Most Amazon sellers can tell you their revenue to the dollar and their referral fee to the percent. Very few can tell you what they actually keep on a single unit after everything Amazon takes. That number — contribution margin per unit — is the one that decides whether the business makes money.
There is a specific moment that repeats across hundreds of Amazon brands: the seller looks at a healthy revenue number, a reasonable gross margin, and a 15% referral fee, and concludes the product is profitable. Then the bank account does not agree. The gap between "looks profitable" and "is profitable" on Amazon is the full cost-to-serve — the complete stack of fees, fulfillment, storage, returns, and advertising that sits between the sale price and the money that actually lands. The referral fee everyone budgets for is usually only 15%. The real total runs 40-60% of the sale price for a typical FBA product once everything is counted. This guide tears the sale apart fee by fee, names the seven costs sellers routinely miss, and rebuilds it into a contribution-margin worksheet that shows what you keep per unit. The companion economics live in how to read your Amazon P&L and the revenue-vs-profit breakdown; this guide is the per-unit teardown underneath both.
The complete set of costs an Amazon sale incurs beyond product cost: referral fee, FBA fulfillment fee, storage, returns and reverse logistics, advertising (TACOS), and hidden costs like long-term storage, removal, and reimbursement leakage. Total cost to serve is typically 40-60% of the sale price for FBA sellers — far higher than the referral fee alone suggests.
Why revenue and gross margin mislead
Revenue tells you how much money moved, not how much you kept. Gross margin (sale price minus product cost) tells you part of the story but stops well short of Amazon reality because it ignores the platform's fee stack entirely. A product can show a comfortable 60% gross margin and still lose money once referral fee, FBA, storage, returns, and ads come out. The two numbers feel like progress while hiding the truth.
Contribution margin is the honest number. It starts at sale price and subtracts every variable cost directly tied to that sale — product, fees, fulfillment, returns allocation, and advertising — leaving the amount available to cover fixed overhead and produce profit. On Amazon, where the platform inserts itself into so many cost layers, contribution margin per unit is the only figure that reliably predicts whether scaling a product helps or hurts.
Revenue is vanity, gross margin is comfort, contribution margin is truth. The number that should drive every Amazon pricing, advertising, and SKU decision is contribution margin per unit — because it is the only one that accounts for the platform's full take.
The referral fee (and why it varies)
The referral fee is Amazon's commission on the sale, deducted automatically on every order. The headline rate everyone quotes is 15%, and that rate does apply to many categories — but it is not universal. Referral fees range from roughly 8% to 17% depending on category, some categories use tiered rates that change above a price threshold, and a handful carry a minimum per-item fee that hits low-priced products disproportionately.
The practical implication is that assuming 15% can quietly distort your unit economics in either direction. A category at 8% gives you seven extra points of contribution margin you might be leaving on the table through over-conservative pricing; a category at 17% takes two points you did not budget for. Always confirm the current referral rate for your specific category in Seller Central rather than carrying the 15% assumption across a catalog that spans multiple categories.
FBA fulfillment fees
The FBA fulfillment fee covers Amazon picking, packing, and shipping the unit to the customer. It is charged per unit and scales primarily with size and weight, which is why two products at the same price can have very different real economics — a light, small item and a heavy, bulky one pay very different fulfillment fees even at identical sale prices.
This is the fee that punishes oversized and overweight products hardest. A dimensional change that moves a product from one size tier to the next can add a meaningful per-unit cost, and that cost recurs on every single sale forever. Sellers who optimize packaging dimensions to stay within a lower size tier often recover more contribution margin than they would from a price increase, because the fulfillment saving compounds across volume without any conversion-rate risk.
Crossing a size or weight tier boundary can raise the FBA fee on every unit indefinitely. Before locking product packaging, check which tier the dimensions land in — shaving a fraction of an inch or an ounce to stay under a threshold can be worth more than a price change.
Storage: the cost that compounds
Storage fees are charged for the space your inventory occupies in Amazon's fulfillment centers, billed monthly and rising sharply during the Q4 peak period. For fast-moving products with healthy sell-through, storage is a minor line. For slow movers, it quietly becomes one of the most destructive costs in the entire stack because it accrues whether or not the unit sells.
The real danger is long-term storage fees, which apply to inventory that has sat too long. A product that overstays starts paying a surcharge on top of standard storage, and if it never sells, the seller eventually pays removal or disposal fees on top of everything already spent. Storage is the cost that turns a forecasting mistake into a compounding loss — which is why inventory discipline (covered in the inventory forecasting guide) is a margin lever, not just an ops concern.
Returns & reverse logistics
Returns carry costs that rarely show up in a simple fee calculation. When a unit is returned, the seller typically loses the fulfillment fee on the original order, pays return processing, and faces the reality that a meaningful share of returned units cannot be resold as new — they get graded down, liquidated, or disposed of. The return is not just a reversed sale; it is a reversed sale plus added cost minus recoverable value.
The way to handle this in unit economics is to allocate a returns cost per unit across all sales, based on the product's return rate. A product with a 5% return rate spreads its return-related losses thinly; a product at 25% (common in apparel and some electronics) carries a heavy per-unit returns allocation that can be the difference between a profitable and unprofitable SKU. Reducing return rate, covered in the return-rate playbook, is therefore a direct contribution-margin improvement.
Advertising as a variable cost
Advertising is where many sellers make their biggest costing error: they calculate margin before ad spend, then treat advertising as a separate marketing line item. On Amazon, where most products require ongoing ad support to maintain visibility, that approach systematically overstates per-unit profitability. Advertising is a variable cost of the sale and belongs inside the contribution-margin calculation.
The right metric is TACOS (total advertising cost of sales), not ACOS. TACOS measures ad spend against total sales, capturing the reality that ads influence organic sales too. If TACOS runs 15%, then on average 15 cents of every sale dollar goes to advertising and comes directly out of contribution margin. A product that looks healthy at 35% margin before ads can land at 20% after a 15% TACOS — still viable, but a very different business than the pre-ad number suggested.
The 7 hidden costs sellers miss
Beyond the headline fees, a set of smaller costs quietly erodes margin. Individually each looks minor; together they can add 5-15 percentage points of cost that never appear in a referral-plus-FBA estimate.
Surcharges on inventory that overstays. Turns a forecasting miss into a compounding monthly loss.
Per-return handling plus the lost original fulfillment fee. Scales directly with return rate.
The share of returned units that can't go back to new. Graded down, liquidated, or disposed.
Fees to pull or destroy aged or unsellable inventory. The final cost on a product that never sold.
Charges tied to how inventory is split and routed into the fulfillment network on the way in.
Penalties when stock runs too thin relative to demand. Punishes the cash-strapped seller twice.
Money Amazon owes for lost or damaged units that you never claim. A cost of inaction, not a fee.
The seventh is unique because it is money owed to you that you fail to collect — effectively a cost of not auditing. The returns and reimbursements guide covers how to recover it; for the purposes of true-cost accounting, unclaimed reimbursements function exactly like a recurring leak in contribution margin.
The contribution-margin worksheet
The worksheet is simple subtraction done honestly. Start at the sale price and remove each variable cost in order. Whatever survives is contribution margin per unit; divide by price for the percentage. The discipline is in counting every layer, not in the math.
The line-by-line order
- Sale price — the actual selling price after any standing discount or coupon, not the list price
- Minus product landed cost — manufacturing plus freight, duties, and inbound to Amazon, per unit (covered in the landed-cost guide)
- Minus referral fee — your category's actual rate, not an assumed 15%
- Minus FBA fulfillment fee — the per-unit fee for the product's specific size and weight tier
- Minus allocated storage — monthly storage divided across the units you expect to sell that month
- Minus allocated returns — return rate times the per-return cost, spread across all units
- Minus advertising — TACOS times sale price
- Equals contribution margin per unit — the number that should drive decisions
Run this for every SKU, not just the catalog average, because averages hide the unprofitable products that a strong hero SKU is quietly subsidizing.
A worked $30 example
Here is the worksheet applied to an illustrative $30 product to show how the layers stack. The exact figures vary by product; the structure is what matters.
| Line Item | Per Unit | % of Sale |
|---|---|---|
| Sale price | $30.00 | 100% |
| Product landed cost | −$7.50 | 25% |
| Referral fee (15%) | −$4.50 | 15% |
| FBA fulfillment | −$4.00 | 13% |
| Storage + returns allocation | −$2.00 | 7% |
| Advertising (15% TACOS) | −$4.50 | 15% |
| Contribution margin | $7.50 | 25% |
The product sells for $30 and keeps $7.50 — a 25% contribution margin, squarely in the healthy zone. But notice that the referral fee everyone budgets for is only one of five deductions, and that advertising alone matches it. A seller who modeled this product as "$30 minus 25% product cost minus 15% referral = 60% left" would have planned around $18 of margin and been short by more than $10 per unit. That gap, multiplied across volume, is the difference between a profitable brand and a confused one.
The referral fee everyone budgets for is only one of five deductions — and advertising alone usually matches it. Plan for 15% and you will be short by ten dollars a unit.
Pricing backward from margin
Once the worksheet is honest, pricing becomes a backward calculation rather than a guess. Instead of setting a price and hoping the margin works, start with the contribution margin you need, add back every variable cost, and the sum is your minimum viable price. Anything below it loses money once the full stack is counted; anything above it is your pricing headroom.
This reframes the most common pricing mistake. Cost-plus pricing — product cost plus a markup — ignores the platform's fee stack and routinely produces prices that look profitable on gross margin but bleed once Amazon's complete take and returns are counted. Margin-backward pricing makes the fee stack visible before the price is set, so a product that cannot clear the minimum viable price gets caught at the planning stage rather than after months of unprofitable sales.
Minimum viable price = target contribution margin + product cost + referral fee + FBA + storage allocation + returns allocation + ad cost. If the market will not bear that price, the product is not viable on Amazon at the target margin — better to learn that before sourcing inventory than after.
Common costing mistakes
Five mistakes show up repeatedly when sellers calculate the cost of a sale. All are preventable with the worksheet discipline.
Treating 15% as the cost of selling on Amazon. Reality is 40-60% all-in. Fix: run the full worksheet on every SKU.
Treating advertising as separate from unit economics. Fix: include TACOS as a variable cost inside contribution margin.
Counting only the units that stick, not the cost of the ones that come back. Fix: allocate return cost per unit by return rate.
A strong hero SKU hides unprofitable products inside the blended average. Fix: calculate contribution margin SKU by SKU.
Skipping long-term storage, removal, placement, low-inventory fees, and reimbursement leakage. Fix: add a 5-15 point allowance and audit reimbursements.
Defending margin against fee hikes
Amazon adjusts its fee schedule periodically, and fulfillment, storage, and placement changes can shift unit economics without warning. A brand that runs on thin contribution margins is fully exposed to these changes; a brand with a margin buffer absorbs them. Defending margin is therefore as much about preparation as reaction.
The three defenses
- Build a buffer — target the higher end of the 20-35% contribution range so a fee increase does not push a product negative. The buffer is insurance, not waste.
- Re-run economics on every fee announcement — when Amazon announces a change, recalculate the worksheet for affected SKUs immediately, starting with the thinnest-margin products since they are most exposed.
- Attack the controllable costs — you cannot control Amazon's rates, but you can control packaging dimensions (FBA tier), return rate (returns allocation), inventory discipline (storage), and reimbursement recovery (leakage). Each is a margin lever that offsets fee increases.
The brands that survive fee changes are not the ones with the lowest costs — they are the ones who know their true cost of a sale precisely enough to react before a thin SKU quietly goes underwater.
The Ecom Profit Box
11 step-by-step PDF guides covering Amazon profitability, listing optimization, conversion, and content strategy.
Grab it free →Know Your True Margin
Book a strategy call. I will help you build the contribution-margin worksheet for your catalog and find the SKUs quietly losing money.
Book a strategy call →The 7 Things to Remember About the True Cost of an Amazon Sale
- The referral fee is usually 15%, but total cost-to-serve runs 40-60% of the sale price once FBA, storage, returns, and ads are counted
- Contribution margin per unit — not revenue or gross margin — is the number that decides whether scaling a product helps or hurts
- Advertising is a variable cost and belongs inside the margin calculation, measured as TACOS, not treated as a separate line item
- Seven hidden costs (long-term storage, returns processing, unresellable returns, removal, placement, low-inventory fees, reimbursement leakage) add 5-15 points
- Run the worksheet SKU by SKU, never on catalog averages, because a hero product hides unprofitable SKUs inside the blend
- Price backward from the contribution margin you need; cost-plus pricing that ignores the fee stack routinely produces money-losing prices
- Target the higher end of 20-35% to build a buffer against periodic fee increases, and attack the controllable costs you can actually move