The formula this calculator uses
The standard safety stock calculation, taught in every operations course:
- Z — the service factor for your target service level. 95% → 1.64, 99% → 2.33. The higher the target, the bigger the buffer.
- σd — the standard deviation of daily demand: a measure of how much your daily sales swing around the average.
- L — lead time in days. The longer you wait for stock, the more demand swing you're exposed to — but note it grows with the square root, so doubling lead time doesn't double the buffer.
If your supplier's lead time also varies, the extended version adds a second term so late deliveries are buffered too:
where d is average daily demand and σL is the standard deviation of lead time. Fill in the optional field above and the calculator switches to this version automatically.
Worked example
A product sells 40 units a day on average, with a daily standard deviation of 15 units. The supplier takes 21 days. At a 95% service level (Z = 1.64):
Hold 113 units of buffer on top of the stock you expect to sell during the lead time (40 × 21 = 840 units), giving a reorder point of roughly 953 units.
The honest bit: where this formula breaks down
This calculator gives you the same answer every planning textbook — and most inventory software — would give you. But the formula has a hidden assumption baked in: it assumes your demand follows a smooth bell curve.
Real wholesale and distribution demand rarely does. It spikes when a big customer places a bulk order. It goes quiet for a fortnight. It jumps ahead of a season and dies after it. When demand has more extreme swings than a bell curve allows for — and in our experience with real catalogues, a large share of products do — the formula quietly under-protects exactly the products where a stockout hurts most, and over-protects the boring stable ones, tying up cash where it isn't needed.
Rather than assuming a curve, Optimal Chain forecasts every product, measures how wrong its own forecasts actually were for that specific product, and sizes the buffer from thousands of resamples of those real misses — enough to cover them at your chosen service level, no more. Products get the protection their real behaviour demands, not what a 1960s formula assumes. It's the method the giant supply chains use, at a price the mid-market can pay.
Practical tips for using the result
- Don't use one service level for everything. Segment: high targets (97.5–99%) for best-sellers, moderate (95%) for the middle, low (90% or less) for slow movers. Blanket 95% is the most common — and most expensive — mistake.
- Measure σd from at least 6–12 months of daily or weekly data, and strip out known one-offs (a single pallet order to a new customer isn't "variability" — it's a customer).
- Watch for stockout-contaminated history. If you were out of stock for three weeks, those zero-sales weeks will drag your average down and shrink the formula's buffer — the exact opposite of what you want.
- Recalculate quarterly. Demand variability and supplier lead times both drift; a buffer set last year is stale.