Ask an owner what their inventory costs and they will tell you what they paid for it. That number is on every invoice, so it gets all the attention. The cost of keeping it, per unit, per year, shows up nowhere, which is how a real expense ends up run by a shrug: somewhere between 20% and 30% of what the thing cost you, probably, whatever.
The guess is respectable. It exists because the real components, added up, usually land in that range. But the range is doing a lot of work. On a storeroom worth of stock, the gap between 20% and 30% is real money, and every inventory decision you make, how much to order, how big a buffer to carry, whether that volume discount is actually a deal, moves when this number moves. It deserves twenty minutes once, instead of a shrug forever.
The number is a stack, not a fact
Carrying cost is several unrelated costs wearing one name. Price each one for your own shelf and the hand-wave disappears.
The money itself. Stock is cash in a less useful shape. What that cash would otherwise be doing sets the rate, and for many small businesses the honest answer is paying down a line of credit at 11%, which makes 11% the starting floor, before a single box is stored anywhere. If inventory sits on borrowed money, holding it is a financing decision that never got made out loud.
Space. A storeroom you already rent feels free, and mostly is, right up until it isn't. The test is the alternative use. Shelving in a basement that would otherwise hold nothing costs you close to zero. A back room that could be a second treatment chair, a private-dining corner, or a rentable bay is charging you its lost use, and that charge belongs on the stock that fills it.
Spoilage and shrink. Whatever expires, breaks, walks off, or gets damaged in handling, as a share of what moves through. A bakery's flour and a bar's citrus run real numbers here. A fastener aisle runs nearly none.
Insurance and the paperwork of owning things. Usually small, rarely zero. Your policy prices your stock; your accountant counts it once a year and bills for the counting.
Going out of style. The quiet killer for anyone seasonal. An outdoor shop's $100 jacket is a $100 jacket until March, and then it is a $60 jacket, and the markdown was a carrying cost the whole time, accruing invisibly while it hung there. Fashion, technology, and anything tied to a season all pay this one. Propane does not.
Stack a plausible outdoor-shop version: 11% for the money, 4% for the space, 2% shrink, 1% insurance, and 8% for the March markdowns spread across the stock that risks them. That is 26%, squarely inside the folk range, which is why the folk range survives. But notice how it got there. A propane dealer running the same exercise might land at 14%. A boutique with a fast season and an expensive lease can clear 35%. Same shrug, very different receipts.
What changes when you know it
Plug your real rate into the ordering math and things move in useful directions. A higher rate than you assumed means stock is dearer to hold than you thought, so the right move is smaller orders, placed more often, and suddenly the delivery fee you have been avoiding looks cheap next to the shelf you have been filling. It also means the volume discount your supplier dangles is weaker than it looks, because the extra units keep costing you money after the discount is spent. A lower rate runs everything the other way.
The sensitivity is gentler than you might fear, which is worth knowing before you agonize. The ideal order size moves with the square root of the carrying rate, so even doubling the rate shrinks the right order by about 30%, and modest errors in your estimate barely move it. All the exercise asks is that a guess spanning a 50% range become an estimate that is honestly yours, and the decisions downstream mostly need the neighborhood, plus a reason to believe it.
There is one carrying cost this arithmetic deliberately leaves out: your own hours. The counting, the reordering, the shuffling of boxes to reach other boxes. That cost is real, it is usually bigger than the shrink line, and it belongs to a different argument about where cut costs actually go, because it behaves differently. The components above scale with the stock. Your hours scale with how the stock is managed, which is a thing you can change without holding one unit less.
Twenty minutes, once
Pull a year of numbers for one category of stock. What you paid to borrow, or what your idle cash earned. The space's next-best use, honestly, even if the answer is nothing. Whatever you wrote off, marked down, or quietly threw away. Add the shares. That percentage is your number, and it will keep being your number until the lease, the loan, or the season changes.
Then hand it to the calculator in place of the guess and watch what shifts. For most businesses the order sizes move a little. For a few, the whole posture changes, usually the ones carrying seasonal goods on borrowed money, where the real rate was sitting near 35% while every decision assumed 20. Those owners were not careless. The number just lived in five different places, and nobody had ever added it up.