Skip to content

Safety stock calculator

How much buffer stock keeps you from running out, at the service level you choose? Enter how much your daily sales and your deliveries vary.

The chance of not running out during a restock cycle.

In units. STDEV over your daily unit sales in a spreadsheet gives it.

From placing an order to having the stock ready to sell.

From your past deliveries. Empty means deliveries arrive when promised.

Result

Fill in service level, average units sold per day, standard deviation of daily sales, average lead time to see the result.

How it works

If sales and lead times were always average, you would need no buffer. Real ones vary: some weeks sell fast, some deliveries arrive late. Safety stock covers that variation up to the level of risk you accept.

The service level is the chance of getting through a restock cycle without running out. The higher you set it, the more stock it takes, and each extra point costs more than the last.

variation over the lead time =
  √(lead time × σ daily sales²
    + daily sales² × σ lead time²)

safety stock =
  z × variation over the lead time
  (z = 1.645 for a 95% service level)

reorder point = daily sales
  × lead time + safety stock

What each term means

Service level
The cycle service level: the chance of no stockout between placing an order and receiving it. It is not the fill rate, the share of units shipped on time, which needs a different formula.
z
The service level's point on the standard normal distribution: 1.282 for 90%, 1.645 for 95%, 2.326 for 99%.
Average units sold per day
Units sold over a recent normal period divided by its days.
σ daily sales
The standard deviation of units sold per day: how far a typical day is from the average.
Average lead time
Days from placing an order to having the stock ready to sell.
σ lead time
The standard deviation of your past lead times, in days. Empty means deliveries arrive when promised.

Which variant: the combined formula for variation in both demand and lead time. With the lead-time variation left empty it becomes the demand-only formula, safety stock = z × σ daily sales × √lead time. It assumes daily sales vary around a steady average, roughly along a normal curve, and that the lead time varies independently of demand. For fast-growing or strongly seasonal products, recompute it from recent numbers. Then use the reorder point in the reorder point calculator to see how many days you have left.

Source: Wikipedia, Safety stock (the service-level formulas and the reorder point) (read 26 September 2026).

Worked example

Example numbers, round on purpose, not a real store:

Service level
95%
Average units sold per day
30
σ daily sales
12 units
Average lead time
21 days
σ lead time
4 days
  1. 1z for 95%: 1.645
  2. 2Sales variation over the lead time: 21 × 12² = 3,024
  3. 3Lead-time variation in units: 30² × 4² = 14,400
  4. 4Combined: √(3,024 + 14,400) = √17,424 = 132 units
  5. 5Safety stock: 1.645 × 132 = 217.1, so 218 units, about 7.3 days of sales
  6. 6Reorder point: 30 × 21 + 218 = 848 units

Without the lead-time variation the same example needs 91 units: 1.645 × 12 × √21 = 90.5. In this example, late deliveries more than double the buffer.

Frequently asked questions

  • What is safety stock?
    Stock you hold on top of the average demand during a lead time, to cover the cycles when sales run faster than usual or a delivery arrives late. Without it you run out in any cycle where demand during the lead time is above average.
  • What service level should I choose?
    It is a business choice: how often you accept running out in a restock cycle, against the cost of the stock that prevents it. Each point costs more stock than the last: with this page's example numbers, 95% needs 218 units and 99% needs 308. You can set higher levels for products where a stockout costs the most and lower ones elsewhere.
  • How do I find the standard deviation of daily sales?
    Export units sold per day for a recent, normal period of a few months, leaving out days you were out of stock, which understate demand. In a spreadsheet, AVERAGE of that column gives the average and STDEV gives the standard deviation. Do the same with your past lead times for their standard deviation.
  • Why does lead-time variation matter so much?
    Because a late delivery keeps you selling from stock at your full daily rate: each extra day of lead time costs a whole day of sales. That is why the formula multiplies the lead-time variation by the average daily sales. In this page's example, the lead-time term is 14,400 of the 17,424 under the square root: most of the variation comes from the deliveries, not the sales.
  • How is this different from max-minus-average safety stock?
    That rule multiplies your highest daily sales by your longest lead time and subtracts the averages. It is simple, but it assumes the worst day and the worst delivery happen together, and the buffer it gives is not tied to any stockout risk. The service-level formula sizes the buffer to the risk you choose.

Related tools

All free tools

Next: which channels bring your orders?

Daily sales are the number everything here rests on. What drives those sales, channel by channel, is the next question. Your GA4 export already holds how long your buyers take and which channels they touch. First finding free, in your browser; the full read is €99.

Get your first finding freeMore free tools