Bear in mind that holding a safety stock is not free and should be carefully computed per type of products. Many companies are building safety stocks just“in case of” but never use them or because they lack planning & scheduling.

**Demand variability / Delivery variability**

In many cases, the demand variability follows a normal distribution, enabling easily to size the safety stock.

The first requirement is to compute the average demand (mean value) over time and the standard deviation.

From that, the required confident interval is needed to size the safety stock: 95% or 98% confidence level, in other way a 2% or 5% stockout risk. Never try to go higher than 98% as the safety stock level would be that huge you’d be in troubles with the accounting team.

Here is the normal distribution on a 95% confident interval:

Let's work out the calculation on the following exemple:

- Average daily demand (d) is the average inventory used per day
- The average supplier leadtime (LT) is the expressed in the same time bucket as for daily demand (e.g calendar or working days)
- The EDDLT is the Expected Demand During the Lead Time, defined as the product of daily demand by the leadtime

Then we need to choose a given service level - here let's go for a 98% - meaning that on average 98% of customer orders are filled out of the current inventory. So we work here with a 2% stockout risk.

We need first to calculate the average EDDLT which is the product of the average demand with the average leadtime.

Then we need to compute the standard deviation of the demand during leadtime, then use it with the willing service level to get the safety stock level.

Then the Reorder point = EDDLT + Safety Stock

Here with a 98% confidence, we have a safety stock of 82 pieces for a 26 pieces used on average per day on a 18 days leadtime from the supplier.

Thus the reorder point is then 465 pieces + 82 pieces so 546 pieces.

**Download below**** the spreadsheet here with all calculation details.**