Safety Stock Calculator
Compute safety stock and reorder point using the standard formula Z × σ × √LT, in units and in weeks of cover. Add lead-time variability for the full version. The result updates as you type.
Inputs
Apply this unit consistently to demand and lead time below.
Mean demand for this SKU over the recent window.
Standard deviation of demand. Rough proxy: (peak − trough) ÷ 4 over recent periods.
Time from PO to receipt. Same time unit as demand above.
Set to 0 to use the simple formula. Any positive value switches to the full formula.
How often you accept stocking out. 95% = once in 20 cycles. Higher = much more capital.
Result
How to use the calculator
Six inputs, one number. The hardest part is usually finding your demand variability; everything else is on a recent supplier email.
- 1
Enter your average demand
Use the time unit that's most natural for the SKU: daily for fast-movers, weekly for everything else. Whatever unit you pick, use it consistently for the variability and lead time too.
- 2
Enter your demand variability
Standard deviation of demand over the last 8 to 13 periods. If you don't have it, a rough rule: if demand swings by about ±X around the average most periods, σ is roughly X. If you run a forecast, a better input is the forecast's error (its average weekly miss × 1.25), because raw demand variability counts the seasonal swings your forecast already plans for. That's what Lumina uses.
- 3
Enter your supplier lead time
Average time from PO to receipt, in the same time unit as demand. If demand is daily, lead time is in days. If weekly, weeks.
- 4
Enter lead-time variability (if known)
Standard deviation of lead time across recent POs. Leave at 0 if you don't have the data and the calculator uses the simpler formula. Enter a value and it switches to the full formula automatically.
- 5
Pick your service level
95% is the standard: stock out one cycle in twenty. 99% costs about 41% more safety stock for the same SKU. Match service level to the commercial cost of a stockout for that specific SKU.
- 6
Read the result
The number is how many units of safety stock to hold, with the same buffer shown as weeks of cover. The reorder point shows when to trigger the next PO. Weeks of cover is the more useful form: a buffer in weeks scales with demand automatically, where a fixed unit count goes stale, and it's the form Lumina sets safety stock in.
The safety stock formula, briefly
The standard simple formula is:
Three inputs do the work. Z is the service-level multiplier: 1.645 for 95%, 2.326 for 99%. σD is the standard deviation of demand per period. LT is average lead time, in the same unit as the demand period.
When supplier lead time also varies, and it almost always does, the full formula accounts for both:
This calculator switches to the full formula automatically the moment you enter a lead-time variability above zero. For long-lead overseas supply chains, the full formula often produces materially higher safety stock. See lead time variability for why.
For the deeper explanation, formula derivation, and where the model breaks, read the safety stock glossary entry. Connected concepts: reorder point, demand variability, and lead time variability.
How Lumina sets safety stock
Lumina runs this same formula per SKU, with three differences that matter at catalogue scale.
- →Variability is the forecast's error, not raw demand. σ is taken from how far the forecast in use has been missing, not from how much demand swings. Raw demand variability counts the seasonal peaks and troughs the forecast already plans for, so for seasonal SKUs it over-buffers; forecast error measures only what the forecast can't see coming.
- →Service level follows the SKU's importance. Each SKU's target defaults from its ABC class (A 98%, B 95%, C 90%) and can be overridden per item, so bestsellers get the buffer they deserve and the tail doesn't tie up cash. An XYZ class (how predictable each SKU's demand is) sits alongside, so planners can see importance and variability together.
- →It's a recommendation, in weeks of cover, that you apply. Lumina shows the recommended buffer in weeks next to the current setting; planners review and apply it for the SKUs they accept. The system never silently changes your stock policy.
What changes the answer
Demand variability (σD)
The biggest single lever for most consumer brands. Halve demand variability and you halve required safety stock. The fastest ways to reduce it: a smoother promotional cadence and better forecasting on top sellers.
Lead-time variability (σLT)
Often the bigger lever and almost always the cheaper one to fix. Stabilising supplier lead times with predictable ordering cadence can reduce safety stock more than any forecasting improvement.
Service level (Z)
Non-linear. Pushing 95% to 99% costs 41% more safety stock for the same SKU. The cost of approaching 100% rises sharply; for most SKUs, 95 to 98% is the sweet spot. Match service level to the commercial cost of a stockout.
Lead time (LT)
Square-root scaling in the simple formula. Doubling lead time multiplies safety stock by √2, about 1.41. Big wins on lead time come from supplier or freight changes, not just buffer sizing.
Common safety stock mistakes
- →One safety stock figure across all SKUs. A flat “two weeks of stock” rule over-stocks slow movers and under-protects bestsellers. Compute per-SKU.
- →Picking a service level without costing it. 99% sounds prudent and costs about 41% more than 95% for the same SKU. Match service level to the commercial cost of a stockout, not a comfort default.
- →Letting safety-stock policy go stale. Demand variability shifts with seasonality and channel mix. Lead-time variability shifts with supplier behaviour. Recompute quarterly at minimum.
- →Treating safety stock as a target, not a buffer. Safety stock is what you can dip into during normal variance. If you never touch it, your buffer is too high.
Frequently asked questions
What does this safety stock calculator do?+
It computes the safety stock and reorder point for a SKU using the standard formula: Z × σ × √LT. If you also enter lead-time variability, it switches to the full formula that accounts for both demand and supply variance.
What's the formula behind the calculator?+
Two forms. Simple: Safety Stock = Z × σD × √LT (where Z is the service-level multiplier, σD is demand standard deviation, LT is average lead time). Full: Safety Stock = Z × √((LT × σD²) + (D² × σLT²)), which adds lead-time variability via σLT.
What service level should I target?+
95% is the default for most consumer brands: stock out one replenishment cycle in twenty. A good rule is to let it follow the SKU's importance: 98% for your A items (the bestsellers you can't afford to run out of), 95% for B, 90% for C and substitutable SKUs. That's how Lumina defaults it, per SKU, with overrides where you know better. Going from 95% to 99% costs about 41% more safety stock for the same SKU.
Should I use daily or weekly figures?+
Whichever fits the SKU's natural sales cadence. Use daily for fast-movers (50+ units/day) where the daily variance is meaningful. Use weekly for slower SKUs where daily numbers are too noisy. Consistency matters more than the choice: use the same time unit for demand, variability, and lead time.
What if I don't know my demand standard deviation?+
Quick approximation: take 8 to 13 weeks of demand, find the highest and lowest, and divide the range by 4. Better: compute the actual standard deviation in a spreadsheet (STDEV over the weekly figures). Best, if you run a forecast: use the forecast's average weekly miss × 1.25 instead of raw demand variability, because it ignores the seasonality the forecast already handles. Lumina does this automatically from its own forecast error.
Why does the result change so much when I bump service level from 95% to 99%?+
Because the relationship is non-linear. Z grows from 1.645 (95%) to 2.326 (99%), which is 41% more safety stock for the same SKU. Pushing to 99.5% adds another 11%. The cost of approaching 100% rises sharply; for most SKUs, 95 to 98% is the sweet spot.
Is this formula good enough for production planning?+
It's the right starting point and is what most ERP and planning tools use. The assumptions (normal distribution, independence between demand and lead time, stable variability) hold roughly for most consumer goods. Where they break, with highly seasonal or intermittent demand or very long leads, more advanced models help. For most scaling brands, this calculator is the floor, not the ceiling. Lumina runs the same formula per SKU with forecast error as the variability input, service levels that follow ABC class, and the result shown as weeks of cover for a planner to review and apply.