How to Size Safety Stock Properly

“Two weeks of cover” applied across the catalogue means the steady item with a reliable supplier is overstocked and the volatile item with an unreliable one still runs out.

The same rule produces both problems, because it ignores the thing that actually determines how much buffer is needed.

Buffer exists to absorb variability

Safety stock is not there for average demand. Average demand is covered by the normal reorder cycle.

It is there for the difference between what you expected and what happened during the replenishment lead time. Two things can differ:

Demand variability — sales during the lead time are higher than forecast.

Lead time variability — the delivery takes longer than expected.

An item with perfectly steady demand and a perfectly reliable supplier needs no safety stock at all. Every unit of buffer is paying for uncertainty in one of those two.

safety stock = Z × √(lead time × demand variance + demand² × lead time variance)

The formula looks unfriendly and its structure is simple: both sources of variability contribute, and they combine as a root sum of squares rather than by adding.

Lead time variability usually dominates

Look at the second term: demand squared multiplied by lead time variance.

Because average demand appears squared, lead time variability is amplified by the demand level. For a high-volume item, an unreliable supplier drives far more buffer than variable demand does.

That has a practical implication that gets missed. The cheapest way to reduce safety stock on a high-volume item is usually to make the supplier more reliable, not to forecast demand better.

A supplier whose lead time varies between 9 and 15 days is imposing a carrying cost on you. That is a negotiable commercial fact, and quantifying it turns a vague complaint into a number you can put in a conversation.

Service level is a cost decision

The Z factor comes from the target service level — the probability of not stocking out during a replenishment cycle.

It rises steeply at the top:

Moving from 95% to 99% costs roughly 40% more safety stock for a reduction in stockouts from one cycle in twenty to one in a hundred.

Whether that is worth it depends entirely on the item. For a component that halts a production line, yes. For one of forty variants of a slow-moving accessory, almost certainly not.

Applying one service level across the whole catalogue is the same error as applying one weeks-of-cover rule, and it is more common than it should be.

Segment before calculating

The practical approach is to segment the catalogue first.

High value, critical — high service level, accept the carrying cost.

High volume, standard — moderate service level, and focus effort on reducing lead time variability, which is where the leverage is.

Low volume, long tail — low service level, or make to order. Carrying buffer for items that sell twice a year is where a lot of dead stock comes from.

That segmentation does more for total inventory cost than refining the formula does.

Common mistakes to avoid

• One weeks-of-cover rule across everything.
• Ignoring lead time variability, which usually dominates.
• One service level for the whole catalogue.
• Recalculating annually when lead times have changed since.
• Treating supplier unreliability as fixed rather than negotiable.

How to do it with Inventory Buffer Calculator

The Inventory Buffer Calculator uses both variability sources.

1. Enter average demand and its variability, and the same for lead time.
2. Set a service level appropriate to that item, not to the catalogue.
3. Read which source is driving the buffer.
4. Where lead time dominates, take the number to the supplier.

Other inventory tools are in the tools directory.

Frequently asked questions

Why not just hold two weeks of cover?

Because a flat rule ignores variability. Two weeks is generous for a steady item with a reliable supplier and inadequate for a volatile one, so the same rule overstocks and understocks simultaneously.

Why is a 99% service level so expensive?

Because the safety factor rises steeply at the top of the distribution. Going from 95% to 99% costs roughly 40% more stock, which is worth it for critical items and rarely across a whole catalogue.

Which variability matters more?

Usually lead time, because it enters the calculation multiplied by average demand. On a high-volume item an unreliable supplier drives more buffer than variable demand does.

Final thought

Find out which variability is driving your buffer before adding to it. If it is the supplier’s, the cheapest fix is a conversation rather than a warehouse.

Try the free Inventory Buffer Calculator