Let me lay out two sets of numbers first. They're hypothetical — plug in your own, the logic holds.
Math one: a SKU sells 3,000 units a month at $8 margin each. Ocean transit runs 35 days; the vessel arrives 10 days late and you're out of stock for 10 days. 3,000 divided by 30, times 10 — 1,000 units you can't sell, $8,000 in margin gone. That doesn't count the ranking slide and the weeks of ad spend it takes to win the organic traffic back. The true cost of a stockout is always bigger than the spreadsheet says.
Math two: the same SKU, 10% more safety stock — 300 extra units. At $12 landed cost apiece, that's $3,600 in tied-up cash. At an 8% annual cost of capital, under $300 a year.
$8,000 of stockout risk against under $300 in carrying cost. Once that math is on the table, "how much safety stock" mostly answers itself: for most sellers the question was never whether to hold buffer, but how much counts as enough. Too much is waste. Too little is gambling.
The Formula, in Plain English
The textbook version: Safety Stock = Z × √(L × σd² + d̄² × σL²). Don't let the symbols scare you — it's three sentences.
First, safety stock covers two "what-ifs": demand running hotter than expected, and the vessel arriving later than expected. With 35 days on the water, there's no dodging either. When Red Sea diversions added roughly two weeks to transit times a couple of years back, sellers who had modeled only demand volatility watched their carefully computed buffers evaporate overnight.
Second, Z is the service-level factor — public statistical knowledge: 1.65 for 95%, 2.05 for 98%. The more certainty you want against stockouts, the bigger the Z, the more inventory you hold.
Third, more volatility means thicker buffers. A SKU moving a steady 100 units a day needs far less buffer than one swinging between 50 and 300 during promos. Step one is always pulling your last 60 days of daily sales and staring at the real volatility — never guess a number.
One trap I have to call out: "transit days" in the formula means door-to-stock days, not the number on the carrier's schedule. Vessel departure, berthing at LA/Long Beach, customs clearance, drayage, receiving, putaway — the clock stops when the unit hits the shelf. I've seen too many sellers plug 35 schedule days into the formula while reality was 42. Good math can't fix bad inputs. If you're unsure, pull your last few shipments' actual records and average them — that number matters more than the formula itself.
95% vs 98%: What Do Three Points Cost?
The question I hear most. Straight answer: moving from 95% to 98% takes Z from 1.65 to 2.05 — safety stock rises 24%. Note that's 24% more safety stock, not 24% more total inventory; with 35 days of cycle stock already in the pipeline, safety stock is only one slice. So don't flinch at "a quarter more" until you've computed the absolute number and seen whether it actually hurts.
My call: 95% is enough for most SKUs. Reserve 98% for two cases — true heroes, where one stockout day keeps you up at night, and anything in the 60 days before peak season, when going dark means handing your best weeks of the year to competitors.
The flip side: C-list long-tail SKUs are fine at 90%, even 85%. A few stockout days on the tail won't kill you. But dead stock sitting on paid warehouse racking bleeds every single day — your 3PL doesn't discount storage because your goods aren't moving.
A Per-SKU Method You Can Run Next Week
Don't set one flat "30 days of sales" buffer for the whole catalog. That's the most common guesswork I see. Do it per SKU — five steps, runnable next week:
Step one: export 60 days of daily sales per SKU. Compute the daily average (AVERAGE) and standard deviation (STDEV) in Excel. If the standard deviation looks ugly, that's reality talking.
Step two: pull your last six shipments' actual door-to-stock transit days. Average and standard deviation. If transit σ tops 5 days, your lane or forwarder is unstable — bump the buffer a tier.
Step three: pick Z. A SKUs get 1.65 (95%), B get 1.28 (90%), C get 1.04 (85%).
Step four: Safety Stock = Z × SQRT(transit days × σ_daily² + daily avg² × σ_transit²). One Excel row, drag down, whole catalog done.
Step five: reorder point = daily avg × transit days + safety stock. Hit that number, place the order, no hesitation.
The ABC playbook: A SKUs — your top 20% by revenue — get the 95% baseline, raised to 98% entering September. They earn the extra cash. B SKUs run at 90% on a normal replenishment rhythm. C SKUs get 85%; if they stock out, emergency air freight or small parcel from China, provided the margin covers the freight. If it doesn't, accept the stockout — or kill the SKU outright.
On peak-season ramp: my practice is adding one buffer tier every two weeks starting in September, maxed out before Black Friday. How much is your call — it depends on your cash flow. Don't copy someone else's factor; you won't be able to execute it anyway.
One practical detail most formulas miss: reconcile the math with shipping economics. If the formula says reorder 200 units but your consolidator's LCL minimum is 500, stock 500. A formula that ignores how freight actually ships is a hobby, not a plan.
And one habit that pays for itself: revisit the numbers monthly. Thirty minutes at the start of each month — actual stockout days, the three SKUs with the worst forecast error, whether transit σ has drifted. Safety stock isn't a compute-once-and-forget number; it's a living input. The sellers I see getting this right treat the spreadsheet as a monthly ritual, not a one-time project — fifteen minutes of discipline beats a quarter of firefighting.
One more rule I give every seller before they touch the spreadsheet: split your SKUs by sales tier before you compute anything. Run the full formula only for your A SKUs — the 20% of items driving 80% of revenue. For B SKUs, a simple rule of thumb (average daily sales × 50 days of coverage) is honest enough. For C SKUs, don't overthink it: one extra carton. I learned this watching a seller burn a full weekend running Z-scores for items selling nine units a month — precision he didn't need, for risk he didn't have.
The Question Every September
Same A SKU. Cash on hand covers safety stock at 95%, but your gut says Black Friday will pop. Do you spend tens of thousands more now to push coverage to 98% and bet against a stockout — or keep the cash and accept the risk?
There's no standard answer. But everyone running an overseas warehouse answers it by hand, every September. My advice is one line: run the formula first, decide with numbers, not feelings.
