PricingMag on the Minimum Price Most Amazon Sellers Get Wrong

Your minimum price is the most consequential number in your Amazon business and almost certainly the least carefully calculated. It is typically produced once, in a hurry, during the setup of something else, using whichever cost figures were within reach — and then it governs every automated pricing decision you make for years.

Here is the reframe that makes people take it seriously: a minimum price is not a safety limit. It is a standing instruction that says *you may sell this product at this price without asking me.* Automation reads it exactly that literally. If the number is wrong, you have authorised a loss and scheduled it to happen repeatedly.

The Calculation Most Sellers Actually Do

The common version looks like this:

Selling price − cost of goods − referral fee = margin.

Set the floor where margin hits an acceptable level.

Three costs are missing, and together they are usually larger than the margin being protected.

Fulfilment cost. For FBA, the per-unit fulfilment fee — which varies by size tier, weight and, increasingly, price band. For self-fulfilled, your actual pick, pack, packaging and postage cost, which is almost always underestimated because packaging and labour get treated as overhead.

Storage and holding cost. Monthly storage, allocated per unit by expected sell-through. Long-term storage exposure on slow movers. The cost of capital tied up in the unit.

Returns cost. Not a catalogue-wide average. Per SKU, because the spread is enormous — and the full cost of a return is the lost sale plus the returns processing plus the unit's condition loss plus the inbound and outbound shipping already spent.

Add advertising where a SKU only sells with ad support, and the floor moves again.

The Calculation That Actually Protects You

Build contribution per unit, then set the floor against it:

Selling Price

− referral fee (category percentage, applied to the total selling price)

− fulfilment fee, including any active surcharge

− allocated monthly storage

− allocated returns cost

− allocated advertising, where the SKU depends on it

− landed cost of goods, including freight and duty

− contribution per unit

The floor is the price at which contribution reaches the minimum you are genuinely willing to accept per unit — which is a commercial decision, not a formula. Some sellers set it at zero contribution for liquidation SKUs. Most should set it meaningfully above zero, because a unit sold at zero contribution has consumed capital, storage, handling and attention for nothing.

Four Inputs That Go Stale and When

A floor is not a permanent object. It is a calculation whose inputs expire, and four of them expire on schedules you can anticipate.

Amazon's fees. The clearest example is recent: the 2026 US referral and FBA fee update was announced in October 2025 and took effect for most sellers on 15 January 2026, with FBA fees rising by an average of about eight cents per unit — framed by Amazon as less than half a percent of a typical selling price — while referral percentages held flat. Then a 3.5% fuel and logistics surcharge was applied to FBA fulfilment fees in the US and Canada from 17 April 2026, a change that does not appear as a line in the published fee schedule. Every floor calculated before those dates understates cost. Details sit in Amazon's selling-partner announcement and its fee change summary.

Your landed cost. Supplier price changes, currency movement, freight rates, duty and tariff changes. Any of these invalidates the floor without touching the listing.

Your returns rate. It moves with seasonality, with listing changes, and with product revisions. A floor built on last year's return rate is a guess about this year's.

Your size tier or price band. Packaging revisions and re-measurement can move a SKU into a more expensive bracket. Price bands mean the fee can change because the *price* changed, which creates a circularity worth understanding: cutting price to compete can push you into a worse fee band, so the floor itself has to be solved against the band it lands in.

Recalculate on events, not on a calendar. A quarterly reminder will miss a mid-quarter fee change by two months.

The Two Mistakes That Cost the Most

Mistake one: one floor logic for the whole catalogue. A single percentage-margin rule applied across mixed products will be right for some and structurally wrong for the rest. Heavy items, low-ASP items and high-return items each need their own treatment, and the SKUs where the generic rule is wrong are exactly the ones that end up parked at their minimum.

Mistake two: treating the floor as hypothetical. Sellers set floors they never expect to be reached and are surprised when automation reaches them routinely. If you would be unhappy selling at your minimum tomorrow, your minimum is wrong today. The test is simple: look at your floor and ask whether you would accept a hundred units at that price. If the answer is no, change the number.

Set the Maximum Too

Minimums get all the attention; maximums do real work and are frequently left empty.

Amazon uses your minimum and maximum price settings as the primary signal for what constitutes a plausible price on your offers. Left blank, Amazon falls back on its own internal data to decide whether your price looks like an error — so an unset maximum is not neutrality, it is delegation. Populated boundaries mean fewer pricing-error deactivations and fewer surprises.

A maximum also caps your exposure under the pricing policies: an unbounded upward authorisation on a marketplace that enforces against unreasonable total prices is an avoidable risk during a demand spike.

The Thirty-Minute Version

If you do nothing else: take your top twenty SKUs by volume. Build contribution per unit properly, with current fee inputs. Compare the result against the floor currently sitting in your pricing tool.

Sellers who run this exercise for the first time generally find two things — several floors below true break-even, and at least one SKU that has been selling at a loss for months with healthy Featured Offer share. That second discovery is the reason this is worth your morning.

If you’re working out what your real minimum price should be, the research team at PricingMag has published the full calculation, worked through cost line by cost line, along with an event-based schedule for keeping it up to date.

Your floor is a promise you made to your automation. It is worth knowing what you promised.