A family bakery came to us this spring wanting to put its cakes on Amazon. The cakes carry a margin of around 60%, and the owner was clear-eyed about the catch: when a box arrives melted, that order goes to zero or to a loss. By the end of that first call, the conversation had turned to the fondant they make to decorate those cakes, which keeps for a long time, ships without drama and carries great margins. Within a few weeks we were launching the fondant, and the cakes were still on the list for later.
That call is a fair summary of what selling food on Amazon asks of a brand. Amazon will happily sell your food. It just runs on its own rules about dates, heat and pack size, and those rules have nothing to do with how your retail buyers taught you to think about the business.
Selling food on Amazon profitably means building the channel around three food-specific constraints before you spend on ads: a pack and price that clear a 30% margin before advertising, inventory that reaches Amazon's warehouse with more shelf life left than your category's minimum, and a fulfillment plan for heat. Brands already stocked in grocery stores should run Instacart next to it, because shelf space is what buys cheap clicks there.
This page is for food, beverage and pet brands that are already on shelves, already making money somewhere, and wondering why Amazon hasn't followed. If you're a first-time seller, the approval section below will get you started, but most of what follows assumes you've got product, distribution and a P&L that somebody upstairs reads.
Should a food brand sell on Amazon if retail is the main channel?
Yes, because your Amazon page is a reference page whether or not anyone at your company manages it. In FMI's 2026 grocery shopper research, 39% of shoppers said they run a web search on their phone while shopping in person, and in my experience a shopper who finds you somewhere else double-checks Amazon close to nine times in ten, even when they buy you at the store. Retail buyers look you up before a category review. Investors look you up in diligence.
Across our whole client base, not one brand's own ecommerce store outsells its Amazon. These are brands with good sites and healthy email lists, and it still holds, because their site is where people go once they want them specifically, and Amazon is where people go while they're still deciding.
The risk of ignoring it is concrete for food. A brand we audited this summer saw its margin on a hero product drop to about 10% for a stretch because a reseller had picked up short-dated product and listed it, and the brand had to lower its own price to compete until the reseller sold out. Nobody at the brand had chosen to be in that fight. I make the longer case, including what a minimum presence looks like if Amazon isn't your priority this year, in Why Sell on Amazon When Retail Is Your Channel?.
What does Amazon require before you can sell food?
Amazon gates food at the product level, so the first step is to search each item in Seller Central's Add Products screen and see whether it says "Sell this product" or "Apply to sell." Amazon's own grocery guide puts it plainly: "If you see an Apply to sell button, then you need approval." Restrictions can sit on an individual ASIN, a brand, a category or a sub-category, and some products need more than one approval.
Per Amazon's "Categories and products that require approval" page, an application usually asks for:
- Purchase invoices dated within 180 days
- Clear product photos showing all sides
- Brand authorization letters, if you're not the brand owner
- Safety certifications and, for some categories, commercial liability insurance
Amazon says review takes 2 to 7 business days. A "Your account does not qualify" response can't be appealed.
Approval is the easy part. Amazon's food and beverages policy is where an established brand has to slow down, because it reads your label and your listing the way a regulator would:
- The label. The full FDA-compliant label in English: brand name, statement of identity, net quantity, the manufacturer, packer or distributor's name and address, an ingredients list by weight, an allergen declaration where it applies, and a Nutrition Facts panel. Organic products need the "Certified organic by" statement from an accredited certifier, and the policy points to federal food law, including 21 CFR 101 for labeling and, as applicable, the current good manufacturing practice rules in 21 CFR 117.
- The listing attributes. Every food listing has to declare that it contains food, its shelf life, whether it's expiration-dated and how the date appears on the product. Once your inventory lands, Amazon tracks its dates off those three expiration fields.
- The claims. No "FDA approved," no FDA logo, nothing that says the product treats or prevents a disease, and no "organic" without naming the certifier on the listing.
- The offer. Product must be new and sealed in the manufacturer's packaging, and units inside a larger sealed, labeled box can't be sold individually.
That last rule matters for multipacks, and we'll come back to it.
How much shelf life does Amazon require for food?
Amazon requires every expiration-dated product to arrive at its fulfillment center with enough shelf life for a customer to finish it, plus 90 more days, and it publishes much longer minimums for most food categories. Amazon's example is a supplement bottle with 240 daily servings, which needs 240 + 90 = 330 days left on arrival.
The 90-day shelf-life rule every guide quotes is Amazon's minimum for chips. For flour and baking supplies it's 270 days, for pasta it's a year, and for candy it's 540 days. Google's AI Overview for this exact search tells sellers "at least 90 days upon arrival," which is right for a bag of chips and wrong for most of the pantry. This is Amazon's own table, read off its "Expiration dates on FBA products" help page on October 6, 2026 (Amazon notes the list is a guide only):
| Product type | Minimum shelf life left when it arrives at Amazon |
|---|---|
| Chips and other snacks | 90 days |
| Dried fruit | 180 days |
| Cereal and granola bars | 180 days |
| Pet food and treats | 180 days |
| Beverages such as coffee, tea, soda and juice | 270 days |
| Salad dressings and oils | 270 days |
| Flour, spices and baking supplies | 270 days |
| Peanut butter, jams and preserves | 365 days |
| Pasta | 365 days |
| Condiments, including salt and pepper | 365 days |
| Candy | 540 days |
| Canned food | 540 days |
| Herbal or mineral supplements | 730 days |
To find your own number, take the larger of two figures: your category's minimum from the table, or the days it takes a customer to finish one unit plus 90. Then add the time between your production run and check-in at Amazon. Most brands have never written that number down. If a brand makes a baking mix with a 12-month shelf life and its product spends four months between the line and Amazon's dock, it has about 240 days left on arrival, and it's already under the 270-day minimum before a single unit sells.
How should food be dated and labeled for FBA?
Every ingestible product needs an expiry or manufacturing date on the unit, and lot numbers alone don't count. Case packs, multipacks and display boxes need the date on the outer packaging and on each item inside. Amazon accepts MM-DD-YYYY, MM-YYYY, or a written month like 15-MAR-2026, and treats "best by" and "sell by" as expiry dates. Two details catch brands out:
- A month-only date is read as the first day of that month. Print "06-2027" and Amazon counts it as June 1, so you lose up to a month of shelf life on paper.
- Each box can hold only one expiry date per ASIN. Two production runs of the same SKU go in separate boxes.
Product that's past its date when it arrives is marked for disposal and can't be returned to you, and inventory that gets close to its shelf-life limit while it sits in the warehouse is set to unsellable. Amazon can send that unsellable stock back to you automatically, but only if you've turned the setting on.
Why shelf life decides your replenishment rhythm
On Amazon, shelf life sets how often you can ship. One specialty better-for-you ingredient brand we work with ships quarterly, which is how its freight works, and the agency before us had planned inventory as if it shipped monthly. We rebuilt the plan around quarterly shipments with buffer and mid-quarter top-ups, and the account went from chronic stranded inventory to being limited only by how fast it could grow. That brand is now up 60% year over year, with conversion up from 16% to 26%, and the full story is in the specialty ingredient brand's turnaround.
The same math runs the other way. Ship a quarter's worth of a 90-day product and you'll be writing some of it off.
Can you sell chocolate and other heat-sensitive food through FBA?
Only in the cooler months, and the list of what counts as "meltable" is longer than most brands expect. Amazon defines meltable products as heat-sensitive consumables at high risk of melting at storage or transit temperatures of 155°F or above, and its current Meltable FBA inventory page names gummies, protein bars, cookies and coated candy alongside chocolate. If you make a protein bar and you've never checked Amazon's Meltable ASINs list, check it this week.
For the 2026 season, Amazon's page says gummies, protein bars, cookies and coated candy can be sent to FBA starting August 31, 2026, with Amazon fulfilling them from September 14, and all chocolate, chocolate-coated nuts and fruit, caramel and gelatin gummies can go in from September 21, with fulfillment from October 5. Meltable stock still in Amazon's warehouses past April 15, 2027 may be disposed of by May 1, 2027. Amazon re-dates this page every season. Pull it fresh before you plan a shipment.
Even if you send Amazon a letter on manufacturer letterhead saying your product holds up at 155°F, Amazon still reserves the right to suppress the offer if customers complain that it arrived melted. So the safest plan for anything that can soften is self-fulfillment through the summer. No chocolate company we know runs FBA through July. They ship it themselves, and when a listing starts getting flagged, they pause it through the worst weeks and switch it back on in autumn. Seasonal pausing is a normal operating pattern for these products.
Self-fulfillment on Amazon (FBM, or Merchant Fulfilled Network in Seller Central) has two rules that decide whether it works:
- Set handling time honestly. The bakery from the top of this page ships only Monday through Wednesday, because weekend transit was damaging a perishable product, and they assumed Amazon would force them to ship daily. A three-day handling time let them keep the rhythm they'd spent years working out.
- Keep your late shipment rate under about 4%. Above that, Amazon can take away self-fulfillment and leave you on FBA only. We've had to win those privileges back for brands, and Amazon judges you on when the tracking number reaches it, so a box that left on time with tracking uploaded a day later still counts as late.
The full cost comparison, including the review protection FBA gives you that self-fulfillment doesn't, is in FBA vs FBM: the shipping cost most brands forget to add back.
What pack size and price should a food brand use on Amazon?
Price every SKU to a pre-ad gross margin of at least 30%, after landed cost and Amazon's fees and before any advertising, and expect to get there with an Amazon-specific pack more often than with a price increase. Under 20%, the product can't carry the advertising it needs to grow, and we decline brands that insist on running margins that thin. The full margin bands, and the subtraction behind them, are in the 30% floor for Amazon profit margin.
The pack question is where food brands lose the most money without noticing, because the retail pack was designed for a shelf and a planogram. My usual line on calls is that you're probably not going to sell the same pack size on Amazon that you sell in retail, or even on your own website. A single unit with a low price can't absorb a per-unit fulfillment fee, and it puts your Amazon price in direct comparison with every grocery shelf in the country.
So the Amazon catalog usually looks different:
- Multipacks and variety packs lift order value, so the fulfillment fee is spread across more product. When we took over a natural snack brand's account, it was built almost entirely on one-pack listings, and part of the fix was launching multipacks and variety 3- and 6-packs on its best sellers. The founder later told us profit was running about 10x what it had been with the previous agency. That story, with the ad numbers, is in the natural snack brand's turnaround.
- Ship a multipack in as a set. If a 6-pack goes into Amazon as six loose units in a master case, Amazon can break the case and sell singles. Box, tape or bag the set and label it as one unit. Amazon's food policy backs this from the other side: units inside a sealed, labeled box can't be sold individually.
- Subscribe & Save belongs on almost every consumable. On the snack account it lifted subscription revenue about 22% in a single month, and on the specialty ingredient account it now runs around 12% of monthly Amazon revenue.
- Keep per-unit price consistent across channels. Amazon checks prices at other retailers and can pull your Buy Box when a single unit somewhere else undercuts the per-unit price of your Amazon multipack. That trap, and the version a distributor can spring on you, gets its own page in The Price Parity Trap.
Fees move by category, size and price point, so pull your own rate in Amazon's revenue calculator before you set a price. I walk through how the fee lines stack in Amazon FBA Fees. If you'd like to see that floor applied to your ad spend, our free four-week trial does it in your own account: we cut the spend that buys sales below your margin floor and move the budget onto terms that earn.
Do Amazon reviews still combine across flavors?
Not reliably. Earlier this year Josh Grillo, our lead brand manager, walked a client through a listing where Amazon had stopped pooling reviews across flavors, so one flavor showed 7 reviews and another showed 89. For a food brand, that turns every new flavor into a new product with a thin review count, sitting next to its own best seller.
His workaround is to test switching the variation theme, to size for example, and he's careful to say test, because a strong flavor can do better standing on its own. The number to protect is 4.3 stars. Between 4.3 and 4.7, Amazon displays a 4.5-star rating, and at 4.2 the display drops to 4 stars and conversion falls away. With an organic review rate of about 2%, a brand that waits needs roughly a hundred sales for every two reviews, so a new flavor needs a review plan from launch day. The tactics we use, including Vine, are in what Amazon Vine costs and what it's worth.
When is your ingredient a better Amazon product than your finished food?
When the thing you make in order to make your product is more shelf-stable, cheaper to ship and higher margin than the product itself. Back to the bakery. Their finished goods were perishable, freight-heavy and risky in summer. The inputs they already made every week, fondant, gum paste, modeling chocolate, royal icing mix and meringue powder, keep for a long time, ship flat and carry better margins. The owner put it better than I can: they wanted to take the ingredients they already had, package them and sell them, instead of only creating the end product.
It changes who the customer is, too. Someone buying a finished cake is buying an outcome, while someone buying professional-grade fondant is buying the means to make their own, and on Amazon that second customer searches, compares and reorders. The bakery's category had a leader that had held it for years more on marketing than on product quality, and a better product with a clean listing can take ground from a leader like that.
Four questions will tell you whether you're sitting on one of these:
- Does anything you make for your own production have a long shelf life on its own?
- Would a home cook, baker or smaller producer buy it if it came in a sensible pack?
- Does it clear the 30% pre-ad margin floor at an Amazon pack size?
- Is the category leader winning on product, or only on having been there first?
If the answers come back yes, the better Amazon business might be one step back up your own production line.
Is Instacart better than Amazon for a food brand?
Instacart pays better than Amazon when your products are already on grocery shelves, because you can only advertise on Instacart against products stocked at its partner retailers, so the brands bidding against you are only the ones already on the shelf beside you. On Amazon, anyone can invent a brand this afternoon, ship in a pallet and bid on your terms by Friday. Instacart's auction is closed to anyone without shelf space, and across the accounts we run, a click costs less there than the same brand pays for comparable intent on Amazon or Google. If you spent a decade getting onto those shelves, that's an advantage no ad budget can buy, and the work it took was harder than anything we do.
The cleanest example we have is Tosí. The brand was in Costco and doing almost nothing on Instacart, and under our management it grew 20x on Instacart sales. In Trevor Hults's words: "By month two our sales were more than 10x over the baseline that we had previously with a ROAS of 4.5."
Side by side, the two channels look like this for a food brand:
| Amazon | Instacart | |
|---|---|---|
| Who can bid against you | Anyone who can list a product | Only brands stocked at Instacart partner retailers |
| What it costs you to enter | Approval, labels, shelf-life and FBA prep | You're already in if a partner retailer stocks you |
| Does paid traffic build organic rank? | Yes, sales velocity from ads feeds organic ranking | No, ad share stays high and you can't taper spend the same way |
| Where most sales happen | Search results and the detail page | The first couple of rows a shopper sees, which Instacart's team puts at roughly 70% to 80% of platform sales |
| Where the budget comes from | Marketing | Often trade and promotions, because you can read the result by SKU and by retailer |
| The signal it's worth managing | Any brand with a sound margin floor | Above roughly 1,000 Instacart partner doors, and almost always above 2,000 |
One caveat matters more than the cheap clicks. Instacart ads don't build organic rank the way Amazon ads do. On one specialty ingredient account we run, ad-attributed sales stay consistently above 80% of the total. That's normal for the channel, and it means you budget Instacart as an ongoing cost.
Two quick tests tell you whether Instacart is already paying you. If you're in Costco and running nothing on Instacart, Costco tends to overperform on Instacart compared with its share of a brand's store sales, so that's usually the biggest single gap in the account. And if you switched on Instacart's automated ads and they're profitable with nobody steering, that's the clearest sign managed spend would return several times more. I'll admit we were slow to take Instacart seriously ourselves. The full case, including the trade-budget argument, is in Instacart advertising for CPG brands, and the newer ad formats are covered in Instacart Carrot Ads: What Changed and Who It's For and the setup in Instacart Is the Channel Retail Brands Keep Leaving on the Table.
Who should run Instacart for a food brand?
Whoever runs it has to own the broker contract question before the first campaign goes live, because Instacart places your ads against a SKU wherever it's stocked, and you can't switch off a retailer inside the platform. A broker contracted for the natural channel and not for Costco or Target can end up driving sales into retailers they aren't paid on. Some retailers, Sprouts among them, run separate sub-accounts, and a different pack size per retailer counts as a different SKU. Those are the levers you've got. It's solvable in the contract, as long as it's solved before the first month's report raises it.
For most brands the choice is between an in-house person, the broker, or an agency like ours, and the honest answer depends on who already owns Seller Central and trade spend. We're an Amazon agency first. Instacart is how we follow a brand onto the next shelf once Amazon is holding, and below about 1,000 partner doors we'll tell you it can wait. The trade-offs between the three options get their own page in Who Should Run Your Instacart: In-House, Your Broker, or an Agency. If you're already on Instacart and want to see how we'd run it, our services page lays it out.
Where does Faire fit next to Amazon for a food brand?
Faire is a wholesale marketplace for selling into independent retailers, so it sits next to Amazon as a route to the health-food stores and specialty shops that build a brand's reputation before the national chains do. It works differently from Amazon in ways that surprise brands, starting with the commission: a retailer who finds you on the Faire marketplace carries a 15% commission, and one you bring to Faire yourself through your Faire Direct link carries none. That, and the rest of the setup, is in Faire Direct: the difference between paying 15% and paying nothing.
The sequencing is the part food brands get wrong. One founder told us on a call that Faire is how she reaches every independent health-food store and functional-medicine practice, with the national chains a year or two out. Building that ground game first gives the chain buyer something to look at, and the sequence is laid out in How to Sell on Faire: The Sequence That Builds a Wholesale Business.
What should a food brand fix first on Amazon?
Fix the economics before the ads: the margin floor per SKU, then the pack, then shelf life and heat, and only then the advertising. That's the order we run every food account in, because ad money spent on a pack that can't clear its fees, or on inventory that's about to go unsellable, buys revenue you can't keep. If your Amazon sales are up and your profit isn't, start with the 30% floor and run your own catalog through it.
If you'd rather watch us do it, apply for the free four-week trial. We read your ads data before we answer, and if we can see the profit in your account, we run your Amazon ads for four weeks at no cost. You'll know by the end whether the channel can pay you. If you're already strong on Amazon and Instacart is the open question, start with our services page instead.
Food brings rules to Amazon that most categories never deal with, and the brands that do well with them are often the ones that already did the unglamorous work of getting onto shelves. If that's you, you're closer than you think.





