Most repricing discussions are about the ceiling: how to win the Buy Box, how to compete without racing to the bottom, which strategy beats which. The floor gets set once during setup, usually as a round number that felt safe, and then never revisited.
That floor is doing more to determine your annual profit than any strategy setting above it. It is the number that defines the worst outcome the repricer is permitted to produce, and a repricer will find that outcome — repeatedly, on your highest-volume SKUs, during exactly the competitive periods when volume is highest.
A minimum price set by instinct is not a safety net. It is a standing instruction about how much money you are willing to lose per unit.
What a minimum price is supposed to be
A minimum price is the lowest price your repricer may set. Its purpose is to stop automated competition from driving a SKU below the point where selling it is worse than not selling it.
Which means the correct floor is derived from one number: true net cost per unit. Not landed cost. Not landed cost plus fees. Every cost that a sale actually incurs.
True net cost per unit = Landed cost
+ Referral fee
+ Fulfillment fee
+ Storage allocation
+ Returns cost (return rate × cost per return)
+ Allocated advertising cost per unit
+ Transaction-level costs (promotions, coupons if running)
Note that referral and fulfillment fees are price-dependent, which makes this slightly circular — the referral fee at your floor is lower than the referral fee at your list price. Solve it at the floor, not at list.
Then:
Break-even price = True net cost per unit
Minimum price = Break-even price ÷ (1 − target minimum margin %)
Why the two costs sellers omit are the two that matter
Almost every seller includes landed cost, referral and fulfillment. Almost none include the other two.
Returns. A return costs you more than the refunded revenue. There is a return processing fee, a refund administration charge, the outbound fulfillment fee you don’t get back, and a unit that comes back unsellable a meaningful share of the time. On a SKU with a 12% return rate, the per-unit returns cost spread across all units sold is not a rounding error.
Allocated advertising. If you spend to acquire the sale, that spend is part of the cost of the sale. A SKU at 18% TACOS carries an ad cost per unit equal to 18% of its selling price — which at your floor price is a smaller absolute number but a larger proportional bite.
Worked example: what an instinct floor costs
Illustrative SKU. Price $39.99, landed cost $12.40, 11% return rate, 16% TACOS.
Cost stack at a $24.99 candidate floor:
| Line | Amount |
|---|---|
| Landed cost | $12.40 |
| Referral fee (15% of $24.99) | $3.75 |
| Fulfillment fee | $5.85 |
| Storage allocation | $0.42 |
| Returns cost (11% × $8.90 per return) | $0.98 |
| Allocated ad cost (16% of $24.99) | $4.00 |
| True net cost | $27.40 |
The floor the seller chose — $24.99, because it looked like a comfortable buffer over a $12.40 landed cost — is $2.41 below break-even.
Now the scale of it. Say the repricer sits at or near the floor for 30% of the month during competitive pressure, on a SKU selling 900 units/month:
Units sold at or near floor: 900 × 0.30 = 270
Loss per unit: $2.41
Monthly loss: $650
Annual loss on one SKU: $7,800
And the sales look fine. Volume is healthy, the Buy Box percentage is good, revenue is up. The only place the problem is visible is net profit per unit — which is why sellers running on revenue dashboards can do this for a year without noticing.
The correct floor, at a 12% minimum margin target:
Break-even: $27.40
Minimum price: $27.40 ÷ (1 − 0.12) = $31.14
The instinct floor was $6.15 too low.
The ad-spend interaction nobody models
Here is the compounding effect. Advertising cost is usually thought of as a percentage of selling price. As the repricer walks the price down toward the floor, the absolute ad spend per click does not fall — your bids don’t drop because your price dropped.
So at $39.99 with a $6.40 ad cost per unit, TACOS is 16%. At $27.99, that same $6.40 is 22.9%. The SKU becomes less profitable per unit and proportionally more expensive to advertise, simultaneously, at exactly the moment the repricer is trying to win volume.
The implication: your floor should be calculated using the absolute ad cost per unit, not a TACOS percentage. A percentage-based calculation understates the ad burden at the floor and produces a floor that is too low, every time.
Best practices
- Calculate the floor per SKU from true net cost. Never account-wide, never a percentage off list price.
- Use absolute ad cost per unit, not a TACOS percentage, for the reason above.
- Include the returns cost. For a high-return category this can be several dollars per unit sold.
- Solve fees at the floor price, not the list price. Referral scales with price.
- Recalculate on every new inbound shipment. New landed cost means a new floor. This is the maintenance step that gets skipped.
- Recalculate after any fee schedule change or dimension remeasurement. A size-tier change can move the fulfillment fee enough to invalidate the floor.
- Set the floor above break-even, not at it. A floor at break-even means the best case for that unit is zero.
- Review how much time each SKU spends at its floor. A SKU pinned at the floor for most of the month is telling you something about the competitive position, not about the repricer.
sellerboard’s repricer is available on all marketplaces and included in all plans, and it works from the same net profit data used to calculate your P&L — which is the point of integrating the two. A floor derived from a live cost stack updates when your costs update, rather than staying at whatever it was set to when the SKU launched.
When selling below break-even is a deliberate choice
There is a legitimate version of this, and it is worth naming so the article isn’t read as “never go below cost”:
- Launch periods, where you are buying rank and have modelled the payback horizon.
- Clearing aged inventory, where the alternative is storage surcharges and eventual disposal, so a below-cost sale is the cheaper of two bad options.
- Competitive defence for a defined window, with a date at which it ends.
The distinction is not whether you go below break-even. It is whether you know you are, for how long, and at what total cost. A floor set below break-even by accident has none of those properties.
Common mistakes
- Setting the floor as a percentage off list price. Bears no relationship to cost, and moves in the wrong direction when you discount.
- Using landed cost as the floor. Ignores every Amazon fee, which is most of the cost stack.
- Setting one floor for a whole catalogue. Guarantees that some SKUs are priced below cost and others are leaving margin unclaimed.
- Never updating after a cost increase. Freight and tariff movements silently push SKUs below their unchanged floor.
- Omitting the ad cost entirely on the grounds that advertising is a marketing expense rather than a unit cost. If the sale required the click, it is a unit cost.
- Trusting the fee estimate rather than the fee charged. Estimated fees and actual charged fees diverge, particularly after remeasurement.
- Monitoring revenue instead of net profit per unit. This is what makes the whole problem invisible.
FAQ
Should my minimum price include advertising cost? If the SKU requires advertising to sell at its current volume, yes. Excluding it produces a floor that is profitable only for organic sales, while most of your units are paid.
How often should I recalculate the floor? On every new inbound shipment at a different landed cost, and after any Amazon fee change or dimension remeasurement. Quarterly at minimum.
What minimum margin should I target at the floor? Enough to absorb estimation error in your returns rate and ad allocation. The floor is a worst case, so building in nothing means the worst case is a loss whenever your estimates are slightly optimistic.
Is it ever right to set the floor below break-even? Yes, for launches, aged-inventory clearance and time-boxed competitive defence — provided it is deliberate and has an end date.
Why does my SKU sell so many units at the floor? Usually because the competitive set is priced near your floor. That is a positioning signal worth acting on, and lowering the floor is the one response that makes it worse.
Conclusion
Repricing strategy determines how much you earn on a good day. The minimum price determines how much you lose on a bad one, and bad days cluster — they arrive during competitive pressure, when volume is high.
The calculation is not complicated. It is landed cost plus every fee, plus returns, plus the absolute ad cost per unit, solved at the floor price rather than the list price, with a margin buffer on top. Doing it once per SKU and refreshing it when costs change is a couple of hours of work that reliably finds money on any catalogue where the floors were set by feel.
