Price divided by how often you will use it
Cost per use is the sticker price divided by an honest estimate of how many times the thing will get used. That is the whole method, and it is the only comparison that turns two different prices into the same unit.
A $240 coat worn four times a week for four winters costs about thirty cents a wearing. A $60 coat worn twice costs thirty dollars a wearing. The cheap one was a hundred times more expensive in the only unit that matters, and it is the one that felt responsible at checkout.
This sits inside the wider question of whether something is worth buying at all, and it is the half of that question you can actually calculate.
Running the numbers
Three inputs, and only one of them is hard.
The price is given, though it should be the price you actually pay rather than the one on the badge. Include delivery where it is real and exclude anything that comes back to you later, because money returned in ninety days is not a reduction in what leaves your account today.
The number of uses is the input everything turns on. More on that below.
The lifespan is the ceiling on the second number. A thing used daily cannot be used a thousand times if it fails at four hundred, so the two interact and the shorter one wins.
Then divide, and where the purchase spans years, divide again by the years. A product costing half again as much and lasting twice as long is cheaper per year, and per-year is the comparison that survives when the two candidates have very different expected lives. That second input is its own question, covered in how long a product should last.
The honest estimate is the hard part
Everyone overestimates how often they will use a thing they already want. This is not a character flaw; it is what wanting something does to a forecast, and knowing about it does not switch it off.
Two disciplines work.
Write the estimate down before you look at the price. Doing it in the other order lets the price anchor the estimate, which is exactly what it does otherwise. A $400 machine quietly produces a higher use estimate than a $90 one, for no reason connected to your week.
Check it against the last comparable thing you bought. Not against your intentions, against what actually happened. The bread maker, the second monitor, the standing desk, the exercise equipment. Most people have a reliable personal multiplier, and it is usually well below one.
A third check for anything expensive: describe the specific occasions on which you will use it, by name. If you can list four, the estimate is probably four rather than the twenty you first thought.
Where it changes the answer
Four situations where the arithmetic reverses the obvious conclusion.
Expensive and used constantly. Frequently the cheapest thing on the page in cost-per-use terms, and the one that feels extravagant. Kitchen equipment, a decent chair, shoes you wear daily.
Cheap and used twice. The classic regret, and it does not register as one because the number was small. Anything seasonal, anything bought for a single occasion, anything bought to support an intention rather than a habit.
Two products at different prices. Cost per use converts both into the same unit, which is what makes them comparable at all. It is the calculation that settles most of the ties in comparing two products.
Renting or borrowing. Once the number is per use, the alternative becomes visible. A tool needed twice is a hire rather than a purchase, and that comparison is impossible while the price is a single figure.
Where it gives the wrong answer
Worth naming, because a method applied everywhere stops being a method.
Capability purchases. A smoke alarm you never need has an infinite cost per use and is the right purchase anyway. The same goes for safety equipment, spares, backups and anything whose value is that it exists when something goes wrong. For those, the question is what the failure would cost, not what each use costs.
Things with a use that is not countable. A rug, a paint colour, a piece of furniture you look at rather than operate. You can force a number and it will not mean anything.
Pleasure. A thing bought because it is a pleasure to own is a legitimate purchase and the arithmetic will always argue against it. Naming that honestly is better than dressing it up in a calculation that was going to lose.
The method's real job is not to veto. It is to stop a low price reading as low cost when it is not.
Subscriptions, where the gap is widest
Everything above applies to recurring spend and is more uncomfortable there, because the number is available and nobody looks at it.
Divide the monthly fee by the number of times you actually opened the thing last month, not by the times you meant to. Most services publish enough usage information to make this a fact rather than an estimate, which is a rare advantage.
Two patterns recur. The service used twice a month at a price that would be fine at twenty, which is a candidate for cancelling rather than for feeling bad about. And the bundle where one component is used constantly and the rest are not, which is worth pricing as the component alone before renewing the whole.
The annual plan deserves its own line. A discount for paying yearly is real only if you would have stayed the full year, and the discount exists partly because many people would not have. Work the cost per use on the monthly plan first, then decide whether the annual price is a saving or a commitment.
The number the price tag hides
Two purchases at the same price can differ by an order of magnitude once the division is done, and the listing cannot tell you which is which.
Take two chairs at $300. One is used eight hours a day for six years and works out at about two cents an hour. The other is a spare used for guests twice a month and works out at about two dollars a sitting. Same price, same category, and a hundredfold difference in what the money bought.
That gap is why a single price is a poor basis for comparison and why the tier you should buy at depends on the use rather than on the budget. It is also the mechanism behind budget against premium: the frequency of use is what decides where on that curve a purchase belongs, which is why the same advice gives opposite answers for a daily chair and an occasional one.
The corollary is worth stating too. A thing used rarely should usually be bought cheaply, and the guilt people feel about buying the basic version of something occasional is misplaced. The expensive version of a rarely used thing is the genuine waste.
The one-line version
Before buying anything above the amount you would spend without noticing: write down how many times you will use it, divide the price by that number, and compare it with the alternative in the same unit. It takes about fifteen seconds and it is the single most useful habit in this whole cluster.
One caution about the habit. The point is to compare, not to justify. A calculation run after you have decided will always produce a number you can live with, which is why the estimate goes on paper before the price does.