The promotion menu, briefly
Amazon offers several promotional formats, and they're easy to confuse, partly because the names have changed over the years. The differences matter because each has its own costs and eligibility rules.
- Lightning Deals run in short bursts, typically a few hours, on the deals pages with a claimed-percentage bar and a countdown. They get high visibility and urgency, and Amazon charges a fee for each deal.
- Best Deals run longer, often a week or two. The presentation is less urgent, but total exposure is far greater, and they also carry a fee.
- Coupons appear as a clickable badge in search results and on the detail page. The shopper ticks it and the saving applies at checkout. You pay the discount plus Amazon's coupon fee.
- Prime Exclusive Discounts are limited to Prime members and show as a price reduction instead of a badge to click. They're usually free to run outside major events, with a fee during peak events such as Prime Day.
- Promotions and price discounts are the plain sale price with a strikethrough, subject to eligibility based on your recent pricing history.
Visibility roughly follows cost, by design. Lightning Deals during a major event buy the most attention and charge the most for it. A coupon is cheap, always available and much less prominent.
In Seller Central, deals, coupons and Prime Exclusive Discounts are each set up from the Advertising menu, which is also where the Promotions page for percentage-off and buy-one-get-one offers sits. Eligible ASINs and recommended deal prices appear when you start creating a deal, and the fee is shown before you confirm.
What each one actually costs you
Every promotion carries three separate costs, and most sellers only think about the first.
The first is the discount. It's the obvious cost, and it's bigger than the headline percentage suggests, because it comes off the top line while your costs stay the same. Twenty percent off the price is not twenty percent of your margin. On a product whose contribution margin is forty percent of the sale price, a twenty percent discount removes half your profit per unit.
The second is the platform fee. Lightning Deals and Best Deals carry a fee per deal that varies by marketplace and season and rises sharply around Prime Day and the Black Friday period. Coupons carry their own fee. Prime Exclusive Discounts are usually free outside event windows. Amazon changes these fees, and it changed them significantly in 2025, so read the figure Amazon shows when you schedule the promotion instead of relying on a number you remember from last year.
The third is the referral fee offset, which is the one piece of good news. The referral fee is a percentage of the selling price, so discounting also reduces what Amazon takes. That softens the blow by roughly the referral rate, commonly around fifteen percent of the discount amount. It's real and it belongs in the model, but it's nowhere near enough to rescue a bad promotion.
How the US deal and coupon fees changed in 2025
In an announcement posted on the Seller Central forums, Amazon changed its US promotion fees for non-peak days from June 2, 2025. Lightning Deals and Best Deals moved from flat up-front fees ($150 and $300 respectively) to $70 per deal per day plus 1% of deal sales, with the variable part capped at $2,000 per deal. Coupons moved from $0.60 per unit sold to an up-front $5 per coupon plus 2.5% of coupon sales. For Prime Day, the same announcement kept Best Deals and Lightning Deals at fixed fees of $1,000 and $500, applied the new coupon structure, and raised the Prime Exclusive Discount fee from $50 to $100 per PED.
Two practical effects follow. A short, low-volume deal is now cheaper to run than it was, while a big-selling deal pays more through the 1% variable fee. And a coupon now costs something even if nobody uses it, though only the $5. Fees in the UK, EU and other marketplaces are set separately, so check the figure in your own Seller Central account.
Under the old structure an unclipped coupon cost nothing. Since June 2025 in the US, it costs the $5 up-front fee and nothing more until shoppers redeem it. That still makes a coupon the cheapest way to test whether a price badge moves conversion on a listing. A failed coupon test costs you a few dollars, while a failed Lightning Deal has already charged its daily fee.
The reference price trap
This mechanic slowly damages sellers who promote often, and it gets far less attention than it should.
Amazon generally requires a promotional price to be at or below the lowest price your item has recently sold for, typically measured over the previous thirty days, often with a minimum percentage reduction on top. The rule exists to stop sellers inflating a price to stage a fake discount, which is reasonable. The side effect is that every discount you run resets the benchmark for the next one.
Run a twenty percent promotion in October and your reference price for November becomes the promotional price. Your list price no longer counts. To qualify for a Black Friday deal you now have to go below the October price. Do this three or four times in a season and you've walked the product down a staircase that's hard to climb back up, because raising the price later costs conversion and rank just as competitors are discounting.
Sellers hit this every year and say Amazon rejected their deal for no reason. The system is doing exactly what it was designed to do. The mistake happened weeks earlier, when someone ran a discount without thinking about what it did to the floor.
The fix is to plan a season's promotions together. Decide the lowest price you're willing to reach at the deepest point of the year, usually the main event, and work backward so that earlier promotions stay above it. Sellers who protect their reference price in September still have room to compete in November.
The break-even math, worked
This is the calculation that should come before every promotion. Take a product at $40 with a fifteen percent referral fee, $6 of FBA fees and $12 of landed cost. The numbers are an illustration; use your own.
At full price, contribution per unit is the selling price less the referral fee, the FBA fee and the cost of goods. That's $40 less $6 in referral fees, less $6 of FBA fees, less $12 of product cost, leaving $16 per unit.
Now apply a twenty percent discount, taking the price to $32. The referral fee falls to $4.80, the other costs don't change, and contribution becomes $9.20. You've kept fifty seven percent of your previous profit per unit.
To make the same total profit you now need to sell $16 divided by $9.20, or 1.74 times as many units. That's a seventy four percent volume increase just to stand still. Under the old flat Lightning Deal fee, $150 spread across, say, 300 units took contribution down to $8.70 and pushed the required lift to about eighty four percent. Under the 2025 US structure, a one-day deal selling those 300 units at $32 pays $70 plus 1% of $9,600, or $166, which is about $0.55 a unit. Contribution lands near $8.65 and the required lift is about eighty five percent, almost the same answer.
The same arithmetic makes coupons look much better. A ten percent coupon on the same product takes the price to $36 and the referral fee to $5.40. Under the old fee of roughly $0.60 per redemption, that left about $12 of contribution, and the required lift was $16 divided by $12, about thirty three percent. Under the 2025 US fee, 2.5% of $36 is $0.90 a unit, which leaves about $11.70 and a required lift of roughly thirty seven percent, before the one-off $5. Either way, a third or so more units is a target a decent listing can realistically hit, which is why shallow coupons are often the most defensible promotion on the menu.
Run this calculation on your own numbers before you schedule anything. The inputs take two minutes to gather and the answer often changes the decision. Sellers who do it once tend to keep doing it, because the first result usually surprises them.
Pre-promotion checklist
- Contribution per unit at full price and at the promotional price, with the current fee from the scheduling screen included.
- The volume multiple needed to break even, written down before the promotion starts.
- Stock cover: enough units to last through the promotion and the weeks after it without running out, since a stockout straight after a deal throws away the rank you paid for.
- The reference price the promotion will leave behind, and whether it blocks a bigger event later in the season.
- Whether the ASIN is in Subscribe and Save, and what the promotional price does to subscriber pricing.
- Ad budgets raised for the promotion window, so campaigns don't run out of money just as conversion peaks.
- The date the price returns to normal, already scheduled.
The halo argument, assessed honestly
The usual defense of an unprofitable deal is that the direct maths understates the benefit. Extra sales lift organic rank, the rank persists after the deal, the review count grows and the product comes out stronger. There's truth in this, and it's also the argument behind a great deal of avoidable loss.
What's true: velocity and conversion do influence organic placement, and a well-run deal often produces a visible rank improvement that outlasts the promotion. For a new product with no sales history, buying velocity to get past the cold start can be a sensible investment, and it's right to treat it as marketing spend instead of a margin failure.
What's less true: the effect fades, often within weeks. Rank gained through a temporary price is partly rank gained because the price was temporarily attractive, and when the price goes back up, conversion falls back toward its usual level. There's also a post-deal trough, because customers who would have bought at full price over the next fortnight have already bought at a discount.
Two more costs rarely make it into the halo argument. Deep discounts attract more price-sensitive buyers, and those customers tend to return goods more often and leave harsher reviews, because their expectations are set by what they paid. And on Subscribe and Save products, a promotional price can reset what long-term subscribers pay, which reduces the value of your most profitable customers without anyone noticing.
Use the halo argument where it applies: launches, real clearance and deliberate event participation. It doesn't justify promoting a mature product that was selling perfectly well at full price.
When deals genuinely work
In four situations the numbers regularly justify a promotion. It's worth being clear about them, because deals aren't bad in general.
- Clearing inventory that costs you money to hold. Aged stock picking up long-term storage charges, or units nearing an expiry date or the end of their season, carry a holding cost that makes a discounted sale better than a unit sitting in the warehouse. The right comparison here is with removal fees and write-offs, because full price is no longer on offer.
- Launching into a category where nobody knows you. Early velocity buys visibility that is otherwise slow and expensive to earn. Budget it as launch investment with a fixed end date.
- Taking part in major events. During Prime Day and the Black Friday window, traffic is unusually high and shoppers are actively hunting for discounts. Deal conversion in these windows can beat normal periods by enough to change the maths materially.
- Defending a rank position under attack. When a competitor discounts hard into your keyword and takes share, a temporary response can cost less than rebuilding rank later. It's a defensive move with a time limit and shouldn't become your pricing strategy.
All four have a purpose other than immediate profit, set in advance, with a measurable end point. Promotions go wrong when the goal is vague, when nobody plans the return to full price and when nobody agreed beforehand what success would look like.
Making the call
Before scheduling any promotion, answer four questions in writing. What is the contribution margin per unit today? What does it become at the promotional price, including fees? What volume multiple is needed to break even? And what are we buying if we don't hit it?
If the honest answer to the last question is "nothing in particular," don't run the deal. If it's inventory clearance, rank for a launch or event visibility you've decided is worth paying for, run it deliberately with a set budget and a return to full price already scheduled.
Then measure what happened. Units during the window will obviously rise, so look past them to total contribution across the promotion and the fortnight afterward, compared with a similar earlier period. That comparison is where the trough shows up, and it's the only number that tells you whether the promotion created value or just moved it around. Sellers who track it build their own record of which promotions work for their catalogue, which is worth more than any general rule about deals.
Amazon's own reporting helps here. Each deal and coupon has a performance view in Seller Central showing units, sales and, for coupons, redemptions, and the Business Reports under the Reports menu let you compare the same ASIN across the weeks before, during and after. Export both, put the fee in, and keep the result next to the plan you wrote before the promotion.
Frequently asked questions
Only when the volume lift clears the break-even, which is higher than most sellers assume. A 20% discount on a product with a 40% contribution margin typically needs sales to roughly double just to hold total profit flat, before the deal fee. Lightning Deals are usually worth running for inventory clearance, launch velocity or event visibility, and rarely as a profit play.
Two costs apply: the discount itself and Amazon's coupon fee. In the US the fee was long around $0.60 per redemption. Under the structure Amazon announced for non-peak days from June 2, 2025, it's an up-front $5 per coupon plus 2.5% of coupon sales. Amazon adjusts fees periodically and other marketplaces set their own, so confirm the current rate in Seller Central before modelling. Beyond the up-front fee, unredeemed coupons cost nothing.
Amazon generally requires a promotion price at or below the lowest price the item sold for in a trailing period, usually 30 days, and often a minimum percentage off as well. If you've already been discounting, you've lowered your own reference price, so the next promotion has to go lower still.
Indirectly. Deals drive sales velocity and conversion, and both influence organic placement, so a successful deal often lifts rank for a while afterward. The effect fades, and it isn't reliable enough to justify a loss-making promotion on its own. Count any rank lift as a bonus and build the business case on other grounds.
A coupon is visible to all shoppers as a clickable badge and carries its own coupon fee. A Prime Exclusive Discount is limited to Prime members, shows as a strikethrough-style saving and usually needs a minimum discount off your recent lowest price. Prime Exclusive Discounts usually carry a fee only during major shopping events.
Some promotion types can overlap, and when they do a shopper may get both discounts together, which can take the price far below what you planned. Before you schedule a deal, check what coupons and promotions are already active on the ASIN and work out the combined price in your break-even model.
