be straightforward. The complexity is created by the uncertainty of demand. It matters whether this uncertainty is high or low. Many companies already struggle with creating a proper baseline forecast, let alone that they have any clue to the level of certainty of this forecast.
A second complexity is the salvage value of the item at the end of the selling season. For a newspaper this salvage value is quite obvious. Yesterday’s newspaper has no value at all or at most a few cents as waste paper. This salvage value will certainly influence the decision of the newsvendor. The higher the salvage value, the less risk involved in buying too many newspapers in the morning. For many items the salvage value is not as obvious. What is the salvage value of a pair of ski trousers at the end of winter? Against what discount can these trousers still be sold?
To make it even more interesting, there is a beautiful circular reasoning in the salvage value. If you estimate a too high salvage value, then the risk involved is assessed too low and you are likely to purchase too many items. Meaning that at the end of the selling season more inventory is leftover, so that you will have to provide higher discounts to get rid of it. These higher discounts mean that in the end the actual salvage value is even lower. Are you still with me?
It is easy to raise objections against applying the newsvendor problem solution. But by ignoring the solution, we do not do ourselves justice. The strength of the model is not just the mathematical formula calculating a theoretical optimum based on a number of debatable assumptions. The strength is in creating a common decision model for sales and purchasing with only one focus: the company’s bottom line. A decision model in which all relevant factors are addressed, so that for instance possible leftovers are already addressed at the beginning of the selling season instead of only at the end. A decision model that is based on actual numbers instead of only on gut-feeling.
The newsvendor model is applicable in all kinds of situations. For instance for promotional activities or for the final purchasing decision of an end-of-life item. Basically, the model can be applied for every one-shot purchasing decision.
Real-world applications show potential large benefits and the practical complexities can be solved by making straightforward assumptions. A recently executed analysis on real customer data shows that by making straightforward assumptions order quantities could decrease by 10%, leading to 15% less leftovers and a 30% higher profit at the end of the season. In these results we haven’t even included the other positive side-effects of having less leftovers, such as the reduced workloads for the planning and sales departments. Of course the model needs to be embedded in a good process, supported by the right, easy-to-use, ICT tools. The newsvendor solution can make an excellent availability during the selling season, but with limited leftovers at the end of the season, happen!
