Summary
This article walks through how standard safety stock is calculated
Applies To
All items in StockIQ that have safety stock calculate, like items with an order policy of order point.
Process
One of the methods that StockIQ uses for calculating safety stock is a fairly standard statistical method.
StockIQ's standard safety stock calculation takes into account 4 important factors to arrive at a recommendation. It is based on:
-
Statistical Forecast Error: items which are highly forecastable, that have low forecast error, don't require as much safety stock since you are able to predict the variability.
- NOTE: For items that don't yet have a forecast error, 100% is used.
- Lead Time: Long lead times require more safety stock, as you are exposed for a longer period over which the variability can occur.
- Forecasted Usage: Items upon which you have higher demand will necessarily need a correspondingly large buffer. This usage quantity is controlled by the Safety Stock Usage Source value calculated by StockIQ, according to your Global Safety Stock Settings and/or any safety stock override you've put in place where you override the Usage Source.
- Target Service Level: In order to achieve a higher service level objective, you need a larger buffer of safety stock to guard against unusual demand.
Here is the process we use to calculate the statistical inventory levels, including safety stock:
1. Convert the Monthly Expected Error percentage to a weekly unit error. The Weekly Average is computed as a forward looking 3 month average of Forecast-0 (The forecast you saved).
Weekly Error = Monthly Error Percent * Forecasted Weekly Usage * sqrt(4.2)
2. Test the Weekly Error to make sure it is not greater than the Weekly Average. If it is, this indicates we are not dealing with a a normal distribution, so to accommodate this we set the Weekly Error to the Weekly Average, or if enabled, we use the Exponential distribution instead of the Normal distribution.
3. If enabled, we compute a Net Service Level based on your order minimums and multiples. The logic behind Net Service Level is this: the larger your average inventory due to large mins or multiples, the less likely at any given time that you will run out of product. For example, if a required multiple of 400 is six months of supply, then you will only ever be out of a product roughly two times a year. So, the required safety stock buffer can be reduced because your high cycle stock is providing service level on its own.
4. Calculate the Statistical Safety Stock amount using the service level goal which is converted to a Z-Score (standard deviations)
Statistical Safety Stock = Z-Score * sqrt(LeadTime / WorkDaysPerWeek * (Weekly Error ^ 2))
5. Calculate the Min based on the lesser of Statistical Safety Stock and whatever the cap settings are for your safety stock values. This is a sliding range between 80% and 500% of demand-over-lead-time depending on the lead time of the part. A purely statistical safety stock can sometimes result in a number that is not always practical or desirable. For this reason, StockIQ allows you to place logical caps on safety stock. The cap is based on lead time, and decreases as lead time increases. The reason for this is that as lead time increases, your ability to expedite for a truly aberrant order increases. As well, demand volume and volatility typically end up capping safety stock on their own in these cases. So, beyond 30 days, a minimum cap, usually 80% of a lead time's worth, is employed. By contrast, down at 5 day lead time, a maximum value of 5 times (500%) of a lead time's worth is allowed, to preserve service level interests.
These caps are configurable in the Global Safety Stock Settings screen.
Capacity Constrained Safety Stock
You can also feed StockIQ a maximum per-item capacity, either in units or days of supply. When this is done, StockIQ will back into the calculation of your safety stock based on your max inventory level, and make you aware if this is resulting in a constrained safety stock value. For more information, see the Item Warehouse Capacity and Shelf Life Planning topic.