Amazon coupon fees are separate from the discount you fund. Amazon's published US fee model, effective June 2, 2025, was $5 per coupon plus 2.5% of coupon sales; verify your current marketplace and account terms before budgeting.
For brands generating $1M+ a year on Amazon, the harder question is whether extra contribution pays for the discount given to shoppers who would have bought anyway. Higher orders can still mean less money. Start with that hurdle, then decide whether the coupon deserves a budget.
Key Takeaways
- The published US model combines a $5 creation charge and 2.5% of coupon sales. Neither replaces the customer discount.
- In the hypothetical example below, 100 discounted orders generate $95 in program fees and $400 in funded discounts, before ordinary selling expenses.
- In the same example, matching the contribution from 100 normal orders requires 170 coupon orders when the $5 fixed charge is kept at campaign level.
- Review 3 things together: contribution, credible incremental demand, and available inventory. A sales increase alone clears none of them.
What does the published Amazon coupon fee model include?
Amazon's US fee announcement, updated March 25, 2025, describes the move from $0.60 per redeemed unit to an upfront $5 per coupon plus 2.5% of coupon sales. A November 2025 Amazon moderator clarification also distinguishes those fees from the coupon budget that funds customer discounts.
Those are dated public records, not a promise that every marketplace or account uses identical terms today. Before launch, open the fee terms in your own account and retain the marketplace, date, fee base and applicable exceptions, including any caps or waivers. A public help-page link that shows only a sign-in screen or navigation is not enough to approve the budget. Use the published model below to understand the calculation, then replace its inputs if your terms differ.
There are three separate money movements. The shopper receives a discount. Amazon assesses the coupon's program charges. Your business still pays ordinary product, fulfillment, advertising and return costs. A report showing one of these is not necessarily showing the others.
The distinction matters when finance and marketing use different exports. Marketing may record the advertised discount budget. Finance may see a program fee later. An analyst can accidentally subtract the discount twice if revenue is already net of that discount, or omit it entirely by using the undiscounted price. Label every column before trusting the total.
How should you calculate the fee without charging $5 to every order?
Keep the fixed charge at coupon level, even when the number of redemptions changes. Under the published model:
Program fees = fixed coupon charge + percentage rate × the applicable coupon-sales base.
The account's definition of coupon sales controls that last input. Do not assume it means the entire ASIN's revenue, the advertised discount budget, or every order during the promotion. Confirm how it treats discounts, cancellations and refunds. Then reconcile the estimate with what Amazon actually charges.
Here is a hypothetical worksheet, not an Amazon fee quotation. Assume one coupon, 100 redeemed orders, a $40 normal price and a $4 funded discount per order. Assume the applicable coupon-sales base is the $36 paid by each buyer.
| Item | Calculation | Amount |
|---|---|---|
| Buyer revenue after the discount | 100 × $36 | $3,600 |
| Funded customer discount | 100 × $4 | $400 |
| Variable program fee | $3,600 × 2.5% | $90 |
| Fixed program fee | One coupon × $5 | $5 |
| Total program fees | $90 + $5 | $95 |
The $400 discount is already reflected in the difference between $4,000 at the normal price and $3,600 received. Do not deduct it again from that $3,600. It belongs in the comparison with the no-coupon scenario, not as a second expense against already-discounted revenue.
Allocated across these orders, the fixed charge is $0.05 per redemption. At only 10 redemptions, it becomes $0.50. That is why a universal claim that coupons become cheaper below a particular product price is incomplete: volume, the eligible sales base and the discount also change the result.
An old spreadsheet charging $0.60 on every redemption can produce a tidy answer to the wrong fee schedule. A new spreadsheet charging $5 on every redemption is worse. Keep the campaign-level and order-level charges visibly separate.
How many additional orders does the coupon need?
Continue the hypothetical example, this time comparing the promotion with the orders you would normally expect. Suppose a normal $40 order contributes $12 after product cost, ordinary Amazon charges, expected returns and the advertising allowance. Those assumptions imply $28 of non-coupon cost per order.
For this illustration only, hold that $28 constant. A coupon order then contributes $36 less $28 less the $0.90 variable coupon charge: $7.10 before allocating the fixed fee. At 100 orders, the $5 campaign charge reduces the average to $7.05. Real referral fees, advertising cost and returns may change with price or traffic; recalculate them rather than carrying this simplified assumption into a live budget.
If the brand would normally sell 100 orders in the same period, normal contribution is $1,200. Selling 100 coupon orders produces $705. The promotion has not created $705 of incremental contribution. It has reduced the period's contribution by $495 under these assumptions.
To match the $1,200 baseline, solve:
Coupon orders × $7.10 − $5 ≥ $1,200.
That requires at least 170 whole orders with these exact inputs: 170 × $7.10 − $5 = $1,202. The simpler per-order average of $7.05 suggests about 171 orders, but that average fixes the $5 allocation at the original 100-order volume. Use the campaign-level formula when quantity changes.
This is the decision the coupon dashboard cannot make for you. Can the offer credibly produce roughly 70 additional orders without raising the other costs or exhausting stock? If not, the extra revenue may merely make the loss harder to notice.
The calculation becomes less forgiving if paid traffic grows too. Suppose the extra demand needs more expensive clicks. The $7.10 pre-fixed-fee contribution is no longer stable. Rebuild the scenario with the new advertising allowance before increasing the budget. Our contribution-margin guide explains that advertising boundary; the coupon worksheet must sit beside it.
How can you tell whether those orders were incremental?
Coupon-attributed orders tell you that a coupon was used. They do not identify how many buyers would have purchased at full price. A loyal customer can redeem the same discount as a shopper the offer genuinely persuaded. Both appear in the promotion total.
For a practical first pass, write down a plausible normal-sales range before launch. Use recent comparable weekdays and note seasonality, availability, price, advertising and competitors' visible offers. If the ASIN normally sells between 80 and 110 units in a comparable week, test the contribution result against both ends. Do not select whichever baseline makes the coupon look best afterward.
A before-and-after comparison remains directional. The coupon week may coincide with a delivery improvement, a new main image or an advertising expansion. Another ASIN is not automatically a valid control either: a different review profile, price or stock position can explain the difference. Call the estimate uncertain when those conditions cannot be separated.
Where the scale of the decision justifies it, ask an analyst to design a defensible holdout or another suitable incrementality study using the controls actually available. Do not label a normal coupon toggle an Amazon randomized experiment. In smaller decisions, use conservative scenarios and a limited exposure budget rather than pretending the measurement is precise.
There is a useful stop condition here. If only the most optimistic normal-sales assumption makes the offer profitable, the evidence is not strong enough to scale it. Keep the test small or spend the money on the actual retail constraint. A discount cannot repair an unclear pack count or an unattractive delivery promise. Use the conversion diagnostic when the offer's underlying problem remains uncertain.
What belongs in the launch and stop decision?
Before activation, record the ASINs, coupon identifier, marketplace, intended dates, normal price, buyer price, current fee terms, discount budget and other active offers. Name the person who can stop the promotion and the person who reconciles charges. Retain a buyer-facing check of the offer that actually appears, not just the settings you intended.
Check promotion combinations in the account. Do not assume a coupon will or will not combine with another discount because a different seller's screenshot says so. If a proposed price change is meant only to manufacture a saving, reject it. Reference-price rules and customer trust are constraints, not obstacles to work around. The price-suppression guide covers the related risk of conflicting channel prices.
Decide the economic floor before the first order. That might be a minimum contribution per order, a maximum total experimental loss, or a defined inventory-clearance objective. Set a review date and a response to an unexpected fee base. The coupon budget should not be treated as proof that all financial exposure stops at exactly that number; verify the current platform behavior and monitor actual charges.
Inventory can overturn a superficially good promotion. In a hypothetical case, the ASIN has 480 sellable units and a replenishment date four weeks away. Normal demand of 100 orders per week needs 400 units before replenishment. At 170 orders per week, the promotion would need 680 units across the same period, leaving a 200-unit shortfall if sustained. The extra contribution this week needs to be weighed against lost contribution during a later stockout. Inbound units without a credible sellable date do not remove the constraint.
At the review, keep the action specific: continue at the existing scope, narrow eligible ASINs, reduce exposure where the account permits, stop, or investigate a reconciliation difference. A slide saying “coupon performed well” does not authorize the next budget.
When can a lower-contribution coupon still be sensible?
For slow inventory facing storage expense, declining demand or an approaching commercial deadline, clearance can change the comparison. Full-price sales at the usual rate may not be the best realistic alternative. A coupon with lower contribution per order could preserve more cash than holding or removing the stock. Compare the alternatives explicitly, including recovery value and avoidable costs.
That is different from saying margin does not matter during clearance. It matters because you are choosing which loss to avoid. Record the inventory quantity, time window, expected recovery and decision owner. Do not let a temporary clearance exception become the standing price strategy for healthy products.
A launch can also justify a bounded learning expense. State what the test must teach, how much the brand will risk and when it stops. Reviews, rank and future repeat purchases are not free credits to add to the worksheet. Count their value only when you have defensible evidence, not because the initial arithmetic is inconvenient.
Where does ALFI fit, and where does it not?
We would rather stop a weak promotion than report a bigger sales number.
ALFI takes responsibility for the connected Amazon operation: advertising, listings, catalog, inventory, pricing and economics. Naeela leads strategy and accountability conversations, with a senior team owning delivery. Our 18-brand ceiling protects the attention needed to remember why a particular SKU should not be pushed this week.
That model fits brands generating about $5M+ a year on Amazon whose promotion decision repeatedly crosses departmental boundaries. Established Shopify or DTC brands generating about $10M+ annually can also fit when they are ready to build Amazon as a serious channel. It requires access to real product costs, stock truth and the authority to act. We will decline a PPC-only mandate that asks us to own profit while excluding the price and inventory decisions that determine it, even when that means losing agency revenue.
If your internal marketplace and finance leads can run this worksheet together, keep the work inside. If the problem crosses the whole operation, ALFI's operating model is the relevant conversation, not a promise of cheaper coupons.
Are Amazon coupon fees included in the customer discount?
No. Amazon's public clarification separates program fees from the coupon budget funding customer discounts. Record the discount through realized buyer revenue, then add the applicable program charges. Avoid subtracting the same discount twice. Check your current account terms because marketplace rules and the precise fee base determine the actual charge.
Is the $5 charge applied to every redeemed order?
In the published US model described here, $5 is charged per coupon created, not per redeemed order. Allocate it across actual redemptions when reviewing average unit economics. Keep it at campaign level when forecasting different volumes, otherwise a changing allocation can distort the number of orders needed to break even.
Does a coupon with higher sales necessarily make more profit?
No. You may discount orders that would have happened anyway, pay additional program charges and acquire more expensive traffic. Compare total contribution with a credible no-coupon baseline. Include returns and inventory consequences. Coupon-attributed sales describe usage; they do not, by themselves, identify incremental demand or establish a profitable promotion.
Should I use a universal product-price threshold for coupons?
No. The fixed fee per redemption changes with campaign volume, while the variable charge depends on the applicable sales base. Funded discounts, ordinary selling costs, normal demand and the value of alternative inventory actions matter too. A price-point shortcut can screen an idea, but it cannot approve the budget.
What should make an operator stop a coupon?
Stop or restrict it when actual contribution breaches the pre-agreed floor, charges cannot be reconciled, the buyer-facing offer differs from the approved offer, or demand threatens stock needed for more profitable sales. Use the controls available in your account. A limited learning or clearance exception needs its own documented reason.
What to do this week
Choose one proposed or active coupon. Reconcile its current fee terms, actual buyer revenue, funded discount and program charges. Build a normal-demand range and calculate the order hurdle at both ends. Put the result beside sellable stock and the next credible replenishment date.
Then make one decision: keep it, narrow it, stop it, or buy a specific piece of missing evidence. More orders are useful only when you know what you paid to get them.
