Date labels: Best by vs use by vs sell by — what’s the difference?

Something changed on the package this year, but it wasn’t the food.

In California, any packaged food made after July 1 that carries a date now has to use one of two phrases. BEST if Used by means the date is about quality. USE by means it is about safety. The words “sell by” are gone from the consumer side of the label. New York’s legislature wrote a similar bill in June, with a 2028 compliance date. A federal version has been sitting in Congress since 2025.

The words changed. The food did not. A date is still chosen by the company selling it, still means whatever category that company decided it means, and for most products is still not required by anyone.

Which leaves the question the new labels were supposed to settle: what do the three dates actually mean, and which one is the one that matters?

The three dates, and what each one is for

Phrase	What it means	Who sets it	What it is not
BEST if Used by	Peak quality: flavor, texture, potency	The manufacturer	A safety date
USE by	The company’s judgment that the food should not be eaten after this date	The manufacturer	A quality suggestion
Sell by	Stock rotation for the store	Retailer or manufacturer	not addressed to you

There is no federal requirement to put a date on most food. Infant formula is the exception, and it is a real one: the date is required, and it should be obeyed. For everything else, dating is voluntary, and more than fifty different phrases have been in use in the United States — “best before,” “enjoy by,” “expires on,” “fresh until,” and the rest.

FSIS, the USDA agency that handles meat and poultry, recommends that companies that do date their products use “Best if Used By,” because research shows shoppers read that phrase as a statement about quality rather than safety. The agency’s own guidance is blunt about what happens after the date passes: food is still safe if it has been handled properly and shows no signs of spoilage.

“Sell by” is the phrase that caused most of the trouble. It was always an instruction to the store about how long to display a product, and it was never a statement about whether the food was good for you. Shoppers read it as one anyway. California’s law keeps the information for retailers in coded form and takes the words off the package, which is the cheapest possible fix: the label that confused people is simply gone.

Who sets the date, and what they’re allowed to make it mean

The date is set by the company selling the food. No federal rule requires a documented method for choosing it — no required shelf-life study, no testing protocol, no filing with a regulator. A manufacturer decides how long its product stays at peak quality, prints that decision on the package, and is not obliged to show how it arrived at it. New York’s bill would add that obligation, requiring companies selling there to substantiate the date with a scientifically valid method.

The date is the manufacturer’s estimate of its own product, printed where you will read it as a fact.

This decision has two audiences. A longer date makes the food easier to sell and easier to keep. A shorter date makes it something you have to come back for. The direction a company chooses depends on which of those pressures is stronger, and nobody outside the company can see the calculation.

The cleanest evidence that date length changes buying behavior comes from a policy change rather than a corporate one. In 2010, New York City repealed a rule that had required milk to be sold within nine days of pasteurization, leaving the date to the companies. Shelf life stretched to about fourteen or fifteen days. Researchers who studied the change found that purchase volume fell by roughly 10 percent, reported waste fell by more than 10 percent, and people actually drank more milk — they bought less of it and let less of it spoil (Yu & Jaenicke, Food Policy, 2021).

The authors frame this as consumer welfare, and it is. But the arithmetic describes a business: a longer date meant households had to buy milk less often. Whether any particular company thinks about that when it sets a date is not something the study can see.

Yann Cornil of UBC’s Sauder School notes that less scrupulous manufacturers might take advantage of the fact that shoppers have to repurchase more often. He is equally clear that the shortcut shoppers use is partly right: a longer shelf life often does track heavier processing, which is what makes the cue so hard to drop.

The published research does not include a case of a company shortening a date to drive sales; what it shows is the incentive and the effect. The defensible version is narrower and harder to argue with: nobody verifies the number, it is set by the party that profits from the sale, and a shorter number sells more food.

There is a second cost, and it lands on food banks. Twenty states restrict the sale or donation of food past its labeled date, so surplus that is still wholesome gets thrown out rather than given away. That is a policy failure with a paper trail, and it is the strongest argument in this whole subject — stronger than any claim about what companies intended.

When the date is real

Everything above can create the impression that dates are noise. They are not, and the exceptions are not subtle.

USE by on perishable food is a safety instruction. Deli meat, soft cheese, smoked fish, cut fruit, sprouts, prepared salads, unpasteurized juice. It is the manufacturer’s judgment rather than a regulator’s ruling — nobody approves the date — but it is the one place where the guess has consequences past texture. Listeria grows at refrigerator temperatures and does not announce itself with a smell or a color. If a package of sliced turkey says USE by Tuesday, Wednesday is not a dare.

Infant formula is the one product where the federal government requires a date, and the reason is nutritional: the date is when the formula is guaranteed to still contain each nutrient in the amount on the label.

Anything you cannot smell or see is a date decision, not a senses decision. Vacuum-packed fish, sealed trays. Your nose should not get a vote here.

Damaged packaging overrides every date on it. A swollen can, a dented lid, a torn seal, a bulging pouch: discard, don’t investigate.

A power outage resets the question. Time and temperature decide, not the print. A full freezer holds temperature longer than a half-empty one; a fridge that sat open through an afternoon is a judgment about hours, not weeks.

The date is a guess about quality that, on perishable food, doubles as a safety rule you cannot verify.

When it’s window dressing

The pantry is where the date comes closest to decoration.

Canned beans, canned tomatoes, canned fish. Dried pasta, rice, oats. Honey, salt, sugar, vinegar. Frozen vegetables and fruit. Hard cheese. These are the cheapest calories in the store and often the longest-dated, and the date on them is a statement about texture, color, and flavor rather than about whether the food is safe.

A bag of dried beans with a date two years out is telling you about the bag. A can of tomatoes stamped 2028 is telling you when the liner and the acid will begin their slow argument. Neither is telling you anything about how processed the contents are, which is a separate question and the one most shoppers think they are answering.

USDA’s position is more permissive than most people assume: food past its date is still safe if it was handled properly and shows no spoilage. For shelf-stable goods, quality fades long before safety does.

The practical test for a quality date is the one British waste campaigners have pushed for years: “look, smell, taste”. Open it, examine it, sniff it, and if it still looks and smells like food, cook a small amount and taste it. If it tastes flat or off, throw it out. That protocol is for quality dates. It is not a way to overrule a use-by date, and it is not a way to rescue a compromised can.

Some foods Americans treat as strictly perishable get a looser treatment elsewhere. British guidance covers yogurt and hard cheese with best-before dates and sensory checks, on the theory that a pot of yogurt a few days past its date is usually a decision your nose can make. Your mileage will depend on your kitchen, your immune system, and how much you trust your own nose.

What the confusion costs

The numbers are large enough that they have become a policy argument in their own right.

Date-label confusion is associated with about 4.3 million tons of wasted food a year in the United States, at a cost of more than $21 billion (ReFED) — roughly $160 per household a year, if you spread it evenly. USDA estimates that 30 percent of the food supply is lost or wasted at the retail and consumer level. In the UK, 1.8 million tons of household food is thrown away each year because it wasn’t used in time.

Date-label confusion is associated with about 4.3 million tons of wasted food a year in the United States, at a cost of more than $21 billion

Deference to the label has grown even as the labels stayed confusing. In a 2025 national survey for ReFED, Harvard Law School’s Food Law and Policy Clinic, and Johns Hopkins, 43 percent of consumers said they always or usually discard food near or past its label date, up from 37 percent in 2016.

Nearly everyone uses the labels: 96 percent of respondents in a January 2026 ReFED/YouGov survey said they rely on date labels to some extent, and 31 percent said they follow them strictly for every food. A third said they only apply dates to foods they consider high-risk, like meat, dairy, and eggs. Another third use them as a general guideline.

The 2026 survey points the other way on behavior. Fifty-five percent of households said they were throwing away less food than a year earlier, and among people whose grocery bills had risen, 77 percent said they were checking and cooking what needed using up. The two findings coexist: more people defer to the date, and more people are trying to waste less.

Why a longer shelf life looks less natural

Shoppers read a longer shelf life as less natural and less healthy, even when the only stated difference is packaging. In one experiment, two granola bars with nearly identical ingredients and nutrition panels still split the same way. That finding comes from seven experiments, plus four more in an appendix, published in the Journal of Consumer Psychology in 2026 (Liu, Cornil & Xu).

When researchers framed a choice between a one-year and a three-year pasta sauce around eating natural food, nearly 70 percent of participants picked the one-year sauce. When they framed the same choice around protecting the environment, that number fell to 43 percent. Two sauces, one changed emphasis, 27 points of preference.

Telling people the longer date came from a better seal did not move them. Asking them to explain out loud why the dates differed did. That is useful: the cue loosens when you make yourself say out loud what the date is actually measuring.

The study is a set of controlled experiments about perception and stated choice. It shows how it works. It does not show that changing your reading of the date changes your behavior in the real world.

Five rules for the aisle

Classify before you judge. A quality date is information. A safety date is a warning. Most of the confusion in the aisle comes from treating the first as the second.

For quality dates on shelf-stable food, the senses decide. Look, smell, taste. The date tells you when the food was at its best, not when it became a hazard.

For safety dates on perishables, follow the print. It is a manufacturer’s judgment rather than a regulator’s, but your nose is not a Listeria detector, and “it smelled fine” is how people end up in bed.

If two packages list the same ingredients and differ only by date, buy the longer one. Then freeze what you won’t finish. The instinct to buy the shorter date because it seems fresher produces exactly the outcome the shorter date predicts: food you throw away. The same logic does not apply across different products — a long-dated snack bar is still a snack bar.

Questions people actually ask

Is it safe to eat food after its best-by date?
Usually yes, if it has been stored properly and shows no signs of spoilage. Best-by is a quality date. The exceptions are infant formula and anything with a use-by date on it.

What’s the difference between sell-by, best-by, and use-by?
Sell-by is a stock-rotation instruction for the store. Best-by is the manufacturer’s estimate of peak quality. Use-by is the manufacturer’s judgment that the food should not be eaten after that date.

Do food dates have to be accurate?
Nothing requires a date on most food, and no federal rule requires a company to show how it chose the one it printed. New York’s pending bill would change that for products sold there, requiring companies to substantiate shelf life with a valid method.

Can I freeze food on its use-by date?
Yes, if it has been kept cold and looks and smells normal. Freezing stops the clock on spoilage bacteria; it does not sterilize food that has already gone over.

Why did California ban sell-by dates?
Because shoppers read a stock-rotation code as a safety warning and threw away food that was still good. The state replaced it with two phrases that mean one thing each.

Citations:

Liu D, Cornil Y, Xu X. “Naturalness against environment-friendliness: The impact of shelf life on food perceptions and purchase decisions.” Journal of Consumer Psychology. 2026;36(4):431–450. doi:10.1002/jcpy.70034

Yu Y, Jaenicke EC. “The effect of sell-by dates on purchase volume and food waste.” Food Policy. 2021;98:101879. doi:10.1016/j.foodpol.2020.101879

USDA Food Safety and Inspection Service. “Food Product Dating.”

California Department of Food and Agriculture. “Food Date Labeling” (AB 660).

Venable LLP. “New York Food Date Labeling Bill Passes Legislature.” June 25, 2026.

ReFED. “How Confusing Food Date Labels Are Hitting Your Wallet.” March 2026; and the February 2025 Harris Poll survey with Harvard Law School and Johns Hopkins.

WRAP. “Food date labelling.”

Congress.gov. S.2541, Food Date Labeling Act of 2025.