The Pumpkin Spice Audit: We Checked 40 Products for Actual Pumpkin

Start with the part everyone gets right, because it matters for everything that follows.

Pumpkin spice, sold on most shelves as pumpkin pie spice, is a baking blend. It's built on cinnamon and ginger, rounded out with nutmeg, cloves and allspice, and sometimes mace. What it does is reproduce the warm, sweet, slightly peppery character of a pumpkin pie's seasoning, which is a different thing from reproducing pumpkin. The blend has deep roots in the spice trade, and a recognizable version of it shows up in Amelia Simmons's 1796 American cookbook, well over a century before anyone put it in a paper cup. The spice retailer Savory Spice traces the combination back through Dutch East India Company trading routes in the 1600s.

Here's the part that gets lost. The blend is named for what it's traditionally used with, not for what's in it, in exactly the way apple pie spice and gingerbread spice are. Nobody is scandalized to learn that apple pie spice has no apple in it. Pumpkin spice just happens to have become famous enough that its name started doing work its ingredients were never meant to do.

And that's where the popular framing quietly slips a gear. "There's no pumpkin in pumpkin spice" is a statement about a jar of seasoning. It gets repeated as though it were a statement about pumpkin spice muffins, pumpkin spice yogurt, and every other thing carrying the name. Those are different products, and they answer the question differently.

We wanted a count that anyone could rebuild, so the method is worth a minute of your time.

The source is FoodData Central, the USDA's public food composition database. Its Branded Foods collection holds label data for roughly two million packaged products, supplied through partnerships with GS1 US, 1WorldSync and Label Insight, and it includes the full ingredient statement as printed on the package. We downloaded the complete April 30, 2026 release and worked from that single file, so every figure below comes from one consistent snapshot.

From there, the steps were these:

• Find the population. We searched every branded record for a product name matching "pumpkin spice" or "pumpkin pie spice." That returned 3,174 records. Collapsing duplicates by brand and product name, and keeping the most recent label record for each, left 1,622 distinct products.
• Read every ingredient list. We scanned each statement for pumpkin, with three deliberate exclusions. A mention of "pumpkin spice" or "pumpkin pie spice" is the blend, not the vegetable. "Pumpkin seed" is a different part of the plant. None of those count as pumpkin.
• Classify the form. Puree or whole pumpkin, pumpkin powder or dried pumpkin, pumpkin concentrate or juice, a pumpkin flavoring only, or nothing at all. Only the first three count as containing pumpkin.
• Record the position. Where pumpkin appears, we noted its place in the ingredient list and how many ingredients the list contains.
• Pick the 40. We set aside plain spice jars, supplements and sugar, sorted the remaining 1,075 products into six aisles, and allocated the 40 slots in proportion to each aisle's share of that pool. Within each aisle we took the largest brand owners first and the most recent label record first, one product per brand owner per aisle, capped at two products per brand across the whole table.

Every one of the 17 positive results was then checked by hand against its raw ingredient statement, because a parser that silently miscounts is worse than no parser at all.

If that approach feels familiar, it's the same one behind our nut butter ranking: work from the actual label data, publish the arithmetic, and let anyone check it.

Of the 40 products we audited, 17 list some form of pumpkin in the ingredient statement, and 23 do not. That's 42.5%.

The number that gives it weight is the one behind it. Across the full population of 1,622 distinct pumpkin spice products, 706 contain pumpkin — 43.5%. Our sample of 40 lands within a single percentage point of the category it was drawn from, which is the most reassuring thing a small sample can do. The 40 aren't the flattering rows or the damning ones. They behave like the shelf.

Neither number is what the trivia promises. A claim that has hardened into common knowledge — that pumpkin spice products don't contain pumpkin — describes a little over half of them. The other half contain pumpkin, and in most of those it isn't a rounding error.

Label data has a way of complicating tidy narratives. It's part of why a study of ultra-processed toddler foods landed the way it did — the surprise wasn't in any single package, it was in what the count revealed.

The gap shows up even inside the coverage that repeats the claim. A 2024 Forbes article headlined "There's No Pumpkin In Pumpkin Spice" reports, a few paragraphs down, that the Starbucks Pumpkin Spice Latte has contained 2% real pumpkin puree from kabocha pumpkins — a figure the company gave a reporter, since US labels are not required to print ingredient percentages. The headline and the reporting are pointing in opposite directions, and the headline is the part that travels.

Here is the table. Brand, product, aisle, the FoodData Central identifier, the date of the label record, whether pumpkin is present, what form it takes, where it sits in the ingredient list, and how the product name handles the flavor wording.

"Sub" marks the four products where pumpkin appears inside a compound ingredient rather than in the main list — more on those shortly.

Scan that table for a second and a pattern starts to surface. It isn't about which brands are honest and which aren't. It's about what kind of food you're holding.

When we ran the pumpkin rate across every FoodData Central category with at least 25 products in it, the split was much sharper than we expected:

There is roughly a seventeen-fold difference between the top of that list and the bottom, and it maps onto something intuitive once you see it. Pumpkin puree is moist, heavy, mildly sweet and faintly orange. In a batter, a dough or a cultured spread, all of those are assets — the puree does structural work, holding moisture in a muffin and body in a cream cheese, on top of whatever it contributes to taste. In a clear syrup or a shelf-stable creamer, every one of those properties is a problem to be engineered around. Pumpkin settles, clouds, and brings a color and texture that a pourable product doesn't want.

So the practical shorthand is close to this: if the pumpkin spice thing in your hand was baked, blended into a batter, or cultured, the odds are strongly in favor of real pumpkin. If it's something you pour, stir in, or unwrap — creamer, drink mix, syrup, candy — the odds run hard the other way. The exception worth noticing is that the drinks category as a whole sits at 34.6%, pulled up by refrigerated items like eggnog and juice blends, which are closer to food than to flavoring.

It's also a reminder that the snack aisle rewards a closer look in general, something that comes up whenever nutritionists weigh in on which packaged snacks are overrated and which quietly deliver.

This is the finding we'd want a shopper to leave with, and it's the one the trivia can't deliver, because the trivia treats the whole shelf as a single thing.

Knowing that pumpkin is in there is one question. Knowing whether it's a real component or a wisp is another, and the label answers it more precisely than most people realize.

Under federal labeling rules, ingredients must be listed by common or usual name in descending order of predominance by weight. The list isn't a catalog. It's a ranking. Second place means there is more of that ingredient by weight than anything below it, and that makes position a genuine, free quantity signal on a package that is otherwise silent about amounts.

In our data, pumpkin generally shows up high. Across the 705 products where a position could be computed, the median position is fourth, and two thirds of the time pumpkin falls within the first five ingredients. In the audited 40, positions run from second to thirty-second, with ten of the seventeen inside the top five. Ahold's pumpkin spice finishing butter lists pumpkin second, behind only the butter itself. Toll House's latte morsels list pumpkin puree second of three inside the flavored chunks.

Two limits are worth keeping in view.

The first is the two percent rule. Ingredients present at 2% or less by weight are exempt from the ordering requirement when they're grouped at the end after a phrase like "Contains 2 percent or less of." Once a list reaches that block, position stops ranking anything — everything inside it is a small, unordered set. So a pumpkin listed at thirty-second of forty-seven, as it is on the Nutri-Grain bar, is telling you something quite different from a pumpkin listed third.

The second is that percentages are optional. If a manufacturer chooses to print one, it has to appear in parentheses after the ingredient as a percent by weight — but nothing requires them to print it, and almost nobody does. That's precisely why a figure like Starbucks's 2% comes from a company statement rather than off a package, and why ingredient position is usually the best signal available.

There's also a place pumpkin likes to hide. In four of our seventeen positive rows, it sits one level down inside a compound ingredient: Lucerne's eggnog carries pumpkin puree inside a "pumpkin spice egg nog base," Market District's caramel popcorn carries pumpkin powder inside a "pumpkin pie seasoning," 365 Everyday Value's latte bites carry it inside a "pumpkin spice latte seasoning," and the Toll House morsels tuck it into those flavored chunks. A quick scan of the top-level list alone would have marked all four as pumpkin-free.

One more quirk explains something you'll notice immediately on a real package. FDA's compliance policy on labeling of seasonings allows the collective term "spices" to stand in for the individual spice ingredients, while non-spice ingredients must be named. So a pumpkin spice product's ingredient list will often read simply "SPICES," naming neither the cinnamon nor the nutmeg. The ingredient list, ironically, tells you far more about the pumpkin than it does about the spice.

Now to the finding that surprised us most, and the rule that explains it.

Here is the pattern in the data. Among the 158 products whose names carry the word "flavored," 35 contain pumpkin — 22.2%. Among the 1,464 products without it, 671 contain pumpkin — 45.8%. Products that advertise a flavor contain the actual food at half the rate of products that don't. If your instinct was that "flavored" is just cheerful packaging language, the numbers point somewhere else.

It isn't a coincidence, and it isn't sloppiness. It's the regulation working as designed.

Federal rules treat the flavor named on a package as a regulated statement rather than free marketing copy. When a label makes any direct or indirect representation about the primary recognizable flavor — by word, or by a picture of the food, or by any other means — that becomes the characterizing flavor, and how it must be named depends on what's actually inside. The result is something like a ladder, and each rung down means less of the real food:

• The plain flavor name. Where a food contains no artificial flavor simulating the characterizing flavor, the name on the front panel is accompanied by the flavor's common name, in letters at least half the height of the food's name.
• "Natural [X] flavored" or "[X] flavored." This is the rung that governs most of the pumpkin spice shelf. It applies when a food is commonly expected to contain a characterizing ingredient and it contains natural flavor derived from that ingredient plus either an amount of the ingredient insufficient to characterize the food on its own, or none of the ingredient at all. The regulation's own example is "strawberry flavored shortcake." The word "flavored" is doing specific work: it signals that the flavoring, not the food, is carrying the character.
• "With other natural flavor." Where both the real characterizing ingredient and other natural flavor reinforcing it are present, that phrase follows the name of the food.
• "Artificially flavored." Where none of the natural flavor comes from the food being simulated, the product is labeled either with the flavor it actually derives from, or as artificially flavored.

Which reframes "flavored" entirely. It isn't decoration. It's a required disclosure, and it's telling you that pumpkin is either absent or present in an amount too small to characterize the food on its own. Our 22.2% versus 45.8% split is that rule showing up in the aggregate, exactly as you'd predict.

Flavor wording is only one of the phrases on a package doing more work than it appears to. The gap between what a label implies and what it actually commits to shows up across the claims food brands put on their labels.

Every count has edges, and ours has a few worth stating plainly.

The data reflects labels on file, not labels on shelves. FoodData Central says that information in Branded Foods is received from food industry data providers, and that downloads are generated every six months and reflect the most up-to-date version of each product at the time the download is generated. The records in our 40 date from March 2021 to November 2023. Seasonal products are reformulated and discontinued constantly, so any individual row is a snapshot with a date on it, which is why the date is printed in the table.

Reformulation is the quiet variable. Recipes drift, sometimes dramatically, which is the whole story behind 1970s supermarket staples and their 2020s versions.

Some judgment calls are ours, and you might make them differently. We counted pumpkin seeds separately from pumpkin: 47 products in the population list pumpkin seeds or seed kernels but no pumpkin flesh, and we scored those as not containing pumpkin. Seeds are a different part of the plant, and a granola with pumpkin seeds isn't what most people mean when they ask whether there's pumpkin in it. That decision holds the popcorn-and-seed-snack category down at 26.4%; count seeds and it would rise noticeably. We also scored "pumpkin flavor" as not-pumpkin, which affects 21 products across the population.

And a boundary we held throughout: this audit reports what ingredient statements say and what the labeling rules require. It doesn't conclude that any product is mislabeled. A pumpkin spice creamer with no pumpkin in it is not doing anything wrong — it's a creamer named after a spice blend, and the spice blend never had pumpkin in it either. The point isn't that anyone is being deceived. It's that a shorthand everybody repeats turns out to describe about half the shelf, and the half it describes is predictable enough to be genuinely useful.

Which is a nicer place to end up than the trivia allows. Pumpkin spice season doesn't have to be a small annual disappointment about a spice blend. There's a real gourd in a lot of it, more often than not in the things you bake with or spread on toast, and the label will tell you which is which if you know what the words are doing.