What is safety stock?
Safety stock is the buffer held against sales coming in higher than expected or the supplier delivering late. If sales run above average between placing an order and the goods reaching the shelf, the buffer keeps the shelf from going empty.
Too little and customers find an empty shelf, losing sales and reputation; too much and money sits in stock while storage and shrinkage costs grow. The right amount depends on three things: how much sales vary, how much the lead time varies and how much risk you accept (the service level).
Safety stock is not an ordering rule on its own: added to the expected sales over the lead time, it gives the reorder point. When stock falls to that level, you order.
Safety stock formula
The simple rule works with maximum and average values; it is a quick start when all you have are sales and delivery records:
SS = (maximum daily sales × maximum lead time) − (average daily sales × average lead time)The statistical formula accounts for the variability of demand and lead time and for the risk you accept (the service level). VidiStock uses this formula in the application:
σLT = √( L × σd² + d² × σL² )
SS = z × σLT
ROP = d × L + SS| Symbol | Meaning |
|---|---|
| d | Average daily sales (demand) |
| σd | Standard deviation of daily sales: how much sales vary from day to day |
| L | Lead time (days): from placing the order to stock on the shelf |
| σL | Standard deviation of the lead time (days): how consistently the supplier delivers |
| z | The value for the service level: 90% → 1.2816 · 95% → 1.6449 · 98% → 2.0537 · 99% → 2.3263 |
| SS | Safety stock |
| ROP | Reorder point |
The simple rule has no link to a service level and is heavily affected by a single extreme day; the result can be too high or too low. If you have sales history, use the advanced mode: paste the history and σd is calculated for you.
Worked example
The same figures as the example on the home page: in the Ankara branch, "Water 1.5 L" sells 12 bottles a day on average, the standard deviation of daily sales is 10, the supplier always delivers in 5 days (σL = 0), there are 26 bottles on the shelf, one order should cover 30 days of demand, the product comes in cases of 10 and the target service level is 95%.
| Step | Calculation | Result |
|---|---|---|
| Demand over the lead time | 5 days × 12 a day | 60 |
| + Safety stock | 1.6449 × √(5 × 10²) = 1.6449 × 22.36 | 36.78 |
| = Reorder point | 60 + 36.78 | 96.78 |
| + Need for the order period | 30 days × 12 a day | 360 |
| = Target stock | 96.78 + 360 | 456.78 |
| − Stock on hand | 26 | |
| = Net requirement | 456.78 − 26 | 430.78 |
| Suggested order | rounded up to cases of 10 | 440 |
Intermediate values keep their decimals; only the final result, the suggested order, is rounded (in practice safety stock is held as 37 and the reorder point as 97 units). There are 26 bottles on the shelf and the reorder point is 96.78: it is past time to order, so 440 bottles are ordered. With the simple rule, taking maximum daily sales of 18 and a maximum lead time of 7 days, safety stock would be 18 × 7 − 12 × 5 = 66 — almost twice the 37 the statistical formula finds for a 95% service level. The calculator above opens with this example.
Limits of the formula
- Intermittent demand: For products such as spare parts that sell a few times a month, sales are zero on most days and a formula that assumes a normal distribution misleads. When you paste your sales history, the calculator detects this with the average demand interval (ADI) and warns you. VidiStock uses Croston, SBA and TSB for these products (methods and accuracy).
- Seasonality and trend: The calculation is based on the historical average. A summer product planned in winter, or a growing product planned with last year's figures, gives the wrong result; here safety stock also counts the predictable variation of demand as "uncertainty" and comes out too high. In the application the forecast error is used instead of σd (what demand forecasting is).
- Lead time variability: If the supplier sometimes delivers in 5 days and sometimes in 9, do not leave σL empty; lead time uncertainty often needs more stock than sales uncertainty.
- One product: This tool is for one product in one warehouse. Recalculating hundreds of products every week, separately per branch and with seasonality, is more work than can be done by hand.