Stockout Cost Calculator

How the number is derived

Stockout cost is what running out actually costs: lost units × contribution per unit, plus the harder-to-measure consequences: substitution to a lower-margin product, an order cancelled entirely, a customer who does not return, and the search ranking or listing position lost while unavailable.

The direct term is easy and understates the total badly. The indirect terms are the reason understocking is more expensive than the arithmetic suggests, and they are why service levels sit as high as they do.

An example

The product contributes $40 a unit and sells 40 a week. A two-week stockout loses 80 units of demand directly, $3,200.

But not all of it is lost. If 30% of those buyers wait, 25% substitute to a product contributing $28, and 45% buy elsewhere: 24 units are deferred rather than lost, 20 units contribute $28 instead of $40 (a $240 shortfall), and 36 units are gone at $40, $1,440.

Direct cost is $1,680 rather than $3,200. Add the ones who bought elsewhere and did not come back: if 20% of those 36 were repeat customers worth $150 of future contribution, that is another $1,080, and the true cost is closer to $2,760.

The usual mistakes

Lost demand is invisible in the data. Sales during a stockout are zero and the system records zero demand, which means the forecast learns the wrong lesson and orders less next time, a feedback loop that turns one stockout into a pattern.

The marketplace effect is also larger than most sellers expect. On Amazon and similar platforms, going out of stock costs ranking that takes weeks to rebuild, so the cost continues well after the stock returns.

What this changes

Compare it directly against the carrying cost of the safety stock that would have prevented it. On the example, preventing this stockout might need 40 extra units costing $180 a year to hold, against a $2,760 event, which makes the buffer decision straightforward.

Then record stockouts explicitly, with dates, so the forecast can be corrected for them. Uncorrected, the demand history systematically understates true demand for exactly the products that sell best.

Why understocking usually costs more than overstocking

Overstocking costs carrying cost and eventually a markdown: a known, bounded, recoverable amount. Understocking costs contribution that is gone, plus customers who may not return, plus a forecast that has learned to order less.

On the example product, holding 100 extra units for a year costs $450 and, in the worst case, a 30% markdown on the residue. Running out for two weeks costs $2,760 and damages the demand signal.

That asymmetry is why service levels are conventionally set at 95% and above rather than at the point where the two costs look equal on paper. The visible cost of excess stock sits on the balance sheet where everyone can see it; the larger cost of shortage never appears in any account at all, which is precisely why businesses systematically under-buy.

The cost that never appears in a stockout calculation is the customer who found an alternative and did not come back. A lost sale is measurable and a lost customer is not, and in categories where switching is easy the second frequently exceeds the first. Which argues for setting service levels higher than a pure lost-margin calculation would suggest.