What is the safety stock formula?

When both sales and supplier lead times vary: safety stock = z × √(L × σd² + d² × σL²), where z is the z-score for your service level, L the average lead time in days, σd the standard deviation of daily sales, d the average daily sales and σL the standard deviation of the lead time.

What service level should I use?

95% (z ≈ 1.64) is a sensible default for most products. Use 98–99% for best sellers where a stockout costs the most, and 90% for slow movers that are cheap to be out of.

How is safety stock different from the reorder point?

Safety stock is the buffer. The reorder point is the stock level that triggers a new order: average daily sales × lead time + safety stock.

Free tool

Safety stock calculator

Work out how much buffer stock to hold when both your sales and your supplier's delivery times vary. Compare the three common methods, see the reorder point each one gives you, and what a higher service level costs. The link keeps your numbers; nothing leaves your browser.

The safety stock formula

Safety stock = z × √(L × σd² + d² × σL²)

z: the z-score for your service level (1.28 for 90%, 1.64 for 95%, 1.96 for 97.5%, 2.33 for 99%).
L and σL: average supplier lead time in days, and how much it varies.
d and σd: average daily sales, and how much they vary from day to day.

The first term inside the root covers sales running hotter than usual while you wait. The second covers the supplier arriving late while you keep selling. If your supplier is always on time (σL = 0), the formula reduces to the familiar z × σd × √L.

Example

You sell 12 units a day, give or take 4, and your supplier takes 21 days, give or take 4. At a 95% service level: 1.64 × √(21 × 4² + 12² × 4²) = 1.64 × √2,640 ≈ 85 units. Ignore the supplier's variation and you would hold only about 31, which is why late deliveries are a common cause of stockouts even when sales forecasts are right.

Which method should you use?

Max − average: needs no statistics, but assumes your worst sales day and slowest delivery happen together, so it usually over-stocks.
Demand variability only: fine when your supplier is reliable.
Demand + lead-time variability: the best default when deliveries vary, and the method Reorder Radar uses.

Getting good inputs

Measure daily sales only on days the product was in stock. Stockout days show zero sales and make demand look lower and steadier than it is.
Measure lead time from the day you place the order to the day stock is ready to sell, across at least three recent deliveries.
With less than about two weeks of in-stock history, variation can't be measured reliably: a flat 25% of lead-time demand is a safer temporary rule.

More detail in the safety stock guide. To turn safety stock into an order date and quantity, use the reorder point calculator.

Want this for every product?

Reorder Radar works out safety stock for every SKU from your Shopify data, excluding stockout days and learning each supplier's real lead-time variation from your received purchase orders. The Shopify App Store listing is in review: leave your email for the install link, or look around the live demo first.