What is Safety stock?

How to calculate safety stock

The simplest, practical safety-stock method covers the gap between your normal and your worst-case demand during lead time. The formula is: (maximum daily sales x maximum lead time) minus (average daily sales x average lead time). The first term is your realistic worst case, the second is your expected case, and the difference is the buffer you need to survive a bad week and a slow shipment at the same time.

Worked example: a SKU averages 20 units a day on a 40-day lead time, but in a spike it sells 30 a day, and a delayed shipment can take 50 days. Worst case demand is 30 x 50 = 1,500 units; expected demand is 20 x 40 = 800 units. Safety stock = 1,500 - 800 = 700 units. That 700-unit buffer is what keeps you in stock when demand and lead time both move against you. Feed it straight into your reorder point so the trigger already accounts for the cushion.

Two inputs decide whether this calculation is useful or fiction. Maximum daily sales should come from your actual worst historical week, not a hunch, and not from a promotional day you'll never repeat. Maximum lead time should be measured from purchase order to the date units became sellable on Amazon, which includes port delays, customs, prep, and the receiving time at the fulfillment center. Sellers who measure lead time to the warehouse door instead of to sellable status routinely underestimate it by a week or more.

Recalculate when the inputs move. A supplier switch, a change of freight mode, a new prep partner, or a step change in velocity all invalidate the old number. Quarterly is a reasonable default cadence for a stable catalog, monthly for anything seasonal or fast-growing.

• •Safety stock = (max daily sales x max lead time) - (avg daily sales x avg lead time)
• •Example: max 30/day x 50 days = 1,500 units worst case
• •Expected: 20/day x 40 days = 800 units
• •Safety stock = 1,500 - 800 = 700 units
• •Reorder point = (avg daily sales x avg lead time) + safety stock = 800 + 700 = 1,500
• •Measure lead time to sellable status, not to warehouse arrival

How much safety stock should you hold?

There is no universal number, because the right buffer depends on two variables: how volatile a SKU's demand is, and how reliable its supply chain is. A steady-selling product from a domestic supplier with a short, predictable lead time needs very little safety stock. A seasonal hero product shipped from overseas, with lumpy demand and a long ocean lead time, needs a much bigger cushion, especially heading into Q4.

The trade-off is always against cash and storage. Every unit of safety stock is cash sitting in the warehouse, and on Amazon it can attract storage fees and, if it ages, aged-inventory surcharges. The goal is enough buffer to protect your in-stock rate on the SKUs that matter, without overstocking slow movers into long-term storage. Size it per SKU, weighted toward your highest-velocity, highest-margin products.

A useful sorting rule: rank SKUs by contribution dollars per day, not by units or revenue. The product generating $180 a day in contribution deserves a buffer that makes a stockout nearly impossible. The one generating $6 a day does not, and holding a deep cushion on it is how sellers end up with capacity limits full of stock nobody is waiting for. Amazon's storage limits make this concrete, because space you spend on slow movers is space unavailable to the SKUs paying your bills.

Safety stock and service level

More formal safety-stock models tie the buffer to a target service level, the probability you want of not stocking out during a given cycle. The higher the service level you target (say 95% vs 99%), the more safety stock you hold, because covering rarer and rarer demand spikes requires disproportionately more inventory. A statistical version multiplies a service-level factor by the standard deviation of demand over lead time.

Most SMB sellers do not need the full statistical model to make good decisions. The practical max-minus-average method above gets you most of the way, and the service-level mindset is the useful part: decide consciously which SKUs deserve a high in-stock guarantee and which can tolerate the occasional stockout, then spend your safety-stock cash where it earns the most.

The disproportion is worth internalizing, because it's where the money is. Moving from a 90% to a 95% service level costs some inventory. Moving from 95% to 99% costs a lot more, since you're now covering demand outcomes that occur rarely. On a product with thin margins, that last few percent of protection can cost more in carrying and storage than the stockouts it prevents. Reserve near-perfect availability for the SKUs where losing rank would be genuinely expensive to rebuild.

What the buffer actually costs: a worked example

Take the 700-unit buffer from earlier and price it. Say landed cost is $7.40 a unit, so the buffer represents $5,180 of capital held permanently, since safety stock by definition is inventory you don't plan to sell through. Imagine storage runs $0.10 per unit per month outside peak season, which is $70 a month or $840 a year, higher in Q4 peak-rate months. These are hypothetical numbers; check the current fee schedule for your size tier.

Add the opportunity cost. If your capital turns three times a year at 20% contribution, the $5,180 tied up in the cushion is forgoing roughly $3,100 of annual contribution it could have earned as working inventory. Total carrying cost of the buffer is therefore closer to $3,900 a year than the $840 the storage fee suggests.

Now compare that to what it prevents. If a stockout on this SKU costs you a week of sales at 20 units a day and $4 contribution per unit, one stockout is $560 of lost contribution, plus rank and ad-momentum damage that's harder to price but real. On those numbers the buffer pays for itself if it prevents around seven stockout weeks a year, which is plausible for a volatile overseas-supplied SKU and clearly not plausible for a stable domestic one. That comparison, done per SKU, is how you size a buffer honestly.

Notice the numbers you needed: landed cost, storage rate, contribution per unit, and realistic velocity. None of them live in Seller Central in usable form. They come out of your books, which is why safety-stock decisions tend to be as good as a seller's amazon fba accounting and no better.

Why safety stock shows up in your books

Safety stock is an operational buffer, but it is also working capital. The units you hold as a cushion are valued in inventory at their landed cost and only become COGS when they sell, so carrying more safety stock means more cash locked on the balance sheet and a thinner cash position even when the P&L looks healthy. That is why in-stock decisions and accounting are two views of the same thing.

Tracking safety stock against true landed cost shows you the real price of your cushion, not just the unit count but the dollars parked in the warehouse. BeanHawk values every unit at landed cost and tracks sell-through by SKU, so you can see which buffers are protecting profitable velocity and which are just slow-moving cash, keeping your inventory valuation and your books cash-accurate.

The pattern this creates catches new sellers off guard: a profitable, growing business with no money in the bank. Growth means larger buffers, larger buffers mean more cash converted into stock, and the P&L records none of that as an expense because it's an asset transfer. Reading profit as available cash is one of the fastest ways to run a healthy business into a cash crunch. Watch the cash conversion cycle alongside profit, especially through the pre-Q4 build.

One more accounting detail. Landed cost has to include freight, duty, and prep to be meaningful, and many sellers record inventory at supplier invoice only, pushing freight straight to expense. That understates inventory value, overstates current-period expenses, and makes every per-unit margin calculation optimistic. If your safety-stock math runs on those numbers, you'll consistently think your buffers are cheaper than they are.

Common mistakes in setting safety stock

Applying one blanket rule across the catalog is the big one. Thirty days of cover for everything means over-buffering steady SKUs and under-buffering volatile ones, and it manages to waste cash and cause stockouts simultaneously.

Using average lead time instead of worst-case lead time is the second. The whole point of the buffer is to survive the bad case, so a formula built on averages defeats itself. Keep a record of actual receipt dates per purchase order and use the slow tail of that data.

Third, forgetting that a promotion changes the maximum. A lightning deal or an ad push raises max daily sales for the period, and a buffer sized for normal demand will evaporate mid-campaign. Build the spike into the plan before the campaign, not after.

Fourth is a records problem rather than a maths one: counting units that aren't actually available. Inbound shipments not yet received, unfulfillable units, and stranded inventory all appear in some inventory views but can't be sold. A buffer calculated against a total that includes them isn't the size you think it is.

Tooling that makes buffers manageable

Below roughly thirty SKUs on one channel, a spreadsheet does this fine. Columns for average and max daily sales, average and max lead time, landed cost, and the resulting reorder point will give you everything in this article, and you'll understand the numbers better for having built it.

Past that, per-SKU maintenance is what breaks. Useful amazon inventory management software should pull real sales velocity, track actual lead times per supplier rather than an assumption, compute reorder points that already include the buffer, and value everything at landed cost so you can see the cash the cushion represents. If it shows units but not dollars, it will not help you make the trade-off in this article.

Multichannel sellers have an extra requirement. Buffers are meaningless if the same physical pool serves Amazon, Shopify, and eBay while each system holds its own quantity, so ecommerce inventory management software that maintains a single source of truth across channels is doing more for your in-stock rate than any forecasting feature.

On the accounting side, whatever you use should feed summarized settlement journals into QuickBooks or Xero and keep SKU detail in a subledger, so your inventory value on the balance sheet matches the units you actually hold. A2X and Link My Books are the common comparisons for the settlement-to-ledger piece; BeanHawk adds inventory valuation and reimbursement recovery. Run any candidate against one real month of your own data before you commit.

Frequently asked questions

What is safety stock in simple terms?

It is extra inventory you hold on purpose, above what you expect to sell, to avoid running out when demand spikes or a shipment is late. It is the buffer that sits underneath your reorder point and keeps a SKU in stock through the bad weeks.

How do I calculate safety stock?

A practical method is (max daily sales x max lead time) minus (average daily sales x average lead time). That difference is the buffer needed to survive a demand spike and a supply delay happening together. Calculate it per SKU and feed it into your reorder point.

How much safety stock should I keep?

It depends on demand volatility and supplier reliability. Steady SKUs with short, predictable lead times need little; seasonal products with long overseas lead times need much more. Weight your buffer toward your highest-velocity, highest-margin SKUs and avoid over-buffering slow movers.

Does safety stock cost me money?

Yes, indirectly. Every unit of safety stock is cash tied up in inventory and can incur Amazon storage fees, plus aged-inventory surcharges if it sits too long. The buffer is insurance against stockouts, so size it deliberately rather than over-ordering across the board.

How is safety stock different from a reorder point?

Safety stock is the buffer itself; the reorder point is the stock level that triggers a new order, and it includes the safety stock plus expected demand during lead time. In other words, safety stock is one of the inputs that builds the reorder point.

How often should I recalculate safety stock?

Quarterly works for a stable catalog, monthly for seasonal or fast-growing SKUs, and immediately whenever an input changes materially: a new supplier, a switch between air and ocean freight, a new prep partner, or a step change in velocity. The formula is only as current as the lead-time and demand data behind it.

Why is my business profitable but short of cash?

Often because growth is converting profit into inventory. Larger buffers and bigger reorders both consume cash without appearing as expenses, since inventory is an asset until it sells. Watching your cash conversion cycle alongside profit, and valuing stock at true landed cost, usually explains the gap.

What software should I use to manage safety stock and inventory value?

Choose something that computes reorder points from real velocity and real supplier lead times, values every unit at landed cost including freight and duty, and keeps one quantity across all channels you sell on. On the books side, look for summarized settlement journals into QuickBooks or Xero with SKU detail in a subledger. Compare BeanHawk, A2X, and Link My Books on those points and test each against a real month of your own data.

Related terms

Go deeper

• Ecommerce Inventory Management: Quantity, Value, Reorder Points