Interactive · Safety stock

How to calculate safety stock and the reorder point

You sell 40 units a day. Your supplier takes 7 days to deliver. So you reorder when stock hits 280, seven days of sales, and it should arrive just as you run dry.

Do that and you run out of stock roughly half the time. Not because the maths is wrong, but because 40 a day is an average. Some days you sell 55. Over a seven day wait, those busy days stack up and you are empty before the van arrives. The fix is a buffer sized to how much demand swings. That buffer is safety stock, and this page shows exactly how to size it.

Technical illustration of a row of warehouse stock bins, with a highlighted reserve section set aside at the back as a safety-stock buffer.
Safety stock is the reserve you set aside for the busy days. Size it too small and you sell out; size it too big and cash sits on the shelf. The rest of this page is how to get it right.

The simulator below runs a real, if simplified, stockroom. Move the inputs and watch the buffer, the reorder point and the stockouts recalculate the way they would in a working system.

The stockout, and the buffer that stops it

Here is the stockroom over 70 days. The blue line is your stock. It falls as you sell and jumps when a delivery lands. The dashed line is the reorder point, the level that triggers a new order. The shaded band is your safety stock.

Turn the service level down to 50% and the buffer disappears, so the line dips below zero on the busy stretches. Those red marks are stockouts, days you had orders you could not fill. Raise the service level and the buffer grows until the line stays up.

Stockroom · 70-day run
Stock on hand Reorder point Safety stock Stockout day
Safety stock66units
Reorder point346units
Fill rate, this runof demand met
The fill rate is what this particular 70-day sample delivered. It hunts around your service target, not on it, because one run is one roll of the dice. Press new demand sample to roll again.

The reorder point

The reorder point is the level that tells you when to order. It has two parts.

The first is the stock you expect to sell while the order is in transit: average daily demand times lead time. At 40 a day over a 7 day wait, that is 280 units. The second is the safety stock, the buffer for demand above average.

Reorder point = (demand × lead time) + safety stock
= (40 × 7) + 66 = 346 units
Order too late and the lead-time demand alone empties you. The reorder point moves the trigger up so the order is placed while there is still stock to sell.

The safety stock formula

Safety stock is not a guess or a flat "two weeks of cover". It is sized to three things: how much demand swings, how long you wait for stock, and how sure you want to be.

Safety stock = Z × σdemand ×lead time
= 1.65 × 15 × √7 = 66 units
Every term moves the buffer in the simulator above. Turn up the demand swing and the buffer grows. Shorten the lead time and it shrinks, but by the square root, not in step.

The three terms, in plain English:

  • Z is the service factor. It converts "I want to be in stock 95% of the time" into a multiplier. It climbs slowly at first and then steeply, which is why the last few points of service cost so much stock.
  • σdemand is the standard deviation of daily demand, a measure of how much it swings around the average. Steady demand has a small sigma and needs little buffer. Spiky demand has a large one and needs a lot.
  • √lead time scales the buffer to how long you are exposed. A longer wait means more days for demand to run high, but the risk grows with the square root of the wait, not in proportion to it.

Service level and the Z factor

Z is the one term people get stuck on. It is just a lookup from your target service level. Being in stock 90% of the time needs a Z of 1.28. Pushing to 99.9% needs 3.09, more than double the buffer for the last ten points.

Service levelZ factorSafety stock*
*Safety stock at each service level for the demand swing and lead time set in the simulator. The jump from 99% to 99.9% costs more buffer than the jump from 90% to 95%.

Where this breaks

On a spreadsheet, none of this updates itself. Sigma is calculated once from last year and left to rot. Lead times drift as suppliers slip. The reorder point is a static number in a cell that no longer reflects how the item sells. So the buffer is too thin on the fast movers, where it costs you sales, and too fat on the slow ones, where it ties up cash.

That is the failure OpsMavix removes. We build operations systems that recalculate demand variability and lead time from your own order history, set the reorder point per item, and flag it before stock runs out, so the buffer stays the right size on its own.

Common questions

What is the safety stock formula?

Safety stock = Z × the standard deviation of demand × the square root of lead time. Z is the service factor for your target service level (1.65 for 95%). Sigma measures how much daily demand swings. The square root of lead time scales the buffer to how long you wait for stock.

How do you calculate the reorder point?

Reorder point = (average daily demand × lead time) + safety stock. The first part covers what you expect to sell while the order is in transit. The safety stock covers demand above average. When on-hand stock falls to that level, you place the order.

What service level should I use?

Most businesses run 90 to 98%. Higher levels cut stockouts but need far more stock, because Z climbs steeply near 100%. Set high-margin or critical items higher, and cheap, easily substituted items lower.

What is the difference between safety stock and the reorder point?

Safety stock is the buffer that absorbs demand above average, a quantity you hold. The reorder point is the trigger level you watch, and it already includes the safety stock plus the stock you expect to sell during the lead time.

OpsMavix builds custom operations and inventory systems for UK manufacturers, wholesalers and distributors.

Book a free Operations Leak Audit, or try the inventory reorder calculator and the interactive bill of materials guide.

This simulator uses a single-item, continuous-review model with normally distributed daily demand and a fixed lead time, the standard textbook case. Real demand can be seasonal or lumpy, and lead times vary too, which raises the buffer further. The arithmetic shown is exact for the model.