By Safock Procurement Research Team8 min read
A clinic network uses about 18 cases of exam gloves a day, give or take 6. The supplier quotes a 7-day lead time, and in practice deliveries arrive anywhere from 5 to 9 days after the order. How much safety stock does it need?
A common shortcut formula gives 26 cases. With that buffer, about one order in four runs out before the next delivery arrives. The reason is that the formula only covers busy days. For this item, 84% of the risk comes from the supplier's delivery times, and that formula leaves it out.
When stock is checked continuously, or at least every day, the usual rule is to place an order when stock falls to a set level: the reorder point.
Reorder point = average daily usage × lead time + safety stock
The first part is what you expect to use while waiting for the delivery. Safety stock covers the times when usage runs higher than expected, or the delivery takes longer than quoted. It decides how often you run out.
Sites that order on fixed days, such as every Monday, set a par level instead: an order-up-to level that has to cover the lead time plus the days until the next delivery. Our guide to par levels covers that case.
Safety stock = z × √( L × σd² + d² × σL² )
The two terms under the square root are the two sources of risk. L × σd² is usage running high while you wait. d² × σL² is the wait itself running long, and it grows with the square of how much you use.
For the gloves, at 95%:
The delivery term is 1,296 out of 1,548, so 84% of the risk comes from the supplier's timing. With 160 cases on the shelf, stock is already below the reorder point, so the order goes in today.
Two shortcut formulas are common in guides and spreadsheets. Here is what each gives for the gloves, and how often that buffer would last until the delivery arrives.
| Method | Safety stock | Value | Lasts the cycle |
|---|---|---|---|
| z × usage variation × √lead time | 26.1 | $1,695 | 75% |
| 25% of lead-time usage | 31.5 | $2,044 | 79% |
| 50% of lead-time usage | 63.0 | $4,089 | 95% |
| Max usage × max lead time − average × average | 204.0 | $13,240 | Almost 100% |
| Full formula at 95% | 64.7 | $4,200 | 95% |
The first formula is right when deliveries always arrive exactly on time. Here they don't, so it covers only the smaller of the two risks. The "max × max" method, with maximums taken as two standard deviations above average, assumes the busiest week and the latest delivery happen together every time, and holds more than three times the stock the full formula needs. A percentage of lead-time usage can land close by chance, as 50% does here, but it has no connection to how variable the item actually is.
| Lead time varies by | Safety stock | Value | Reorder point |
|---|---|---|---|
| 0 days | 26.1 | $1,695 | 152 |
| ±1 day | 39.5 | $2,562 | 166 |
| ±2 days | 64.7 | $4,200 | 191 |
| ±3 days | 92.6 | $6,008 | 219 |
Halving the supplier's delivery variation, from ±2 days to ±1, frees $1,638 of stock. Removing usage variation entirely, which no site can do, would free only $357. On-time delivery is worth measuring and writing into a supplier SLA.
| Cycles without running out | z | Safety stock | Value |
|---|---|---|---|
| 90% | 1.28 | 50.4 | $3,272 |
| 95% | 1.645 | 64.7 | $4,200 |
| 97% | 1.88 | 74.0 | $4,803 |
| 99% | 2.33 | 91.5 | $5,940 |
| 99.9% | 3.09 | 121.6 | $7,891 |
The first few points of service are cheap and the last few are expensive. Choose the level per item, based on what running out actually costs. Gloves in a clinic justify a high level. A slow-moving spare part that can be couriered in a day may not.
For the gloves, a longer lead time barely moves safety stock: 61.6 cases at 3 days and 69.8 at 14. Most of the buffer here protects against variation in when deliveries arrive, and that doesn't grow with the length of the wait. The reorder point does grow, because more usage happens while you wait: from 116 cases at 3 days to 322 at 14.
Enter an item's usage, lead time, how much each varies, and what you have on hand. The defaults are the gloves.
Safety stock for one item · same maths as the full calculator
160 on hand is at or below the reorder point: order now.
Open with these numbersFree tool
Safety Stock & Reorder Point Calculator
Enter averages or paste a usage history, see safety stock, reorder point and par level side by side with the cost of each service level, or calculate a whole item list at once and download it as a CSV.
Runs in your browser. No sign-up.
Safety stock assumes the delivery arrives complete. If it arrives short, record it on a goods received note so the shortfall is chased and doesn't quietly eat the buffer.
Extra stock held above what you expect to use while waiting for a delivery, to cover days when usage runs higher than usual or the delivery arrives late. It is the buffer that decides how often you run out.
Multiply a service-level factor (z, such as 1.645 for 95%) by the variation in usage over the lead time. When delivery times vary too, that variation is the square root of lead time × usage variation² plus average usage² × lead-time variation². For 18 cases a day ±6, a 7-day lead time ±2 days and 95%, safety stock is about 65 cases.
Reorder point = average daily usage × lead time + safety stock. When stock on hand (plus anything already on order) falls to that level, place an order.
It depends on what running out costs. Each step up costs more than the last: in the example in this guide, moving from 95% to 99% adds about 40% to safety stock, and 99.9% nearly doubles it. Use higher levels for items that stop work when they run out.
Safety stock is the buffer itself. A reorder point is usage over the lead time plus safety stock, used when stock is checked continuously. A par level is the order-up-to level for sites that order on a schedule, and it includes its own buffer covering the lead time and the days between deliveries.