The retailer that sells almost nothing at a profit

Two companies invented this format separately and then merged. The Price Company began operations in San Diego in 1976, pioneering the membership warehouse club; Costco Wholesale Corporation began operations in Seattle in 1983 with what its own filings call a similar concept. In October 1993 they combined into a single company with, at the time, over 15 billion dollars of annual sales and more than 200 warehouse clubs. It traded as Price/Costco until shareholders approved the current name in January 1997.

Costco's own statement of the concept is the place to start, because it is unusually blunt about the direction of the causation. The filing says the company operates on the concept that offering low prices on a limited selection of nationally-branded and private-label products in a wide range of categories "will produce high sales volumes and rapid inventory turnover." Low prices are not a reward for efficiency achieved elsewhere. They are the input, and volume and turnover are what they buy.

The limited selection is the mechanism. Costco says it carries fewer than 4,000 active stock keeping units per warehouse, which it notes is significantly less than other broadline retailers, inside a building averaging about 147,000 square feet. Fewer items means larger orders per item, which means better prices, which brings more volume. The filing states the consequence plainly: these volumes and turnover let it "operate profitably at significantly lower gross margins... than most other retailers."

Then comes the sentence most readers skip, and it is the whole of working capital in fourteen words: "We often sell inventory before we are required to pay for it." A pallet arrives, sells, and the cash from selling it is in hand before the supplier's invoice is due. The customer funds the inventory.

So a retailer has deliberately given away gross margin and arranged not to finance its own stock. It has to earn its profit somewhere, and it charges people for the right to walk in. Costco is direct about why that matters, in the management discussion rather than the business section: "The membership format is integral to our business and profitability. This format is designed to reinforce member loyalty and provide continuing fee revenue."

So the question to hold through the rest of this: if you strip the membership line out of the income statement, is what remains a good business, a break-even one, or a machine that exists to sell memberships?

Two revenue lines, and they are not remotely alike. Net sales of merchandise and fuel were 269,912 million dollars in fiscal 2025. Membership fees were 5,323 million, about two percent of the total, which is small enough that a casual reader skips it.

Now look at what each line costs to produce. Merchandise cost 239,886 million dollars to buy, leaving gross margin of 30,026 million, which Costco reports as 11.12 percent of net sales. Selling, general and administrative expenses took 24,966 million, or 9.25 percent. So the merchandise operation earns roughly two points of margin before anything else happens.

The membership line has almost no cost of goods at all. A renewal is a card swipe against a warehouse network that already exists, staffed by people already on shift. Whatever the fee is, most of it survives to operating income intact. That asymmetry is the whole business model, and it is why the small line is the load-bearing one.

The fee also arrives early. A member pays for a year in advance and then shops through it, so Costco holds cash it has not yet earned and pays suppliers on terms after the goods have already sold. This is the same structure an insurer has, where premiums come in before claims go out, and it is the reason a retailer with two points of merchandise margin can fund its own expansion.

One more thing the fee does, and it is easy to miss: it changes who walks in. Someone who has paid 65 dollars for the year has a reason to consolidate their shopping rather than split it across three stores, because the fee is only recovered through use. The membership is not just a revenue line. It is the mechanism that produces the volume the low prices were supposed to buy.

Start with the member, because the member is the unit. Costco ended fiscal 2025 with 81.0 million paid members holding 145.2 million cards, and the annual fee in the United States is 65 dollars for a Gold Star or Business membership. A member can pay a further 65 dollars to upgrade to Executive, which pays back two percent of qualifying purchases up to a stated maximum reward of 1,250 dollars a year.

That upgrade is where the interesting arithmetic sits. Executive members were 38.7 million of the 81.0 million paid, about 47.8 percent of the base, and they accounted for roughly 73.6 percent of worldwide net sales. So half the members do three quarters of the shopping, and they are the half that pays twice. The reward is not a discount Costco absorbs unwillingly. It is a payment for the behaviour it most wants.

Now the warehouse. Costco ended the year with 914 warehouses and 269,912 million dollars of net sales, which is about 295 million dollars of net sales per warehouse. That figure needs a caveat: Costco publishes its own average-sales-per-warehouse metric at 272 million dollars, and the filing does not reconcile the two. E-commerce is about seven percent of net sales and is not attributable to a building, which plausibly explains most of the gap, though the filing does not say so. Cite whichever you use, and do not present the computed figure as Costco's.

Either way the number is enormous. A large supermarket does a fraction of it, and the reason is the same limited selection: 147,000 square feet holding a few thousand items sells each item in volumes a supermarket never sees. Comparable sales grew six percent in fiscal 2025, eight percent excluding fuel and currency, and e-commerce grew sixteen percent.

The renewal rate is the last unit figure, and the one to watch hardest: 92.3 percent in the United States and Canada, 89.8 percent worldwide. Read Costco's own definition before you use it. The rate is a trailing calculation covering renewals in the period seven to eighteen months before the reporting date, so it tells you about last year's members, not this month's.

The temptation with a membership business is to size the market by counting households. Resist it, because the constraint here is not demand for cheap paper towels. It is buildings.

Costco opened 27 warehouses in fiscal 2025, three of them relocations, for 24 net new against a base of 914. That is under three percent unit growth, and it is not a decision about ambition. A warehouse needs a large parcel, highway access, and a catchment dense enough to fill 147,000 square feet with people making a deliberate trip. Those parcels are scarce in exactly the places with the most members.

So the honest sizing runs the other way. Take the units it can plausibly add, multiply by sales per warehouse and by the fee income those members bring, and you get a growth rate in the high single digits rather than a market-share story. In fiscal 2025 the two engines were both modest and both real: six percent comparable sales, and membership fee revenue up ten percent, of which the September 2024 fee increase accounted for about forty percent of the growth.

That last figure is the most useful sentence in the filing for a sizing exercise. It says the fee is a lever management can pull, that pulling it once moved membership income by a measurable amount, and therefore that a chunk of future profit growth does not depend on opening anything at all. It also tells you the lever is finite. There are only so many times you can raise a price whose whole purpose is to make people feel they are getting a deal.

The geographic answer is where the room actually is. Costco's United States and Canada operations were 86 percent of net sales and 84 percent of operating income in fiscal 2025, on 739 of the 914 warehouses. Other International runs 175 warehouses producing 38,266 million dollars of total revenue. If the format travels, the arithmetic of the next thousand warehouses is mostly written outside North America, and the renewal rate there is the thing to watch, because Costco says its worldwide rate is dragged down by growth in newer international markets.

Three reported segments, and one disclosure limit worth knowing before you try to compare them. Costco reports segment total revenue, which bundles membership fees in with merchandise and fuel, and it does not disaggregate membership fees by segment. So you cannot compute a segment's merchandise margin from this filing, and any per-segment gross margin you see quoted somewhere else was built on an assumption.

What the segments do tell you is where the profit sits and how differently the three behave. The United States produced 200,046 million dollars of total revenue and 6,878 million of operating income. Canada produced 36,923 million and 1,849 million. Other International produced 38,266 million and 1,656 million.

Read those pairs as ratios and Canada is the standout: on revenue only slightly below Other International, it earns materially more operating income. A mature market with 110 warehouses, high renewal rates and no expansion drag converts revenue better than a young one still opening stores. That is not a permanent verdict on either. It is what the two look like at different points in the same build-out.

The concentration underneath is worth naming. California alone was 26 percent of United States net sales. A membership warehouse business is a bet on specific metropolitan areas, and the largest of those bets is one state's housing costs, wage growth, and regulation.

One number, and it is not comparable sales. If membership fees are the profit and the fee is only collected from people who renew, then the renewal rate is the business, and everything else is the reason it stays high. Costco says as much in its own risk factors: the extent to which it grows the base, increases Executive penetration and sustains high renewal rates "materially influences our profitability."

So run the failure honestly. At 92.3 percent renewal in North America, the base replaces itself with room to spare. Take that to the mid eighties and two things happen at once: fee income falls, and the members who leave are disproportionately the low-frequency ones whose fee was almost pure margin, so profit falls faster than revenue. The merchandise operation, earning about two points after SG&A, cannot absorb that.

The second failure is subtler and it is priced into the model. The fee only feels fair while the prices visibly beat the alternative, so every fee increase spends some of the goodwill the low prices bought. The September 2024 increase delivered about forty percent of the year's membership income growth, which is the argument for pulling the lever. It is also a reminder that the lever works by asking members to pay more for the privilege of a bargain, and there is a level at which that sentence stops being persuasive.

Third, watch what the renewal rate is made of rather than the rate itself. Costco tells you the worldwide figure is dragged down by newer international markets and by memberships sold online through digital promotions, which "renew at a slightly lower rate on average." A promotion that buys sign-ups at a lower renewal rate improves this year's member count and dilutes next year's fee income. Growth bought that way looks identical to growth on the way in and different on the way out.

Finally the boring one, which is that the fee cannot be raised into a recession while the merchandise margin is two points and fuel is about ten percent of net sales. The model is genuinely defensive, but its defence is that members feel they are saving money. That is a feeling, and feelings are cyclical.