The number nobody writes down is the one that matters. A stockout produces no invoice, no adjustment, no line in a report. The order that would have arrived simply does not, and the only trace is a slightly flatter week that gets attributed to something else.
This page covers what a stockout actually costs once you count the orders you never saw, why the visible costs are the small half, and how to size it well enough to make a buffering decision with.
What is a stockout, and what does it cost?
A stockout is any moment a customer wants a unit you cannot supply. Its cost is the margin on the units you would have sold, plus whatever you spend recovering — expedited freight, a short run, a discount to win the customer back. The largest component is usually the demand you never observed.
That last point is what makes stockouts hard to manage rather than merely unfortunate. A write-off is visible: units, cost, a date. A stockout is an absence, and absences do not appear in reports. A brand can run out of its best SKU three times in a quarter, see revenue roughly hold because customers substituted, and conclude that stockouts are not really a problem — while the substitution quietly teaches those customers that the shelf is unreliable.
| Component | Visible? | Typical size |
|---|---|---|
| Lost margin on the sale | No — the order never existed | The obvious half |
| Expedite and recovery | Yes — it has an invoice | Small, and the only part most brands count |
| Substitution that sticks | No — looks like normal churn | Usually the largest, and the slowest to show |
| Retailer or channel penalty | Sometimes contractual | Varies, occasionally severe |
| Position in a channel's ranking | No | Compounds while you are out |
The third row is the one worth arguing about internally. A customer who wanted your product, found it unavailable, and bought a competitor's has not just cost you one sale — they have run an experiment on your behalf and got a usable result. Most of them come back. Some do not, and the ones who do not are invisible in exactly the same way the original lost order was.
How do you size it without perfect data?
Estimate the run rate immediately before the stockout, multiply by the days you were out, and apply your margin. It is deliberately rough. The purpose is not accounting precision but a number big enough to compare against the cost of carrying more stock, which is the decision it is meant to inform.
Two refinements make the estimate less wrong without much work. Use the run rate from the days just before, not a monthly average, because the item was probably selling well — that is often why it ran out. And count the days from when the last unit left, not from when somebody noticed, which is frequently a different date and always an earlier one.
Then resist the temptation to refine further. A stockout estimate accurate to the nearest few hundred is enough to answer the only question it needs to answer: is this bigger or smaller than what the extra buffer would cost me? Precision beyond that is effort spent on a number you are going to compare against another estimate anyway.
What is a backorder, and how is it different from a stockout?
A stockout is a sale you lose. A backorder is a sale you keep but cannot ship yet: the customer orders anyway, and you fulfil when stock lands. The difference in meaning is lost versus deferred, and that practical definition decides whether buffering is worth paying for.
The costs differ in kind, not just in size. A stockout costs margin you never earn. A backorder costs cash timing, a cancellation rate, and the support load of explaining a date you may miss. Split shipments add freight nobody planned for. None of it appears as a write-off, and all of it is real.
Whether a backorder actually saves the sale depends on the buyer. A committed one — a contract, a subscription, something made to order — waits, and backorders are cheap insurance. An impulse or easily substituted purchase does not wait. There the backorder converts into a cancellation, and you have paid the support cost on top of losing the margin. Where your cancellation rate on backorders runs high, hold the buffer instead. Where it runs low, accepting backorders is cheaper than carrying stock.
What do you do with the number?
Compare it against the cost of carrying the buffer that would have prevented it, and let the comparison set the buffer rather than instinct. For dated goods the comparison has a second term: the buffer that prevents the stockout may age out, in which case you would be paying for both failures.
That second term is what separates food and drink from the textbook version. Elsewhere the trade-off is stockout cost against holding cost, and holding cost is mostly capital. Here it is stockout cost against holding cost plus the probability the extra units expire, which is why the answer is often ordering more frequently rather than ordering more — the same conclusion the reorder point's shelf-life cap reaches from the other direction.
Where the number lands high and the shelf life is short, the honest fix is usually upstream: a shorter or more reliable lead time from the supplier does more than any buffer, because it shrinks the window the buffer has to cover in the first place. That is the pattern the operator’s playbook keeps running into: the fix usually sits upstream of where the pain shows.
What should you do next?
Price the last one, then use it to set the buffer.
Measure
- Find the last SKU that hit zero and for how many days.
- Take the run rate from the days immediately before.
- Apply your margin, and stop refining.
Compare
- Cost the buffer that would have prevented it.
- For dated stock, add the odds that buffer expires.
Act upstream
- Where the number is high and shelf life is short, work on lead time.
- Order more often before ordering more.
Frequently asked questions
What is a stockout?
A stockout is any moment a customer wants a unit you cannot supply. That includes shelf-empty at a retailer and a channel showing unavailable, not just a zero in your own warehouse — the customer experiences all three identically.
How do you calculate the cost of a stockout?
Start with lost margin on the units you would have sold, then add the recovery costs you actually incur: expedited freight, a short production run, or discounting to win the customer back. The hard part is not the arithmetic, it is the demand you never observed.
Is it cheaper to hold more stock or to run out occasionally?
Depends on the product's shelf life. For durable goods, holding more is usually the cheaper insurance. For dated goods it stops being true at the point where the extra buffer ages out, and then you are paying for both failures at once.