Safety Stock and Reorder Level: How They Work Together, With Real Numbers

Reorder Level and Safety Stock Are Not the Same Number

The reorder level (also called the reorder point) is the stock quantity that tells you to place a new order right now. The safety stock is one of the three numbers baked into that trigger — it's a cushion inside the formula, not a separate reorder rule of its own.

The formula that ties them together: reorder level = (average daily unit sales × lead time) + safety stock. Get the first two inputs right and safety stock still has to absorb everything you didn't predict — a supplier running four days late, a listing that suddenly ranks and starts selling twice as fast, a carrier missing an appointment at the FBA dock.

Confuse the two and you'll usually make one of two mistakes: setting your reorder level too low because you forgot to add safety stock at all, or setting safety stock so high 'to be safe' that it quietly turns into dead cash sitting in a warehouse.

The Formula, With Real Numbers

Say a product sells 18 units a day on average, based on the last 90 days of sales — not one good week. Lead time, measured from PO placement to units checked in and sellable at FBA, has averaged 21 days over the last six shipments, with the slowest at 27 days.

Safety stock using the lead-time variability method: (max lead time − average lead time) × average daily sales = (27 − 21) × 18 = 108 units. That's the buffer.

Reorder level = (18 × 21) + 108 = 378 + 108 = 486 units. That's the number that should trigger the reorder alert — not the 378 you'd get if you forgot the safety stock term. The 108-unit gap, on a slow boat from an overseas supplier, is roughly a week and a half of extra cover.

Which Safety Stock Formula Should You Use

There isn't one correct safety stock formula — there are three, and each answers a different question:

• Basic / 50% rule: half of average lead-time demand. Fast to calculate, fine for stable, low-value SKUs, wrong for anything seasonal or newly launched.
• Lead-time variability: (max lead time − average lead time) × average daily sales. Good when your supplier or freight is the unpredictable part — the example above.
• Demand variability (Z-score): Z × standard deviation of demand × the square root of lead time. Better when the product itself swings — a seasonal spike, a viral mention, a competitor going out of stock. It needs enough sales history for a standard deviation that actually means something.

Picking the wrong method for the situation is a bigger error than any rounding inside it. A brand with volatile demand using the 50% rule will stock out during every spike; a brand with stable demand using the Z-score method at a high service-level target will carry safety stock it never needed.

When the Number Turns Out Wrong

If you followed the formula and still stocked out, the formula probably isn't the problem — one of its inputs went stale. Check in this order:

• Lead time drifted and nobody re-ran the number. A supplier that took 18 days a year ago and now reliably takes 24 will blow through safety stock every cycle, because the reorder level was calculated on the old, shorter figure.
• Average daily sales moved while the input stayed frozen at last quarter's number, especially after a price change, a listing update, or a review milestone that shifted conversion.
• Safety stock was set once and never revisited. It isn't a constant — it needs the same refresh cycle as the sales average.

If the number itself was right and you still stocked out, look at the trigger, not the math: was the reorder actually placed the day the level was hit, or did it sit in an approval queue for a week? A correct reorder level with a slow purchasing process behind it fails the same way a wrong reorder level does.

The Most Common Mistake (We've Made It Too)

The mistake we see most — and one we've made ourselves early on — is calculating lead time from the supplier's quoted number instead of the actual, measured one. A supplier quotes 15 days. Actual time from PO to sellable FBA inventory, including production, freight, customs, and check-in, runs 24. Every reorder level built on the quoted number is short by nine days of coverage, and it only shows up as a stockout weeks later, disconnected from the decision that caused it.

Across the $500M-plus in managed revenue we've reviewed across 100+ brands, that gap — quoted lead time versus measured lead time — is the single most common reason a reorder level looks fine on a spreadsheet and fails on the shelf. It isn't a formula error. It's a stale input treated as a constant.

Where This Fits Into Amazon Inventory Management

Nothing above requires software — it's arithmetic, and a spreadsheet built on clean sales and lead-time history will get you a defensible reorder level and safety stock number. Where it breaks down is at scale: fifty SKUs, seasonal shifts, and three suppliers whose lead times move independently of each other, tracked by hand, tend to go stale between reviews.

Dr. Stock, from Fable 5 (part of Full Circle), watches that drift on Amazon-specific inventory — reorder timing, storage and aged-inventory charges, the removal-versus-liquidation call, and FBA fee or reimbursement errors — and flags it before it becomes a stockout or a cash-flow problem. Purchasing decisions always go to a human, whatever autonomy setting the client chooses. If the leak is in ad spend rather than inventory timing, that's a question for Dr. PPC, not this one.