Most sellers do not have a pricing strategy on Amazon. They have a repricer, and the repricer has a floor, and the price is wherever the last competitor left it. That is a reaction, and a reaction to a market of sellers who are all reacting to each other ends in the same place every time: the floor, with the margin gone and the rank no better than it was.
A strategy is the decision above the repricer: where the product sits against the market on purpose, what a promotion is for, and what Amazon's own pricing rules will let it do. Those rules come first, because they decide who is allowed to be seen at all.
Amazon's Marketplace Fair Pricing Policy says it monitors the prices of items in its store, including shipping, and compares them with other prices available to its customers. A price it judges to be significantly higher than recent prices offered on or off Amazon is a practice that harms customer trust, and Amazon's response runs from removing the Featured Offer, through removing the offer, to suspending the account in serious or repeated cases.
The practical version of that policy is the Pricing Health page in Seller Central, which lists every offer currently ineligible to be the Featured Offer because it is not competitive with the lowest price at other reputable retailers, which Amazon calls the Competitive External Price, or because it is priced significantly above the recent prices Amazon has seen. An offer on that list is still live and still costs storage; it simply is not the one the buy button points at.
So the first input to a strategy is not the competitor on the listing. It is the lowest price for the same product anywhere Amazon can see, and a brand whose own site, or whose wholesale customers, undercut its Amazon price has set a ceiling on Amazon without meaning to. Channel price discipline is a pricing strategy in itself.
Against sellers on the same listing there are three positions, and each is a decision with a cost. Lead takes the Featured Offer most of the time, at the lowest margin, and invites the next seller to go a cent lower, which is how a listing races to the floor. It pays for a launch, where the share of the buy button is what ranks the product, and for lines where the seller has a structural cost advantage nobody else can follow. Match shares the Featured Offer on the other inputs, fulfilment, shipping speed, seller metrics and stock, and holds the price where the market is; it is the default for a seller whose account health is better than the competition's, because the tie goes to the better seller. Hold keeps the price above the market and accepts a smaller share of the button in exchange for margin; it works for a brand that owns the listing, controls its distribution and has the reviews, and it does not work for a reseller of a product ten other people also stock.
The mistake is choosing lead by default. A product that leads at a loss to win the button is buying rank with money it does not have, and the rank it buys belongs to the listing, which every seller on it shares. Which position a product is in should be a line in the plan, reviewed monthly, and the repricer's floor and ceiling should be set from it, which is the subject of
a repricing strategy for marketplace sellers.
A price cut is a small change to the price and a large change to the profit. The referral fee is a percentage of the price, so it falls a little; the fulfilment fee, the cost of goods and the advertising per unit do not move.
Worked through. A product sells at $30. A 15 percent referral fee takes $4.50, fulfilment $5.00, the product costs $9.00 landed and advertising averages $3.50 a unit, leaving $8.00. Cut the price by a tenth, to $27. The referral fee falls to $4.05 and nothing else moves, so the unit now makes $5.45. A tenth off the price has taken almost a third off the profit. The figures are illustrative; the fees are each product's own.
What the volume has to do. To earn the same profit, the product now has to sell about 47 percent more units. A cut that lifts sales by a fifth has made the business smaller while the sales report shows it growing.
So every cut answers one question before it is made: will it move units by that much, through the Featured Offer, conversion or rank, and for how long. For a launch the answer can be yes. For a mature product with a steady share of the button it rarely is.
Amazon itself prices dynamically on its retail offers, changing prices as demand, stock and competitors' prices move, and it makes a version of the same tool available to sellers for nothing. Automate Pricing is included in the Professional selling plan, needs an active offer on the product, and runs one of a handful of rules: match, undercut or sit slightly above the current Featured Offer; match, undercut or sit above the lowest price in the store; track the Competitive External Price beyond Amazon; a sales based rule that lowers the price when stock sits unsold; and a business rule that competes for the Business Featured Offer. A minimum price is required, a maximum is optional, and Amazon's own advice is to set both. Rule changes take up to an hour to apply and later price updates usually land in under 15 minutes.
The tool is not the strategy. Every rule needs a floor, and the floor is the strategy's first number: the price below which the unit loses money once the referral fee, the fulfilment fee, storage, returns and advertising are paid. Most sellers set the floor from the cost of goods and forget the rest, which is why
what it costs to sell on Amazon comes before this article, and
the fee schedule is the reference the floor is built from. A repricer that is allowed to go below the true floor is a machine for selling at a loss faster than a person could.
A promotion on Amazon is a temporary price, and it is worth being precise about what it costs and what it earns, because the fees changed in 2025. Coupons, the clipped discount shown on the listing and in search, moved on 2 June 2025 from a per unit fee to an up front fee of $5 per coupon plus 2.5% of the sales the coupon is redeemed on, as Amazon announced in Seller Central. A Prime Exclusive Discount has to be at least 10% off the non member, non promotional price, on a product in new condition with a rating of three stars or none, and the event versions of it, the Prime Day and holiday discounts, carry a flat fee per campaign that Amazon publishes ahead of each event.
What a promotion earns is a temporary lift in conversion, and conversion is what Amazon's ranking reads. So a coupon on a product that is already converting is mostly margin given away, while a coupon on a launch, or on a product whose sessions are fine and whose conversion is not, buys the sales velocity that moves rank, and the rank stays after the coupon ends. That is the test for every promotion: does it change a number the algorithm reads, or only the number the customer pays. A discount that does the second is a price cut wearing a badge.
The other thing a promotion does is set a reference. Amazon's fair pricing policy names misleading reference prices as a practice that harms customer trust, and a product that runs a discount most of the year has, in Amazon's eyes and the customer's, a normal price that is the discounted one. The strike through only means something if the price it strikes through was real, recently, and for most of the time.
For a product people buy again, Subscribe and Save sets a second price beside the first. Amazon's page says products are enrolled with a 0 percent seller-funded discount that the seller can raise to 5 or 10 percent, and that Amazon funds a 5 percent discount on Subscribe and Save orders of five items or more. Automatic enrolment needs FBA and a brand in Brand Registry, among other conditions, and Amazon says there are no selling costs beyond the usual fees.
The seller-funded discount comes off the floor. A product whose floor sits 8 percent below its everyday price cannot fund a 10 percent subscriber discount without selling every subscription at a loss.
What it buys is the repeat order without a second advertising click. Start at the level the floor allows rather than the highest: a subscriber who would have reordered at 5 percent is margin given away at 10.
The listing shows a price, and often a reference price beside it, and the customer reads the two together. A brand that controls its listing can anchor deliberately: a list price that is genuinely charged elsewhere, a pack size that makes the per unit price read as fair, and a variation ladder in which the size most customers want is priced to look like the sensible middle. None of that lowers the price. It changes what the price is compared with.
What it cannot do is invent the anchor. The reference price has to be a price the product actually sold at, and Amazon checks. The honest version of anchoring is a stable everyday price held for months, so that a real promotion, when it comes, is measured against something the customer believes.
Manage Your Experiments tests a listing's content rather than its price, so a price test on Amazon is a sequence: one price, then another, and a comparison that has to survive everything else that changed in between.
One change only. No coupon, advertising change or new competitor inside the test. At least two weeks at each price, the same weekdays in each, away from deal events.
Read profit per session, not units. Units almost always fall when the price rises, which proves nothing. Multiply each period's unit session percentage by the profit per unit. For the product above, at $27 with 12 percent and $5.45 a unit, each session earns about $0.65; at $30 with 10.5 percent and $8.00, about $0.84. The higher price wins while selling fewer units.
Then watch the sessions. Fewer units can mean less rank and, a few weeks later, fewer sessions. A price that wins per session and loses a fifth of its traffic has not won, so read the test again a month after it ends.
A pricing strategy fits on one line per product: the true floor, the position against the market, the everyday price, the promotion calendar and what each promotion is meant to move, and the review date. The repricer is set from that line and not the other way round. Rank is checked against price monthly, because a price that holds margin and loses the button on every session is not a strategy either; what moves rank at the listing level, and how price sits among the other inputs, is in
what moves rank on Amazon.
The same plan, with different fees and a different buy button, is what
a Walmart pricing strategy needs, and a brand selling on both should write the two lines side by side, because a Walmart price that undercuts the Amazon price is a Competitive External Price Amazon will find.
For the brands we run, pricing is a weekly job inside our
Amazon marketplace management: the floor rebuilt when a fee changes, the position reviewed against the Featured Offer share, promotions run for a reason and measured against rank rather than against the discount, and the Pricing Health page checked before anything else, because an offer that is not eligible to be seen has no price at all.