Home » Glossary Terms » Safety Stock

Safety Stock

Safety Stock - Featured Image
TL;DR
Safety stock is the buffer inventory you hold above expected lead time demand to absorb variability in sales and lead time. The classic formula uses a service-level Z-score multiplied by combined demand and lead-time variability. For FBA, lead time has to include FBA receive, and the buffer should be split between FBA on-hand and AWD-pullable.

Definition

The safety stock formula Amazon FBA sellers need isn’t padding for the average case. It’s protection against the cases where demand spikes OR lead time stretches OR both. Without it, your reorder point math is set up to stock out in roughly half of all reorder cycles, because actual demand is always either higher or lower than the average. The fastest way to set yours is with our free safety stock calculator.

The classic safety stock formula Amazon FBA operators use treats demand variability and lead-time variability as independent random variables and computes a buffer that holds against both. The Z-score reflects your target service level: 90% = Z 1.28, 95% = Z 1.65, 97% = Z 1.88, 99% = Z 2.33.

For FBA, you’re protecting against three specific failure modes: trailing demand running hotter than the average (a TikTok hit, a competitor stockout, a Q4 ramp), supplier delays (Chinese New Year, port congestion, factory quality holds), and FBA receive delays (Q4 backups can add 7-14 days to publish time).

Safety Stock Formula Amazon FBA

FORMULA (FULL)
Safety stock = Z × √( (LT × σdemand2) + (D2 × σLT2) )
where:
Z = service-level Z-score (1.65 for 95%)
LT = average lead time (days)
σdemand = standard deviation of daily demand
D = average daily demand
σLT = standard deviation of lead time (days)
// Simplified (when lead time variance is negligible):
// Safety stock = Z × σdemand × √LT

Example: a high-variability review-driven SKU

A seller with average daily demand of 24 units, daily demand standard deviation of 7 units (high variability, review-driven category), average lead time 60 days, lead-time standard deviation 8 days (occasional supplier delays), targeting 95% service level (Z=1.65):

σ_demand²          = 7² = 49
LT × σ_demand²      = 60 × 49 = 2,940
D²                  = 24² = 576
σ_LT²              = 8² = 64
D² × σ_LT²        = 576 × 64 = 36,864

Sum                 = 2,940 + 36,864 = 39,804
√39,804           = 199.5
Safety stock        = 1.65 × 199.5 = 329 units

So ROP = (24 × 60) + 329 = 1,769 units.

Now drop lead-time variability from 8 days to 3 days (better supplier or domestic switch):

σ_LT²              = 3² = 9
D² × σ_LT²        = 576 × 9 = 5,184
Sum                 = 2,940 + 5,184 = 8,124
√8,124            = 90.1
Safety stock        = 1.65 × 90.1 = 149 units

Cutting lead-time variability slashes safety stock by more than half. That’s why most FBA operators see big working-capital wins from switching to more reliable (often domestic) suppliers, not just from negotiating shorter lead times.

Why safety stock matters for FBA sellers

FBA’s dirty secret: receive variability can dwarf supplier variability in Q4. A shipment that “arrived” at the FC on Nov 10 may not be sellable until Nov 24. Build that into σ_LT, especially for Q4 cover. A second factor: split safety stock between FBA on-hand and AWD-pullable. AWD inventory bought into the safety-stock budget protects against demand spikes without paying full FBA storage rates.

The other reason this number matters: it’s the single biggest controllable component of working capital. Fast-moving SKUs with stable supply can run on 7-14 days of safety stock. Volatile SKUs with unreliable supply can need 45+ days. Mapping safety stock per-SKU to actual variance, instead of using a uniform days-of-supply rule, can free 20-30% of inventory cash.

Where this shows up in Profit Hawk
Profit Hawk computes safety stock per SKU using your actual demand and lead-time variance over rolling windows, and shows you how the number changes if you tighten supplier reliability or adjust service level. Start a free trial.

Common mistakes

  1. Using a single days-of-supply rule. “30 days of safety stock” sounds reasonable but ignores the underlying variability. A SKU with stable demand and reliable supply needs much less. A noisy SKU needs much more.
  2. Setting service level uniformly across the catalog. A $4 contribution-margin SKU at 90% costs less than a $40 contribution-margin SKU at 90%. Map service level to economic value, not category.
  3. Treating Q4 with last year’s safety stock. Demand variance widens in Q4. Recalculate σ_demand using only Q4-equivalent windows.

Related terms

Frequently asked questions

What service level should I target?

90-95% covers most SKUs. Push to 97-99% on flagship SKUs where stockouts compound losses (BSR drop, lost reviews, ad waste).

Should safety stock be in FBA or AWD?

Split it. The minimum needed to cover normal demand spikes goes in FBA. The deeper buffer (covering longer lead-time variability) sits in AWD where storage is cheaper.

How do I estimate demand variability with a short sales history?

For the first 90 days post-launch, use industry benchmarks for the category (often sigma is 30-40% of mean). After 90 days, recalculate from your own data on a rolling basis.

Does safety stock change with seasonality?

Yes. Both demand variability and lead-time variability can spike in Q4 (demand variability rises with promotional activity; supplier and freight delays widen). Calculate separate Q4 and non-Q4 safety stocks.

What's the trade-off between safety stock and stockout cost?

Higher safety stock costs cash and storage fees. Lower safety stock costs lost sales, BSR drop, and ad spend on out-of-stock units. The optimum balances expected stockout cost against carry cost; service level is a proxy for that balance.

Keep going

[ph_glossary_nav]

Nine free Amazon FBA calculators — plain English, no signup.