Safety Stock Formula: The Simple Rule and the Statistical Method

Safety stock is the stock you keep so that a busy week or a late truck doesn't turn into an empty shelf. Too little and you stock out; too much and cash sits in the warehouse. There are two common ways to size it.

Method 1: Cover part of the lead time

Safety Stock = Average Daily Demand × (Lead Time × 50%)

This rule says: keep enough extra stock to cover half of the supplier's lead time at your normal selling rate. A SKU selling 21 units a day from a supplier with an 8-day lead time gets 21 × 4 = 84 units of safety stock.

It is easy to explain and needs only two numbers you already have. Its weakness is that it treats a steady SKU and an erratic SKU the same way when their average is the same.

Method 2: The statistical formula

The textbook formula sizes the buffer from how much demand and lead time actually vary:

Safety Stock = Z × √( L × σd² + d̄² × σL² )

Z   = service-level factor (1.65 for about 95%)
L   = average lead time in days
σd  = standard deviation of daily demand
d̄   = average daily demand
σL  = standard deviation of lead time in days

The first term inside the root covers demand swings during the lead time; the second covers the supplier arriving early or late. If your supplier is perfectly punctual, σL is zero and the formula reduces to Z × σd × √L.

Worked example

Illustrative numbers for one SKU at one warehouse:

d̄ = 20 units/day, σd = 6 units/day
L = 10 days, σL = 2 days, Z = 1.645 (95%)

L × σd² = 10 × 36 = 360
d̄² × σL² = 400 × 4 = 1,600
√(360 + 1,600) = √1,960 ≈ 44.3
Safety Stock ≈ 1.645 × 44.3 ≈ 73 units

Notice that the lead-time term (1,600) dominates. For many distributors, an unreliable supplier costs more safety stock than unpredictable customers do. That is a good reason to measure supplier OTIF and lead-time slippage.

Which one should you use?

• Use the simple rule when you have little history, or you need a number everyone on the team can check by hand.
• Use the statistical method for fast movers with enough history to measure variability, and for suppliers whose delivery dates wander.
• Use neither blindly for intermittent SKUs (many zero days, occasional large orders). The standard deviation of mostly-zero demand is misleading; those SKUs need an intermittent-demand forecast such as Croston's method.

Common mistakes

• Using the supplier's quoted lead time instead of the one they actually deliver.
• Setting one safety-stock percentage for the whole catalog.
• Never revisiting it: demand and suppliers change, and safety stock should follow.

Supply Chain Manager recalculates safety stock and the reorder point for every SKU and warehouse from your own data and shows the inputs next to each recommendation. To see it on sample data, open the live demo.