Last month we began our discussion of safety stock. Safety stock is “insurance” inventory you maintain for a product to prevent stockouts due to unexpected demand or delays in receiving a replenishment shipment. Most systems calculate safety stock quantities based on a number of day’s supply or a percentage of anticipated lead time usage. These methods are fairly easy to calculate, however they apply one calculation to all of your products.

We have found that a better way to calculate safety stock in many cases is to look at the average deviation or difference between the forecast and actual usage over the last several months. Items with an accurate forecast (i.e., little difference between the forecast and actual usage) get just a little safety stock. However those items with a less accurate forecast (i.e., with a significant difference between the forecast and actual usage) get more safety stock. The amount of safety stock is fine-tuned according to accuracy of the forecast.

To utilize this method:

  1. calculate the deviation between the forecast and actual usage over the previous several months. Look at this item example:


Mon. Forecast Actual Usage Deviation
Feb 50 60 10
Mar 76 80 4
Apr 80 70 -10


In February, the demand forecast for the product was 50 pieces and actual usage was 60 pieces resulting in a deviation or difference of 10 pieces. In March, the demand forecast was 76 pieces and actual usage was 80 pieces, which produced a deviation of four pieces. The average deviation for this item is:


(10 + 4)/2 = 7 pieces per month


You’ll see that items with an accurate forecast, have a smaller deviation and a smaller amount of safety stock. Whereas, items with an inaccurate forecast have a larger deviation and a larger amount of safety stock.


Note that the deviation for April, in which demand exceeded usage, is not considered in our calculation of safety stock. Why? Because if our prediction of what customers want exceeds actual usage, we certainly don’t want to add more safety stock to inventory! We probably have more than enough of the item on the shelf already.


Now, to take things a bit farther

  1. Multiply the average deviation by a deviation multiple. The deviation multiple used is dependent on the customer service level we want to provide. Customer service level is defined as the percentage of line items for the product shipped complete, in one shipment, by the promise date. The higher the multiple, the more safety stock we’ll maintain and the higher the customer service level. Generally, we’ve found that the following multiples will usually achieve the corresponding levels of customer service:





Approx Customer

Service Level

2 95.0%
3 97.5%
4 98.5%
5 99.0%


If our goal is a 95% customer service level, we’ll multiply the average deviation by a multiple of two (7 x 2 = 14 pieces). Using this method provides more safety stock to those items whose usage is more erratic. But be careful! Using a higher deviation increases the amount of safety stock (i.e. insurance inventory) on your shelf as well as your investment in stock inventory.

Next month we will examine a way to fine tune your deviation multiples to achieve your desired level of customer service at the lowest possible inventory investment.