How Costco turned higher wages into lower costs

In April 2018, at a chamber of commerce lunch in Issaquah, Washington, Costco's then chief executive Craig Jelinek told a story against himself. He had once gone to the company's co-founder, Jim Sinegal, with a problem. "Jim, we can't sell this hot dog for a buck fifty. We are losing our rear ends." Sinegal's reply, as Jelinek recalled it, was short. "If you raise the effing hot dog, I will kill you. Figure it out."

So Jelinek figured it out. Costco began making the hot dogs itself at its meat-processing plant in Tracy, California, and later added a second facility in Morris, Illinois. The combo has cost $1.50 since 1985. Costco sold 245.1 million of them in the 2025 financial year, and in March this year the current chief executive, Ron Vachris, filmed himself eating one and told the camera the price would not change while he was around.

The story tends to get filed under branding, and there is no shortage of case studies to file it alongside. Costco is widely regarded as the best-run membership business in retail, and the numbers validate the claim. Total revenue passed $275 billion in the financial year to August 2025. The renewal rate in the US and Canada stood at 92.2 per cent at the end of the third quarter of this year, which is a figure most subscription businesses would take on any terms. Cheap prices, an annual fee, Kirkland Signature, a food court priced to reinforce the value promise rather than maximise its own margin. Most people know the model.

But "figure it out" is not a branding instruction. It is a constraint handed down with no route out of it, and a constraint like that is only survivable if something underneath the business is strong enough to absorb it. The more useful question about Costco is not what it charges. It is what it refuses to touch, and what makes that refusal affordable when every competitor in the sector faces the same cost pressure and gives way.

The membership fee is the financial stabiliser. In the 2025 financial year Costco collected $5.32 billion in fees against operating income of $10.38 billion. Slightly more than half of its operating profit therefore came from customers paying for the right to return, collected largely in advance and recognised across the year.

That does not mean the merchandise makes no money. It means Costco does not require every item to carry the margin expected from a conventional retailer. The fee stream gives it room to hold prices down, while limited product choice, enormous buying volumes and a deliberately simple warehouse operation make the labour inside each building unusually productive.

The wages sit inside that system. Costco can pay top-of-scale hourly workers $32.90 because it has designed the work to carry the rate, and the rate helps preserve the experience that keeps the system productive. Membership creates the room. Operational discipline earns it. Retention compounds it. Values are the language used to describe the choice afterwards.

It would be wrong to make that sound serene. In January 2025 the Teamsters, representing around 18,000 Costco workers across six states, voted by an 85 per cent margin to authorise a strike, alleging the company had harassed organisers and shut them out of stores. One worker of 37 years told reporters this was not the Costco he had joined. Costco announced a raise for its non-union staff the day before the strike deadline, a deal was reached hours before it expired, and some members then called the outcome a sellout. The compact is being tested, and holding it is not free. That is worth saying plainly, because a story this tidy invites suspicion and it should.

Now the part that carries commercially. The pay decision was not taken recently. Sinegal learned it from Sol Price at FedMart and built it into Costco from the early 1980s. The results being admired this year sit roughly forty years downstream of it.

Tony Barzar is what that gap looks like from the inside. He joined Price Club in 1986 collecting trolleys and later moved to the checkout, where he has stayed. The Wall Street Journal reported in July that he earns $32.90 an hour and has more than $1 million in his retirement account. He has turned down supervisory roles because he prefers working the till. Costco's chief financial officer, Gary Millerchip, has said that many thousands of the company's American hourly workers hold 401(k) balances above $1 million.

That tenure is doing commercial work. Turnover at Costco after the first year runs at around 7 per cent, against a retail industry average close to 60 per cent. Inventory shrink runs at 0.1 to 0.2 per cent of sales, against a national retail average of 1.4 to 1.6 per cent. Some of that second gap is structural, because a member-only warehouse with one entrance and a receipt check is hard to steal from. But structure does not notice a mispriced pallet or a double-scanned item, and someone who has worked the same floor for twenty years does. The canopy everyone is admiring is standing on roots that were watered in 1990. Costco has also resisted the assumption that tenure only creates value when it leads to promotion; Barzar’s knowledge remains economically useful precisely where the customer encounters it.

A wage is a rate you can see on a payroll. Turnover is a cost you have to go looking for.

The obvious objection is that all of this is affordable because the business model is superb, and that the wages are a consequence rather than a cause. The nearest competitor is the test of that. Sam's Club has the same warehouse format, the same membership fee and a parent company with more scale than Costco has. It also ran turnover at around 44 per cent against Costco's 17 per cent on Wayne Cascio's 2005 figures. Then John Furner, running Sam's Club, pushed wages up against resistance from his own HR and finance functions, and within two years productivity was up 16 per cent, turnover down 25 per cent and sales up 25 per cent, according to Zeynep Ton at MIT Sloan. Furner went on to run Walmart in the United States. A competitor holding every structural advantage ran the experiment on itself and arrived at the same answer.

Cascio's study remains the cleanest translation of that into money. In 2005 Costco's average hourly rate was $17 against roughly $10.11 at Sam's Club, more than forty per cent higher. Costco did $43.05 billion of American sales with 38 per cent fewer employees. Operating profit per hourly worker came to $21,805 at Costco and $11,615 at Sam's Club. The higher rate bought the lower cost.

The same arithmetic runs today in a different currency. Renewal is the profit pool. Close to half of operating income now depends on members choosing to pay again, renewal depends on the experience inside the building, and that depends on who is standing in it and how long they have been there. For the investors and acquirers reading this, that is a diligence question rather than a values question. A target's headline margin can look identical whether its service rests on twenty-year staff or on a workforce that turns over twice a year, and the second version carries a repair bill that has not been written down anywhere.

Costco is not exempt from that scrutiny either. Its shares trade at a heavy premium, on a trailing multiple near 56 at the time of writing, and the stock has been flat to down over the past twelve months while the wider market rose. Very little room for error is priced in. The model is not invincible. It is legible, and the thing carrying it is visible if you look at the right line.

None of this requires a warehouse or a billion dollars. It requires two moves.

The first is to find out what turnover actually costs you. Not the recruiter's invoice. The weeks the role sat empty, the manager hours spent covering it and then hiring for it, the training, the errors in the first six months, the customer who noticed and said nothing. McKinsey put the cost of replacing a single frontline retail worker at around $10,000. Whatever your equivalent number turns out to be, it is very likely larger than the raise you decided you could not afford, and it is currently sitting across four cost codes where nobody has to own it.

The second move is to protect the decision structurally rather than personally. Decide which condition your future performance depends upon and remove it from the category of costs that can be cut reflexively. That may mean ring-fencing a portion of recurring revenue, setting a minimum staffing ratio, fixing a pay position against the market, or requiring the full replacement cost to appear beside every proposed labour saving. Write the rule down. Give it authority beyond a bad quarter and beyond your own mood during one.

Which leaves one question worth carrying out of this.

What is the single condition your current results actually depend on, and are you funding it as an asset or trimming it as a cost to make this quarter's numbers? If you own the business, ask it of what you have built. If you are about to buy one, ask it of what you are about to inherit, and then ask how long the answer has been true.