Safety stock is the extra inventory you hold beyond expected demand, sized to absorb two kinds of surprise: customers buying faster than forecast and suppliers delivering slower than promised. The practical calculation for most sellers: worst-case demand during lead time minus average demand during lead time.
The simple method that covers most sellers
You do not need the z-score formula to get a working buffer. You need four numbers per SKU:
- Average daily sales (30-day window is fine).
- Maximum realistic daily sales (your best day in the last quarter, promotions included).
- Average lead time in days.
- Maximum realistic lead time (the slowest your supplier has actually been).
Safety stock = (max daily sales x max lead time) - (average daily sales x average lead time)
Worked example: a SKU averages 4 sales a day and peaked at 9. Your supplier averages 12 days and has slipped to 18. Safety stock = (9 x 18) - (4 x 12) = 162 - 48 = 114 units. That number will look high, it is pricing in both worst cases hitting at once. Many sellers deliberately hold a fraction of it based on how painful a stockout is for that SKU versus how expensive the units are to hold.
The statistical method, briefly
The textbook formula is safety stock = Z x standard deviation of demand over lead time. Z is a service-level factor: 1.65 for 95 percent, 2.05 for 98 percent. It is worth adopting when you have clean daily sales history and enough volume for the standard deviation to mean something, usually hundreds of orders per SKU. Below that, the simple method plus judgment beats false precision.
What safety stock is protecting you from
- Demand spikes: a mention, a competitor stocking out, a marketplace featuring your listing. Ecommerce demand is spikier than retail because discovery is algorithmic.
- Supplier variance: production delays, shipping congestion, customs. Lead-time surprise is usually the bigger driver for imported goods.
- The compounding case: the spike and the delay together. That is the scenario the formula’s worst-case terms exist for.
What safety stock cannot fix: stale stock counts between your channels. If eBay thinks you have 5 when you have 1, no buffer at the warehouse prevents the double-sell, that is a sync problem, solved differently.
Multichannel: two different buffers, two different jobs
Sellers who run several channels need to separate two ideas that often get merged:
- Purchasing safety stock is physical: extra units in the warehouse, sized by the formulas above, protecting fulfillment.
- Listing buffer is virtual: advertising fewer units than you physically hold, so channels list 8 when you have 10. It protects against the sync gap on fast movers, the last-two-units race where oversells cluster.
They are complementary. Purchasing safety stock without a listing buffer still oversells at the thin tail; a listing buffer without purchasing safety stock just stocks out honestly. Unifystock handles the second kind natively: set a stock buffer once and every connected channel (Amazon, eBay, Shopify, WooCommerce, Etsy, OpenCart at launch) advertises the reduced number, while your true count stays intact in your store.
When to review the buffer
Safety stock is not set-and-forget. Re-run the numbers when:
- Velocity shifts by more than about a third in either direction, watch your days of cover trend as the early signal.
- Suppliers change, new supplier means new lead-time variance, and their promised lead time is not data yet.
- Season turns: the buffer that fit September is wrong for December in both directions.
- The channel mix moves: a SKU suddenly doing most of its volume on a new marketplace inherits that marketplace’s demand volatility.
Common questions
What is a good starting point if I have no sales history?
Hold two to four weeks of expected demand as your buffer, then replace guesses with data at the first review. New-product forecasts are wrong by default; the buffer exists to make that survivable.
Is more safety stock always safer?
No. Every buffered unit is cash you cannot spend and shelf space with a holding cost, and for seasonal or trend goods, an obsolescence risk. The right buffer is the smallest one that makes stockouts rare, not the largest one you can afford.
Should the buffer be the same on every channel?
Your listing buffer can be. Some sellers hold a larger buffer on the channel where penalties are harshest, Amazon metrics being the usual example. Unifystock applies your buffer to what channels see, so the policy is consistent instead of hand-managed.
How does safety stock interact with reorder points?
The reorder point is average demand during lead time plus safety stock. The buffer raises the trigger so that a late supplier or a strong week does not eat your last units before the purchase order lands.
Put the buffer where the risk is
Unifystock keeps every channel’s stock in step in real time, applies your listing buffer everywhere at once, and flags the thin-stock SKUs where buffers matter most. Plans start at $49/month with unlimited orders. See pricing.