What holding stock actually costs
Ask a seller what storage costs and you get the warehouse invoice. That is usually a quarter of the real figure. The rest is invisible because nobody sends a bill for it — and it is the part that decides whether a bulk discount was worth taking.
The four parts
Holding cost is conventionally expressed as a percentage of average stock value per year. For a small e-commerce operation it usually lands between 20 and 30%. It breaks down like this:
| Component | Typical | What it is |
|---|---|---|
| Capital | 6–15% | What the money would earn elsewhere, or what the overdraft costs |
| Storage | 2–8% | Rent, shelving, the warehouse invoice, FBA long-term fees |
| Risk | 6–12% | Obsolescence, damage, theft, expiry, markdowns to clear |
| Service | 1–3% | Insurance, counting, system overhead |
Storage — the only one you get an invoice for — is the smallest item on the list in most catalogues. That is why the intuition “the warehouse is cheap, so more stock is nearly free” is wrong so often.
Capital is not free even when it is yours
The commonest objection is that money already in the business costs nothing. It costs whatever it would have earned in its next best use — the advertising you did not buy, the supplier discount you could not prepay for, the loan you had to take because the cash was in cartons.
A practical rule: if you are borrowing, use the interest rate you actually pay. If you are not, use the return your marketing generates. A shop turning 1 € of advertising into 3 € of contribution has a very expensive warehouse, because every euro in stock is a euro not spent on customers.
Why it changes what you order
A supplier offers 8% off for taking six months of stock instead of two. The discount looks obvious.
Take 20 000 € of goods sold evenly over a year. Ordering every two months leaves an average of about 1 700 € on the shelf; every six months, about 5 000 €. At a 25% holding cost that is 425 € a year against 1 250 € — an extra 825 € to save 8% of 20 000 €, which is 1 600 €.
Still positive, but the margin is half what it looked like, and it assumes you sell all of it. Add one slow-moving line marked down to clear and the deal is gone. This is the arithmetic behind the rule that discounts on slow movers are usually a trap and discounts on fast movers usually are not — the same reasoning as in what a discount really costs, applied to buying instead of selling.
Where it shows up in the other numbers
Safety stock. Every unit of cushion carries the holding cost for as long as it sits there. That is the real price of raising a service level from 95 to 99.9% — the reorder point calculator shows how much stock the last few points buy.
Lead time. Shortening delivery reduces both cycle stock and safety stock. Air freight often looks indefensible per kilo and defensible per year once the frozen capital is counted — particularly for high-value, low-weight goods.
Product costing. Holding cost belongs in the landed cost of a unit, not in general overhead. Left in overhead it is spread evenly across fast and slow lines, which hides exactly the products that are eating the money — see how to cost a product.
A number you can use tomorrow
You do not need a precise figure. Take 25% as a working assumption, or build your own: cost of capital, plus warehouse cost divided by average stock value, plus last year’s write-offs and clearance markdowns divided by the same.
Then apply it as a monthly rate — 25% a year is about 2% a month — and ask of every purchase how many months of cover it buys. A line with eleven months of stock is carrying roughly a fifth of its own value in cost before it sells.
Frequently asked questions
Yes, and the storage part is easier to see because it is itemised — including long-term storage surcharges, which are deliberately punitive. The capital and obsolescence parts are identical to a warehouse of your own.
Not into gross margin, which should stay comparable across products. Put it in the contribution view, per unit and per month held. That is where it changes decisions.
Almost everything does commercially, even when it does not physically. Packaging changes, a newer model arrives, the listing drops in the ranking. Very stable goods justify a lower risk component, not a zero one.
The classic EOQ formula balances ordering cost against holding cost. It is a reasonable starting point, but it assumes steady demand and no quantity discounts, so treat its answer as a magnitude rather than an instruction.