Why Safety Stock Is Unsafe

00:08 Introduction
00:33 How would you define safety stock?
02:05 When did these ideas come about? Why is this something the market is hang up on?
04:04 Where are the key difficulties with safety stock?
08:57 Why is the probabilistic approach better than the safety stock one?
13:08 How do you deal with customers that need this extra safety value that is safety stock?
14:43 Is there something between the complexity of the probabilistic approach and the lesser accuracy of the safety stock one?
18:53 Why is the wastage in safety stock something that people haven’t tried to improve?
22:28 What is the key message today?


Safety stocks are an inventory optimization method that enforces an extra quantity of stock beyond the expected demand in order to maintain a target service level. This method relies on key statistical assumptions about the demand forecast, most notably that the error is normally found in the distribution.

In the 1920’s people came up with the idea of inflating demand predictions in order to more safely cover periods of high demand. This method, commonly known as safety stock, gained a huge following and as a result was hard wired into many ERP’s to protect against variability in both demand and lead times.

Despite the fact that this method was conceived before computers were even invented and that both technology and supply chains have transformed so dramatically since, saftey stock still remains a staple approach in many supply chains. Furthermore, even though it is a relatively simplistic concept, it is unfortunately often used far too conservatively, resulting in overstock and wasted inventory.

In this episode of LokadTV, we explore just how this approach became so extremely popular, why it is still so widely used and where this methodology fails. In addition, we discuss the modern alternatives that are available to a supply chain professional.

Often, when people use safety stock, they blame the results of forecasts for any errors rather than the tool itself. The root cause of the problem is the fact that normal distribution will always underestimate risk. Uncertainty is inevitable and needs to be embraced, yet safety stock methods don’t usually approach this the right way.

Here, we try to explain why many professionals never really notice when they are overstocked and why they add coefficients to safety stock to overcompensate for stockouts. We investigate the idea that uncertainty can be seasonal, as well as demand. Finally, we try and see how businesses can achieve less stockouts with less stock.