The fees that wreck Amazon margins are rarely the ones sellers price for. Everyone builds the referral fee into their model. Almost nobody builds in the aged inventory surcharge, the inbound placement charge, or the returns processing on a category with a twenty percent return rate. Those land in the settlement report, get netted against your deposit, and never appear as a line item anyone looks at.
Here are eight, in roughly the order sellers discover them, which is not the order of size.
1. The referral fee, which is not one number
Sellers quote “fifteen percent” as though it applies across the board. Amazon’s published fee schedule at sell.amazon.com/pricing is category-dependent and sometimes price-dependent within a category. Home and Kitchen is fifteen percent. Electronics Accessories is fifteen percent on the portion of the sale price up to one hundred dollars and eight percent on the portion above. Clothing and Accessories is tiered at five, ten, and seventeen percent depending on the price band.
If you sell across categories and model a single blended rate, your margin forecast is wrong by the spread. For a catalog spanning apparel and electronics accessories, that spread is wide enough to flip a product from profitable to not.
2. The thirty cent referral minimum
Most categories carry a minimum referral fee of thirty cents per item. On a twenty dollar product this is invisible. On a four dollar add-on item, thirty cents is seven and a half percent, not the category percentage you budgeted, and it does not scale down.
This is the single most common reason a low-price SKU that looks fine in a spreadsheet loses money in reality. Anything under about six dollars deserves a hand calculation before you list it.
3. Fulfillment fees that do not scale with price
FBA fulfillment is priced on size and weight, not on what you charge. That means the fee is close to fixed while your margin is variable, so the cheaper the product, the larger the share of revenue fulfillment takes. Amazon documents the current fulfillment cost structure in its Fulfillment by Amazon material, and the specific per-unit rates change on a published schedule.
Two practical consequences. First, any fee figure you memorized is a figure with an expiration date, so date it or recheck it. Second, the size tier boundaries matter more than the rates, because a product that crosses a dimensional threshold by a quarter inch of packaging can jump a tier. Repackaging to stay under a boundary is a real margin lever, and it is one of the few that costs almost nothing to pull.
4. Monthly inventory storage, which is seasonal
Storage is charged on the volume your goods occupy, and the rate is higher in the fourth quarter than the rest of the year. Sellers who build inventory in October for a holiday push are paying the elevated rate on every unit that does not sell through.
The failure mode is subtle. A product with a good unit margin and a slow sell-through can carry a perfectly healthy per-unit profit and still be a bad use of capital, because the storage accrues monthly against a unit that is not moving. Unit margin does not capture this. Only a view that ties storage cost back to the SKU does.
5. The aged inventory surcharge
Units that sit past defined age thresholds attract an additional long-term surcharge on top of monthly storage. This is the fee that punishes the decision you already regret, and it compounds: slow product accrues surcharge, surcharge makes liquidation more attractive, liquidation at a discount confirms the product was a mistake.
The operational answer is unglamorous. Set an age alert, and act on it while the discount required to move the units is still small.
6. Inbound placement charges
Amazon charges for distributing your inbound shipment across its network, with the amount varying by how you choose to split the shipment. Sending everything to one destination is cheaper on your end and more expensive on the placement charge. Splitting across multiple destinations flips it.
Sellers frequently pick the option that minimizes their own freight and never compare it against the placement charge that choice triggers. The two need to be evaluated together, and the answer differs by shipment size.
7. Returns processing
Return costs vary by category, and for categories with structurally high return rates the processing charge is a permanent line in your cost structure rather than an occasional event. Apparel is the obvious case. So is anything with sizing, fit, or color expectations.
The margin error here is treating returns as a deduction from revenue instead of a cost per unit sold. If a category returns at twenty percent, the true cost of selling a hundred units includes processing on twenty of them plus the units that come back unsellable. Price for the cohort, not the transaction.
8. Removals, disposals, and the cost of being wrong
Getting inventory out of the network costs money too, whether you remove it to your own address or dispose of it. Sellers rarely budget for this because it only happens on products that failed, and nobody plans for the product to fail.
The useful framing: every purchase order carries an embedded exit cost. On a product you are confident about, that exit cost is negligible in expectation. On a speculative launch, it is a real number and it belongs in the decision.
Where all of this actually shows up
Every fee above appears in the settlement report, netted against your sales, on a roughly two-week cycle. That report is the only place the complete picture exists, and it is formatted for reconciliation rather than for analysis. Most sellers glance at the deposit amount and move on, which is precisely how a fee that grew forty percent year over year goes unnoticed for three quarters.
If you have never taken one apart properly, this walkthrough on reading the two-week statement line by line covers what each section contains and which lines are the ones worth watching. The exercise is worth doing manually once even if you automate it afterward, because the manual pass is what tells you which of these eight is actually your problem.
The one habit that catches all eight
Track fees as a percentage of gross sales, by month, as a single trend line. Not the dollar amount, which grows with volume and tells you nothing, but the ratio.
Fee load as a share of revenue should be roughly flat if nothing has changed. When that line moves up two points and your product mix did not shift, something on this list is the cause, and the trend line will tell you which month to go looking in. It is a five minute monthly check that most sellers never set up, and it is the difference between finding a fee problem in April and finding it at year end.