The Reorder Point Formula

It answers one question: at what stock level do I raise the purchase order? Everything else on this page exists to make the three inputs honest.

At 40 units a day with a 21-day lead time, demand during lead time is 840 units. Add a 300-unit buffer and the reorder point is 1,140 — 28.5 days of cover, of which 7.5 days is the buffer.

What Counts as Lead Time

This is where most reorder points go wrong, and it is not a subtle error — it is usually weeks.

Lead time is purchase order raised to units available to pick. For an imported product that is:

• Production at the factory
• Inland transport and export clearance at origin
• International freight
• Customs clearance at destination
• Port to warehouse
• Receiving, inspection and put-away — the step almost everyone forgets, and rarely less than a few days

Use the supplier’s quoted production time alone and your reorder point can be out by a month. Every stockout that follows will look like a demand forecasting problem and will not be one.

The Two Safety Stock Methods, and Why They Disagree

Most calculators pick one and never mention the other. On identical inputs they can differ by a factor of four, so it is worth knowing which one you are being handed.

On the worked example — 40 units a day with a standard deviation of 12, a 21-day lead time varying by 4 days, a busiest day of 65 and a worst-case lead time of 30 — the statistical method at a 95% service level returns about 278 units. Max-minus-average returns 1,110.

Neither is wrong. The max method insures against your busiest day and your slowest shipment landing together for the whole lead time, which is unlikely and expensive to hold — at $9.89 a unit and a 25% holding rate, that difference is about $2,060 a year of holding cost per SKU. The statistical method prices the risk instead of eliminating it.

A third figure worth seeing: the simplified formula that counts only demand variability returns just 90 units on the same inputs. It is the version most often quoted online, and it quietly assumes your supplier is never late.

Service Levels and What They Actually Cost

Service level is the share of replenishment cycles you expect to survive without a stockout. It enters the formula as a Z value:

• 90% → Z = 1.28
• 95% → Z = 1.65 — the usual default for core lines
• 97% → Z = 1.88
• 98% → Z = 2.05
• 99% → Z = 2.33
• 99.9% → Z = 3.09

Safety stock scales linearly with Z, so 95% to 99% costs about 41% more buffer and 95% to 99.9% costs 88% more. The right level is not the highest you can afford — it is the point where the cost of a stockout exceeds the cost of holding the stock that prevents it. That answer differs by SKU, which is why a single company-wide service level is usually leaving money in two places at once.

EOQ — and When to Ignore It

D is annual demand in units, S is what it costs to place and receive one order, H is what it costs to hold one unit for a year. At 14,600 units a year, $450 an order and a $9.89 unit at a 25% holding rate, EOQ is about 2,305 units — roughly 6.3 orders a year, every 58 days, tying up $22,800 of cash per order.

At the EOQ, annual ordering cost and annual holding cost are equal. That is the whole idea: it is the point where the two curves cross.

It is a reference point, not an instruction. EOQ assumes steady demand, a flat unit cost and no constraints, and real ordering has none of those. Supplier minimums, container fill, price breaks, cash position and shelf life all legitimately override it. What it is genuinely useful for is catching two expensive habits — ordering far too often, and buying a year of stock to chase a discount that costs more in holding and cash than it saves on the unit.

Why This Belongs Next to Your Ad Account

Inventory planning looks like an operations problem until you are paying to sell something you cannot ship.

• A stockout on a scaling SKU is a paid media problem first. Spend keeps running, conversion rate collapses, and the algorithm learns from a broken signal. The damage outlasts the stockout.
• Safety stock is capital, and capital is fungible with ad spend. Holding $11,000 of buffer at a 25% carrying cost is $2,750 a year that could have been acquisition budget. Worth deciding, not defaulting.
• Lead time sets how far ahead your growth plan has to be right. If you need 21 days to restock and you are planning a promotion, the inventory decision comes before the media plan, not after it.
• Order quantity moves your landed cost. Freight and fixed import fees spread across more units, so the EOQ you choose changes the unit economics your landed cost and break-even ROAS are built on.

What This Calculator Cannot Tell You

• It assumes demand is roughly stable. The statistical method rests on normally distributed demand. A launch, a promotion or a seasonal peak breaks that assumption — plan those explicitly rather than trusting a reorder point through them.
• It does not know your supplier minimums or container sizes. The EOQ it gives you may be unorderable. Round to what you can actually buy.
• It treats all stockouts as equally costly. They are not. Losing a subscription customer is worse than delaying a one-off order, and that should change your service level by SKU.
• It does not model price breaks. If your supplier discounts at volume, the true optimum may sit above the EOQ — compare the saving against the extra holding cost and the cash it locks up.
• It cannot tell you whether your standard deviations are real. Rough inputs give rough answers. If you do not have the numbers, the max-minus-average method needs no statistics and is the safer starting point.