Inventory Carrying Cost: How to Calculate Your Own Rate

You already paid your carrying cost last year. It is spread across a bank statement, a storage unit invoice, an insurance renewal, the hours somebody spent receiving and counting, and a pallet of something you eventually cleared below cost. Because it arrived in six places, nobody added it up, and it never made it into a single ordering decision.

This post is about doing that addition. The site's guide to the inventory metrics worth tracking summarises carrying cost in a paragraph; this is the long version, and it is mostly a question of which lines from your own accounts belong in it. What to do once the total makes you uncomfortable is a different job, covered in releasing the cash tied up in stock you already own.

What carrying cost includes

Carrying cost is what it costs to hold stock for a period, on top of what you paid for the stock. It accrues while units sit, which is what makes it easy to ignore: nothing happens when the pallet arrives, and then something small happens every week it does not leave.

The standard operations-research definition is narrower than most blog versions of it:

"The holding cost (sometimes called the storage cost) represents all the costs associated with the storage of the inventory until it is sold or used. Included are the cost of capital tied up, space, insurance, protection, and taxes attributed to storage."
Hillier and Lieberman, Introduction to Operations Research, chapter 19.

Practitioners group the same ground into four buckets, and this post uses those: capital, storage, service and risk. The first two are in that definition verbatim. Service covers insurance, protection and tax, plus the labour of handling and counting. Risk covers what units lose in value while they wait, which the textbook definition leaves implicit and which does most of the damage in a small catalogue.

Three things get miscounted into it and do not belong: the purchase price of the stock, which is the investment being carried rather than a cost of carrying it; inbound freight, conventionally folded into unit cost already; and the pick, pack and postage on an order you have sold, which happened because the unit left, not because it stayed.

Beyond that there is no agreed component list. Two careful people can total the same store's carrying cost differently and both be defensible, which is the first reason a borrowed percentage cannot be compared with yours.

Capital cost

Capital cost is the price of the money sitting in your stock. If you financed the inventory, this one is not an estimate: it is the interest you were charged on the borrowing that funded stock, and it is on a statement.

If you funded it yourself, capital cost is an opportunity cost, and stores either skip it or invent it. Skipping makes stock look free to hold. Inventing usually means picking an aspirational rate of return, which produces a large number with nothing behind it.

The version that survives scrutiny is concrete: the rate you would pay to borrow the same amount back if you needed it tomorrow, or the return on the next thing that money would have funded. Capital is also the component that scales with unit cost rather than unit count, which matters when you allocate the total across products.

Storage, service and risk

Storage is the space the stock occupies: warehouse or storage-unit rent, the storage line on a 3PL invoice, shelving. One honest note, because this is where the calculation usually gets padded. If the stock sits in a spare room you already pay for, with no alternative use and no bill attached, its storage cost is close to zero and you should write zero. An imputed rent on a garage you were never going to let out is a number invented to make a case.

Service is insurance on the stock, any tax assessed on inventory where you operate, and the labour of holding it: receiving, counting, moving boxes, reconciling variances. The labour half is what stores underestimate most, because it never appears as an inventory line item. It appears as somebody's Tuesday. Count the hours and multiply by what they cost.

Risk is the value stock loses while it waits: shrinkage, which shows up as the variance at your counts; damage, which shows up as write-offs; and obsolescence, which shows up as units you eventually sold below what you paid. Obsolescence is invisible until it is realised, so measure it backwards, using what you actually cleared under cost last year rather than a forecast of how much of today's stock will go stale.

Building your own rate

The calculation is addition and one division. The work is collecting the lines, and it is an afternoon with a year of invoices in front of you.

Total those and you have your annual carrying cost in dollars, which is already useful. To turn it into a rate:

Carrying cost rate = total annual carrying cost ÷ average inventory value at cost

Two rules keep the result comparable with itself next year. Value the denominator at cost, never at retail, or the rate quietly shrinks by whatever your margin is. And build average inventory from monthly snapshots rather than a start and end figure, which is the standard shortcut but a poor one for any store with a seasonal buy in it.

For allocating cost to individual products, a second form is more practical:

Carrying cost per unit per year = total annual carrying cost ÷ average units on hand

A worked example

The figures below belong to an invented store. They are here so the arithmetic is visible end to end, not because they are typical of anything.

Checking the two composite lines: 3 hours a week of receiving and counting is 156 hours a year, and 156 × $22 = $3,432, which with the $480 policy makes $3,912 of service cost. The discontinued line cost $9 and cleared at $3, so 140 × $6 = $840, which with $520 of count variances makes $1,360 of risk. Add the buckets: $2,400 + $3,120 + $3,912 + $1,360 = $10,792 for the year.

That store held an average of 2,900 units across its catalogue over the same twelve months, so:

$10,792 ÷ 2,900 = $3.72 a unit a year, or $0.31 a month. Treat that flat figure as a first pass. It charges the same amount to a small cheap item and a large expensive one, when capital scales with what a unit cost and storage with the space it takes; the careful version allocates capital by each product's share of inventory value and storage by the space it occupies. For the same reason, do not divide the flat figure by one product's unit cost. That mixes a cost-driven component with a volume-driven one and returns a number that means nothing.

Now put a product against it. Cedar & Fig, 250g sells 5 units a day at $18 on a $7 unit cost, and the store holds an average of 165 units of it. At $3.72 a unit a year, that is 165 × $3.72 = $614 of carrying cost, against 1,825 units a year at $11 of margin, which is $20,075 of gross margin. Carrying cost is a rounding error on a product moving that fast.

Change one input and it stops being one. A gift tin at the same $7 cost and $18 price, held at the same average of 165 units, but selling one unit every five days, moves 73 units a year for $803 of gross margin. Its carrying cost is the same $614, because it occupies the same shelf and the same money, and that leaves $189. Nothing changed except speed. Holding 165 units of something selling at that rate is more than two years of cover, and $614 a year is what that decision costs in cash.

Using the rate

Carrying cost is not a report line. It is an input to three decisions otherwise made on instinct.

Order sizing. Every order quantity trades ordering cost against holding cost, and holding cost is this number. It is the H term in the economic order quantity formula, which cannot be calculated without it, and it is what tells you whether accepting a supplier's minimum order quantity is a bargain or an expensive way to buy a discount.

Hold or discount. When stock is not moving, the question is whether waiting costs more than discounting. Carrying cost is the price of waiting, and it accrues whether the units move or not. Working out how much excess you are holding gives you the unit count; multiplying by your monthly rate turns it into a figure you can set against a markdown.

Buying discipline. Most carrying cost is created at the moment of the order, which is why preventing overstocking does more for the total than any storage renegotiation will.

One thing the rate does not feed: GMROI ignores carrying cost entirely, so a slow mover with a weak GMROI is worse than its GMROI says.

Why borrowed percentages mislead

You will find a single tidy percentage of inventory value quoted almost everywhere carrying cost is discussed, and its provenance is worth knowing before you plan around it. The figure traces back to a trade-magazine article published in 1995 by Helen Richardson in Transportation & Distribution, not to any study of ecommerce, and the range that article gave was wide enough to be useless as a target. What circulates now is a midpoint that lost its range somewhere in thirty years of retelling. The supply-chain trade body it is usually attributed to does not publish it anywhere we could find, across two separate attempts to trace it.

The deeper problem is not provenance. The number is not portable even when it is honestly measured. Two stores holding identical stock can have carrying costs that differ by a multiple, and every driver is structural: one financed the stock and one did not, one rents a unit and one uses a room it already had, one sells a category that goes stale in a season, one counts the receiving hours and one has never counted them. A percentage is the output of those choices, so importing it while keeping your own inputs gives you a number that describes nobody.

The comparison that works is against yourself. Total your components this year, total them the same way next year, and the movement is real information: whether the storage bill grew faster than the catalogue, whether obsolescence is creeping, whether handling hours rose when you added a location.

The reason the total keeps climbing is usually upstream of storage. It is buying too much too early, one order at a time. StockCue forecasts demand per SKU from your own sales history, on every plan including Free, which decides how many units are sitting there in the first place. It will not total your invoices; nothing does that except you and an afternoon.

Frequently Asked Questions

What is included in inventory carrying cost?

Four groups: the capital tied up in the stock, the space it occupies, the service costs of holding it (insurance, any tax on stock, and the labour of receiving, counting and moving it), and the risk of losing value while it waits (shrinkage, damage and obsolescence). A standard operations-research definition names capital, space, insurance, protection and taxes; most merchants add handling labour and obsolescence, because both are real and both are theirs. There is no universally agreed component list, so name the components you counted whenever you quote your figure.

What is a typical inventory carrying cost percentage?

There is no defensible answer for a small Shopify store. The percentage quoted most often traces back to a trade-magazine article from 1995 rather than to any study of ecommerce, and the original gave a range far wider than the tidy figure people repeat from it. Totalling your own components and dividing by your own average inventory value takes an afternoon with your invoices, and it is the only version of the number that can safely inform a decision.

How do I calculate carrying cost for one product?

Total your store's annual carrying cost first, then allocate it. The quick method divides that total by the average number of units you hold, giving a cost per unit per year, then multiplies it by the product's own average units on hand. The careful method allocates the capital component by unit cost and the storage component by the space a unit takes, which matters as soon as one product is much bulkier or much more expensive than the rest of the catalogue.

Does carrying cost include the cost of the inventory itself?

No. The purchase price of the stock is the investment being carried, not a cost of carrying it. Carrying cost is what holding that investment costs you over a period on top of the purchase: interest or opportunity cost on the money, storage, insurance, handling, and value lost to shrinkage and obsolescence. Counting the stock itself double-counts it and produces a rate above 100 percent, which is usually the first sign a calculation has gone wrong.