Amazon Fee Structure for Sellers

Amazon fee math is layered

Amazon margin is rarely just product cost plus a referral fee. The real stack usually includes referral fee, fulfillment cost, inbound freight or prep, returns pressure, and advertising.

That is why many products look profitable on a rough spreadsheet but weak after you model the full order path.

Common fee buckets

When Amazon still works

• You have enough gross room before ads.
• Fulfillment and inbound are stable enough to forecast.
• Your ad cost per order still leaves positive contribution.

The referral fee is the floor, not the whole story

Amazon's US pricing table shows that referral fees vary by category and can use a percentage, price tiers, and a minimum amount. A single 15% assumption is therefore not a reliable substitute for the SKU-specific preview.

FBA and inventory costs depend on the product and program. Record fulfillment, inbound placement, storage, aged-inventory, return, removal, or other charges only when the current preview, report, or transaction shows they apply.

Returns and advertising quietly erode margin

Two costs are easy to leave out of a spreadsheet and painful to discover later. Returns carry a cost even when Amazon handles them: you may not be able to resell the item, and in FBA you can still pay removal or disposal fees. A category with a high return rate needs a fatter margin to absorb it.

For advertising, use spend and completed orders under a consistent attribution definition. Keep paid and non-paid orders separate, then evaluate a blended scenario; dividing all spend by only attributed orders answers a different question than dividing it across every completed order.

Model the full order before you commit

The practical test is simple: take the selling price, subtract the referral fee, the FBA fulfillment cost, a share of inbound and storage, an allowance for returns, and your advertising cost per order, then subtract the product cost. What remains is your true contribution per unit. If that number is thin or negative, no amount of volume fixes it — scaling a money-losing unit just loses money faster.

Run this before you reorder inventory, launch a campaign, or drop your price to win the buy box. A product that clears the full fee stack with room to spare is one you can grow; one that only works when you ignore fulfillment and ads is a trap that gets more expensive as it scales.

Signs a listing is quietly losing money

• You have not recalculated true contribution since the referral fee or fulfillment tier last changed.
• Ad spend per order has crept up but you are still using an older, lower cost assumption.
• Return rate has risen but you have not added a bigger allowance to the margin model.
• The product only looks profitable when storage and inbound freight are left out of the math.

FBA versus self-fulfillment: a different fee shape

Choosing self-fulfillment instead of FBA does not remove the fee stack, it just changes its shape. You trade FBA's per-unit pick, pack, and ship fee and storage charges for your own warehousing, packing labor, and shipping costs, which can be cheaper for bulky or slow-moving items but harder to staff consistently at scale.

The referral fee and advertising costs stay largely the same either way. Before switching fulfillment methods to chase a thinner fee stack, model your own fulfillment cost per order honestly, including labor time, packaging, and the shipping rate you can actually get, not a best-case estimate.

A worked example: referral fee plus FBA on one SKU

For an explicit illustration, suppose a SKU produces $25 seller revenue and Amazon's current preview shows a $3.75 referral fee and $4.50 fulfillment fee. With $7 product cost, $1 inbound allocation, and $2 attributed advertising, modeled order contribution is $6.75 before overhead and omitted costs.

The $3.75 and $4.50 figures are example inputs, not claims about a size tier or current universal rate. Replace them with the exact preview and reconcile the output to a transaction.