In this article, learn about:
Why intended use decides which category a product falls in
How classification sets the labeling format, required item attributes, and product identifiers
How misclassification turns into receiving rejections, chargebacks, and warning letters
Walk the health, beauty, and wellness (HBW) aisle and it reads like one department. A moisturizer, a pain reliever, a bar of soap, a jar of multivitamin gummies, all within a few feet of each other, like one category split into five distinct flavors.
That impression isn’t accurate, however. Cosmetics, OTC drugs, soap, supplements, and medical devices are five separate regulatory categories that happen to share an aisle. Each one demands a different label, a different set of item attributes, and sometimes a different product identifier. What decides how a product gets regulated is intended use, set by the claims a brand makes and the ingredients it uses.
For instance, a moisturizer is a cosmetic until it claims to treat acne, at which point it becomes a drug. Or a gummy is a supplement until it claims to cure a disease, at which point it is misbranded.
Treat all of HBW as one bucket and the item data inherits that mistake. The result shows up as misbranding and FDA warning letters, as receiving discrepancies when the record does not match the product, and later as chargebacks and deductions. This piece gives a working model of all five subcategories, understanding that classification drives labeling, item data, identifiers, and every retailercompliance rule that follows.
How Does Intended Use Decide the Product Category?
Intended use is the answer to a simple question: What is this product for? Regulators read that answer from two places.
First, the claims. Everything a brand says about the product counts, not just the label. The marketing copy, the website, the social posts, and even a customer-facing FAQ can all establish intended use. A claim made on Instagram can reclassify a product sitting on a shelf.
Second, the ingredients. Certain active ingredients carry an intended use on their own. Put a recognized drug active in the formula and the product can be treated as a drug regardless of what the marketing says.
Take two products that look nearly identical on the shelf. A deodorant masks odor, which is a cosmetic job, so a deodorant is a cosmetic. An antiperspirant reduces sweating, which changes how the body functions, so an antiperspirant is an OTC drug with an active ingredient and a Drug Facts panel. The same split shows up in shampoo: A basic shampoo cleans hair and is a cosmetic, while an anti-dandruff shampoo treats a condition and is a drug.
A brand can move a product from cosmetic to drug without changing the formula, just by changing what it promises. The product is no safer or less safe than it was. It made a different claim, and the claim is what gives the product its classification.
The Five Categories, Side by Side
Here are the five classes that have legal definitions, laid out against the four things a supplier has to get right for each one: what puts a product in the class, how it has to be labeled, the item attributes the class demands, and the identifiers it carries.
Category | What puts a product here | Label format | Signature item-data attributes | Identifiers |
Cosmetic | Cleanses or beautifies, with no claim to affect body structure or function | Ingredient list (descending order), net quantity, responsible party | Ingredient list, claims flags, fragrance and allergen flags, net contents | GTIN/UPC |
OTC drug | A recognized therapeutic claim, or an active drug ingredient | Drug Facts panel: actives, purpose, uses, warnings, directions | Active ingredient and strength, dosage form, Drug Facts content, warnings | NDC, plus GTIN/UPC |
Dietary supplement | Supplements the diet; structure/function claims only, never disease claims | Supplement Facts panel, plus the required FDA disclaimer | Supplement Facts attributes, serving size, claims flags, allergen statements | GTIN/UPC |
Soaps | Nonvolatile matter is mostly an alkali salt of fatty acids, cleaning comes from those compounds, and it is sold solely for cleansing | CPSC consumer-product labeling rather than FDA cosmetic labeling | Cleansing-only claim, net contents, responsible party | GTIN/UPC |
Medical devices | Diagnoses, treats, or monitors without acting on the body chemically (diagnostics, wearables, at-home test kits) | FDA device labeling, scaled to device class; indications for use and directions | Device class, UDI, intended use/indications, premarket status | UDI (built on a GTIN) |
Cosmetics were long the lightest-touch category, judged mostly on ingredient honesty and accurate net contents. MoCRA ended that. Facility registration, product listing, safety substantiation, and adverse-event reporting now apply, so "cosmetic" no longer means low-obligation. It means a specific, and newly larger, set of obligations.
OTC drugs carry the heaviest data load of the five, because the monograph system dictates not just the label but the allowed actives and concentrations behind it. The clearest tell is the identifier: The day a product needs a National Drug Code (NDC), it has stopped being a cosmetic or a supplement, whatever the marketing still says.
Dietary supplements get wide latitude for structure-and-function claims and zero latitude for disease claims. A single sentence promising to treat a condition converts a lawful supplement into an unapproved, misbranded drug, with none of the monograph work that a real OTC drug would have done to earn the label.
Soap is carved out of the cosmetic category entirely, as long as it is genuinely soap: mostly an alkali salt of fatty acids, cleaning by virtue of those compounds, and sold for nothing but cleansing. Such products answer to the Consumer Product Safety Commission, not the FDA. The catch is how narrow that definition is. Add synthetic detergents or any claim beyond cleansing, like moisturizing, and it becomes a cosmetic; claim it treats acne and it becomes a drug. Most things labeled "soap" are not soap in the eyes of the law.
Medical devices are a separate category entirely, covering diagnostics, at-home test kits, and wearables with medical function. What defines a device is that it works physically or mechanically rather than chemically in the body, and that triggers a separate track: a risk-based device class, a Unique Device Identifier (UDI) alongside the retail code, and a premarket pathway such as a 510(k). Moving from supplements into a test kit is entering your product into a new category.
The Boundary Cases in Health, Beauty, and Wellness
The trouble starts with words that sound like classifications but aren't, and with products that refuse to sit in one class:
"Cosmeceutical" implies a regulated middle ground between cosmetic and drug, but no such category exists. A product is one or the other, and a brand using the word has usually just described a cosmetic while hinting at drug claims it can't make.
"Personal care" is also a shelf label with no legal weight.
Dual-status products genuinely belong to two classes at once. An anti-dandruff conditioner treats a scalp condition and conditions hair, so it carries the Drug Facts panel of an OTC drug and the cosmetic ingredient labeling of a conditioner. These are two full sets of obligations stacked on one item, and a record built for only one of them is incomplete by definition.
The trickier cases are when the physical inventory stays put but the classification shifts underneath it:
Claims creep: A product is received and shelved as a cosmetic, then marketing adds a line to the website promising it "clears breakouts." That single sentence reclassifies it as a drug while the inventory and item data still say "cosmetic." Nothing on the pallet changed, but the product is now misbranded, and the record no longer matches reality.
Ingredient-driven reclassification: Adding an ingredient with a recognized drug use, like fluoride or an acne-fighting active, can tip a product into drug status on its own, even if the marketing copy never changes. The claim isn't the only thing that signals intended use; some ingredients signal it by themselves.
In both directions the lesson is the same. Classification is not set one time at launch. It moves whenever the claims or the ingredients move, and the item data has to move with it.
Why Classification Is an Item-Data Problem
Up to this point the article has been about regulatory class. But a misclassification rarely costs you anything as a legal abstraction. It costs you when the product data is wrong, because the data is what retailers actually act on. Classification runs down a chain, and each link sets the next:
Intended use (the claims and ingredients) sets the subcategory: cosmetic, OTC drug, supplement, soap, or device.
The subcategory sets the labeling format: Drug Facts panel, Supplement Facts panel, cosmetic ingredient list, and so on.
The labeling format sets the required attributes: the specific data fields that label depends on, like active ingredient and strength, or serving size.
The required attributes set the identifiers: a GTIN/UPC for everything, an NDC once it's a drug, a UDI once it's a device.
The identifiers and attributes then have to be mapped to each retailer's own compliance rules. Retailers set their item-setup and receiving requirements themselves, often through shared data standards like GDSN, and those requirements track regulatory class without being derived from it.
If the item record says one thing and the product on the pallet is something else, that mismatch shows up at a specific point that costs you money:
A wrong or missing identifier stalls item setup. An OTC drug loaded without an NDC is an incomplete record, and the item usually won't make it cleanly through onboarding.
Missing required attributes trigger receiving discrepancies. If the class demands attributes the record doesn't carry, the shipment gets flagged at the dock and rejected.
A label that doesn't match the product's true class is a misbranding problem that sits with the physical product and its packaging, not the data record. The item data is a separate failure: When the record disagrees with the product, the cost is a retailer one (a rejected shipment or a chargeback), even if the label itself is fine.
Each of these issues show up as chargebacks and deductions and eventually becomes a fee. And because the deduction usually arrives weeks after the shipment, it is often hard to trace back to the classification error that caused it.
The Supplier Playbook in Health, Beauty, and Wellness
The risk in health, beauty, and wellness comes from treating classification as a labeling detail handled late, rather than the decision that drives the item record.
Five practices keep a product on the right side of that decision:
Classify by intended use before you build the item record, not after. The subcategory determines which fields and identifiers the record needs, so decide it first. Building the record and backfilling the class later is how products get set up as cosmetics that should have been drugs.
Carry the classification through the item data accurately and consistently across every retail connection. A product that is an OTC drug for one retailer is an OTC drug for all of them. The class and its attributes should match everywhere you sell, not drift from one connection to the next.
Keep claims, formula, and declared class in sync, and treat a claims change as a data event. When marketing adds a claim or R&D changes the formula, the classification can move with it. Route those changes back to whoever owns the item data instead of letting them ship straight to the website.
Validate that the required attributes and identifiers match the class under each retailer's rules. Confirm the drug items carry an NDC, the devices carry a UDI, and the class-specific attributes are populated the way each retailer's item setup expects.
Re-audit when marketing language or the formula changes. Classification is not a launch-day decision. Recheck it on any reformulation or claims update, because that is exactly when a compliant product becomes a misbranded one.
The common thread is that classification is a living attribute of the product, not a one-time label. Suppliers who treat it that way keep their records accurate as products change.
Where This Gets Operational
Getting the classification right is one thing, but carrying it cleanly to every retailer is another. That is the work SPS Commerce Assortment handles. It centralizes your product data, standardizes it to each trading partner's attribute requirements, validates it against their business rules to catch and correct errors, and delivers it in the format each partner expects. For a regulated product, that means the class-specific attributes a retailer requires get checked before the data goes out, rather than discovered as a problem at the dock.
When a mismatch slips through anyway, the cost lands downstream as deductions and chargebacks. SPS Commerce Revenue Recovery works the other end of that problem, identifying, validating, and disputing those deductions to recover the dollars behind them. It doesn't stop at recovery, though. It surfaces the root causes, including the product-setup and data issues that misclassification creates, so the same deductions stop recurring.
Getting the classification and the data right up front is what keeps a product on the shelf and out of the warning-letter and chargeback pile.