Safety stock is the buffer inventory you hold on top of expected demand, so a demand spike or a late delivery doesn’t immediately become a stockout. It’s insurance measured in units: too little and you lose sales; too much and you tie up cash.
Safety stock, in plain English
If demand were perfectly steady and suppliers perfectly punctual, you'd need zero buffer. Neither is true. Safety stock exists to absorb the two things you can't control: demand swinging above forecast and deliveries arriving late.
The standard formula
The classic textbook calculation multiplies a service factor (how rarely you're willing to stock out) by how much daily demand varies, scaled by lead time. At a 95% target the factor is 1.64; at 99% it's 2.33.
Worked example
A product sells 40 units/day on average, daily variation (standard deviation) is 15 units, the supplier takes 21 days, target service level 95%:
1.64 × 15 × √21 ≈ 113 units of safety stock. Try your own numbers in our free safety stock calculator.
The catch nobody mentions
The formula assumes demand follows a smooth bell curve. Real wholesale demand spikes, goes quiet and jumps when a big customer orders — so the formula tends to under-protect the erratic, high-stakes products and over-protect the stable ones. That's why more advanced operations size buffers from each product's real forecast-error history instead of an assumed curve.
Common safety stock mistakes
- One blanket service level for everything — best-sellers deserve more protection than slow movers.
- Setting it once and forgetting it — variability and lead times drift; recalculate quarterly.
- Computing variability from stockout-contaminated history — weeks at zero because you had no stock aren't "low demand".
Optimal Chain doesn't assume a bell curve. It forecasts every product, measures how wrong its own forecasts actually were for that product, and sizes the buffer from thousands of resamples of those real misses — exactly enough protection at your chosen service level, per SKU. You can see the reasoning for every number, and switch stocking policies (lean / balanced / safety-first) to watch the cash-vs-risk trade-off move in real time.