The Cost of Stockouts vs Overstock
Two failures with opposite symptoms and one shared cause. Why the stockout is almost always more expensive than the overstock — and why every business over-corrects toward the overstock anyway, because only one of the two arrives as a number on a report.
Overstock has an invoice. Somebody bought it, the cash left, the goods are on the shelf, and it appears in your stock valuation every month as a slightly uncomfortable number. Stockouts have no document at all. A customer asked for something, you did not have it, and no record was created anywhere — which is why every business you will ever visit is carrying more overstock than it should and cannot tell you what its stockouts cost.
This asymmetry is not a measurement quirk. It is the mechanism by which working capital gets quietly consumed, because the visible error gets managed and the invisible one gets repeated.
What a [stockout](/glossary/stockout) actually costs
Four layers, and most businesses count only the first — which is why the number always comes out too small to justify acting on.
The four layers of a stockout
The lost margin on that sale
The margin on the units the customer wanted. Real, straightforward, and the smallest of the four.
The rest of the basket
The customer came for one thing and would have bought four. A missing item does not lose one line, it loses a visit — and in retail and wholesale that multiplier is typically two to four times the missing line itself.
The next visit
A customer turned away twice learns that you might not have it. They do not announce this. They simply check somewhere else first, and the loss is now every future basket rather than this one.
The emergency purchase
Somebody buys it locally at retail, or pays for expedited delivery, or splits an order across two suppliers to get it faster. That premium is real, is usually 15–40% over your normal cost, and is almost never attributed to the stockout that caused it.
One stockout on a fast-moving line, fully counted
Against a carrying cost of maybe KES 500 a month to have held two weeks more of that item. Which is the whole argument: on fast-moving lines the arithmetic is not close, and businesses under-stock them anyway because the carrying cost is visible and the stockout cost is not.
What overstock costs, honestly
Less than the finance-textbook figure and more than nothing. The classic 20–30% annual carrying cost assumes warehousing, insurance and capital costs that most Kenyan SMEs do not incur separately — the store is already rented, the shelf is already there.
| Component | Real for an SME? | Rough annual rate |
|---|---|---|
| Capital tied up | Yes, and it is the main one | 15–25% — your actual borrowing cost, or what the cash would have earned elsewhere |
| Obsolescence and expiry | Yes, and varies wildly by category | 0% on hardware, 100% on anything dated |
| Shrinkage and damage | Yes — more stock means more of both | 1–4% |
| Warehousing and insurance | Usually already paid for | Near zero at the margin, until you need more space |
| Handling and counting cost | Yes, quietly | More lines means longer counts and more errors |
So somewhere between 18% and 30% a year for most goods, and effectively total loss for anything with a date on it. Which is real money and is still, on fast-moving lines, smaller than the stockout it prevents.
Where the two costs actually cross
Which error is cheaper, by line type
Fast-moving staples
High basket effect, cheap to hold, predictable demand. Never be out of these. Hold more than the model says.
Critical spares and consumables
A missing part stops a machine or a job. The stockout cost is downtime, which dwarfs the holding cost of any spare.
Mid-range regular lines
The genuine balance point, and where a reorder point earns its keep. Roughly two-thirds of your item list.
Slow-moving, high-value
Order on demand. Holding one of these for eight months costs more than the sale you occasionally miss.
Anything dated or perishable
Overstock is total loss, not a carrying cost. Being out is annoying; expiring is a write-off with a name on it.
The practical consequence: a single service-level target across a whole item list is wrong at both ends. The A items should be over-stocked relative to any formula and the dated items under-stocked, and only the middle third deserves arithmetic.
The one measurement that changes behaviour
A tally sheet at the counter. Every time a customer asks for something you do not have, one line: date, item, roughly how much they wanted. Four weeks.
It is the cheapest and most persuasive measurement available in inventory management and almost no business does it, because the loss is invisible by nature and nobody thinks to make it visible by hand. After four weeks you have the number that no system will produce for you, and it typically reorders your priorities entirely.
What the reports here will and will not tell you
Dead stock and inventory ageing, turnover and days-on-hand, ABC ranking by consumption value, and a batch-expiry view with value at risk — all of those exist and all of them are about the overstock side. There is no stockout log anywhere, because a stockout leaves no record to log. The tally sheet is not a workaround for a missing feature; it is the only possible source of that data.
What to do with the two numbers
-
Rank the item list by consumption value
The ABC report bands items on cumulative share of consumption value — the lines making up the first 80% of spend, rather than the top 20% of rows. Those are your A items and they are where stockouts hurt.
-
Raise the reorder points on the A items, deliberately over-holding
Against whatever the formula suggests. On a fast-moving staple, two extra weeks of cover costs a few hundred shillings a month and prevents a loss of tens of thousands.
-
Order the slow-moving high-value lines on demand
Accept the occasional missed sale. Holding a KES 80,000 item for eight months is a worse trade than losing one sale a year on it.
-
Put anything dated on a shorter leash than feels comfortable
Expiry is total loss rather than carrying cost, and there is a daily expiry scan and a 30/60/90 report with value at risk — but nothing writes stock off automatically, which is deliberate. A write-off should have a name on it.
-
Re-read your tally sheet at 90 days
The same four weeks of counting, after the reorder points have been tuned. That comparison is the only honest verdict on whether any of this worked.
Our take
Count your stockouts for four weeks on a sheet at the counter — it is the only source of that number and it will change your priorities. Then deliberately over-hold your fast-moving staples and critical spares against whatever the formula says, order slow high-value lines on demand, and keep anything dated on an uncomfortably short leash. A single service-level target across a whole item list is wrong at both ends of it.
See the overstock, then count the stockouts
Dead stock and ageing, turnover and days-on-hand, ABC ranking on cumulative consumption value, and expiry with value at risk — plus an honest statement that the stockout side has to come from a tally sheet, because no system can log an absence.
See plans & pricingFrequently asked questions
Which is more expensive, a stockout or overstock?
On fast-moving lines, the stockout, usually by a wide margin — and businesses over-correct toward overstock anyway because only one of the two arrives as a number on a report. On slow-moving high-value lines and anything dated, overstock is worse, and for perishables it is total loss rather than a carrying cost. A single service-level target across a whole item list is wrong at both ends.
What does a stockout actually cost?
Four layers, and most businesses count only the first: the lost margin on that line; the rest of the basket the customer would have bought, typically two to four times the missing line; the future visits lost when a customer learns you might not have it; and the emergency restock premium of 15–40% over your normal cost, which is almost never attributed back to the stockout that caused it.
What is a realistic carrying cost for a Kenyan SME?
Somewhere between 18% and 30% a year for most goods, and effectively total loss for anything dated. The textbook 20–30% assumes warehousing, insurance and capital costs many SMEs do not incur separately — the store is already rented. The dominant component is capital tied up, at your actual borrowing cost or what the cash would have earned elsewhere.
How do we measure stockouts?
A tally sheet at the counter for four weeks: date, item, roughly how much the customer wanted. No system can produce this, because a stockout leaves no record — it is defined by an absence. It is the cheapest and most persuasive measurement in inventory management, almost nobody does it, and it usually reorders your priorities entirely.
What should we do differently once we have both numbers?
Rank the item list by cumulative consumption value, then deliberately over-hold the A items against whatever the formula suggests — two extra weeks of cover on a staple costs a few hundred shillings a month and prevents losses of tens of thousands. Order slow-moving high-value lines on demand and accept the occasional miss. Keep dated stock on a shorter leash than feels comfortable.
