US marketplace. Fee figures reflect Amazon’s published 2026 rate cards, last verified August 2026.
Selling below cost to build rank is defensible only when you can state, in advance, the total loss you will absorb, the rank position it should buy, and the number of months at full price required to earn it back. If a $4.58 loss per unit across 800 launch units costs $3,664 and your steady-state contribution margin is $9.67, you need 379 additional units of organic sales — beyond what you would have sold anyway — just to break even on the tactic. Most loss-leader launches fail not because the strategy is wrong but because nobody calculates that number, and the payback is measured in units the seller was going to sell regardless.
The tactic has a real mechanism behind it. Sales velocity influences organic ranking, ranking drives sessions, and sessions at a decent conversion rate produce sales that sustain the ranking. Buying velocity below cost is buying a position in that loop. The question is never whether the mechanism exists — it does — but whether the price you pay for entry is recoverable.
What does a loss-leader launch actually cost per unit?
The loss is larger than the gap between price and COGS, because Amazon’s fixed per-unit fees do not shrink when you cut the price, and your advertising cost per unit is at its highest precisely during launch.
Large standard-size product at 1.25–1.5 lb, $6.90 landed COGS, 15% referral fee. Both the $29.99 target price and the $18.99 launch price sit in the $10–$50 fulfillment band, so the fee is $5.42 base or $5.61 with the 3.5% fuel and logistics surcharge — identical at either price.
| Line | Target price $29.99 | Launch price $18.99 |
|---|---|---|
| Selling price | $29.99 | $18.99 |
| Referral fee (15%) | −$4.50 | −$2.85 |
| Fulfillment fee (incl. surcharge) | −$5.61 | −$5.61 |
| Landed COGS | −$6.90 | −$6.90 |
| Storage and returns allowance | −$0.61 | −$0.61 |
| Advertising | −$2.70 (9% TACoS) | −$7.60 (40% TACoS) |
| Contribution per unit | +$9.67 | −$4.58 |
Note where the damage concentrates. The price cut costs $11.00 of revenue but only returns $1.65 in referral fee savings — the fulfillment fee, COGS, and storage are unchanged in dollar terms. Meanwhile launch-phase advertising nearly triples per unit because you are bidding for placements without ranking or review history. Advertising, not the discount, is usually the larger half of a launch loss.
One 2026 detail worth checking before setting a launch price: fulfillment fees are banded at under $10, $10–$50, and over $50. If your launch price crosses below $10, the product picks up the Low Price FBA discount — an average $0.86 per unit — which changes the loss calculation. If your target price crosses above $50, the fee rises. Run the fee at both prices rather than assuming it is constant.
How do you calculate the payback?
Three numbers, in this order:
1. Total launch investment. Loss per unit × launch units. At −$4.58 across 800 units: $3,664.
2. Incremental units required. Total investment ÷ steady-state contribution margin. $3,664 ÷ $9.67 = 379 units.
3. Payback period. Incremental units ÷ incremental monthly sales. The word “incremental” is the whole discipline here. If the SKU would have sold 150 units a month organically without the launch push and now sells 260, the incremental figure is 110 — not 260. At 110 incremental units a month, payback is 3.4 months.
If you cannot estimate the counterfactual, you cannot evaluate the tactic. A launch that produces 260 units a month against a SKU that would have found 240 on its own is a $3,664 payment for 20 units a month, and payback runs past 18 months — well beyond the horizon over which any ranking position is stable.
When does the tactic actually work?
| Condition | Why it matters |
|---|---|
| Repeat-purchase product | Consumables and replenishables let you recover on the second and third order, not just the first. This is the single strongest predictor of loss-leader success. |
| High steady-state contribution margin | A $9.67 contribution recovers a $3,664 investment in 379 units. A $2.15 contribution needs 1,704 units for the same investment. |
| Defensible position once ranked | Patent, brand registry, exclusive supply, or a genuinely differentiated product. If three competitors can copy you next quarter, you bought a temporary position at a permanent cost. |
| Adequate inventory depth | Going out of stock mid-launch surrenders the ranking you just paid for, and the low-inventory-level fee applies below 28 days of historical supply — $0.32 to $2.09 per unit depending on size tier and how far below you fall. |
| A defined end date | The price must return to target on a schedule, with a plan for the volume drop that follows. |
Conversely, the tactic is close to indefensible on one-time-purchase commodity products in unprotected categories with thin steady-state margins. That describes a large share of the launches where it is attempted.
What happens when you raise the price back?
This is the step that undoes most launches, and it is predictable.
Moving from $18.99 to $29.99 is a 58% increase. Conversion rate will fall — the magnitude depends on the category and the competitive set, but a meaningful drop is certain. Lower conversion reduces sales velocity, which affects organic rank, which reduces sessions, which pushes you back toward paid traffic at exactly the moment you were counting on organic sales to repay the investment.
Practical mitigations:
- Step the price up. $18.99 to $22.99 to $26.99 to $29.99 over six to eight weeks, watching conversion and rank at each step. Slower, but it lets you find the point where conversion breaks rather than discovering it all at once.
- Use coupons rather than list price cuts. A coupon preserves the reference price and can be withdrawn without a visible price increase on the listing. It costs a redemption fee, but that is usually cheaper than rebuilding price perception.
- Do not raise price while reviews are still thin. The full price has to be justified by social proof. Raising before the review count supports it converts a launch loss into a stalled listing.
- Budget for a post-launch ad increase. TACoS will not drop to steady state the moment the price goes up. Assume an intermediate period.
What are the alternatives to selling below cost?
Loss-leading is one of several ways to buy velocity, and rarely the most efficient.
| Approach | Cost profile | Trade-off |
|---|---|---|
| Below-cost list price | Loss on every unit, including units that would have sold at full price | Blunt; no targeting; damages price reference |
| Coupons | Discount plus redemption fee, only on redeemed units | Preserves list price; visible badge lifts click-through |
| Higher ad spend at full price | Loss concentrated in advertising, targetable by keyword | Buys the same velocity while protecting margin on organic sales |
| Lightning Deals and deal placements | Fee plus discount, time-boxed | Volume spike in a controlled window rather than an open-ended loss |
| Vine and review programmes | Fixed cost plus free units | Buys social proof rather than velocity — often the actual bottleneck |
Aggressive advertising at full price frequently beats a below-cost list price for the same total spend, because the loss lands only on the units that needed help. A below-cost price discounts every unit, including the organic sales you were already winning.
Common mistakes
- No stated budget or exit. “We’ll run it low until it ranks” is not a plan. Set the total loss you will absorb and the date the price returns to target before the first unit ships.
- Counting all post-launch sales as incremental. The payback calculation is only valid against the counterfactual. Without a baseline estimate, every launch looks successful.
- Forgetting the fixed-dollar fees. Cutting price by 37% does not cut costs by 37%. Fulfillment fees, COGS, and storage are unchanged within a band, so contribution margin falls far faster than price.
- Underestimating launch TACoS. Launch advertising commonly runs three to five times steady-state. Budget for it explicitly rather than discovering it in the settlement report.
- Running out of stock. The fastest way to convert a launch investment into a pure loss, and it also triggers the low-inventory-level fee. Model the higher sell-through rate into the reorder before the promotion starts.
- Confusing this with a genuine retail loss leader. In physical retail, a loss leader draws a customer who fills a basket with profitable items. On Amazon there is no basket — the customer buys your discounted unit and leaves. The mechanism you are buying is ranking, not attachment, and it is a weaker and more perishable asset.
Evaluating this honestly means seeing per-SKU contribution margin before, during, and after the promotional window, with advertising attributed at the SKU level. Analytics platforms including sellerboard break profit down per SKU and per period, which is what makes the incremental-versus-baseline comparison possible after the fact.
Frequently asked questions
Does Amazon’s algorithm actually reward sales velocity?
Sales velocity and conversion rate are widely understood to influence organic ranking, and Amazon has never published the weighting. Treat velocity as one input among several — relevance, conversion, reviews, availability — rather than a lever you can pull in isolation.
How long should a loss-leader launch run?
Long enough to establish ranking and accumulate reviews, typically four to eight weeks, with a defined end date set in advance. Open-ended below-cost pricing stops being a launch tactic and becomes a structural loss.
Is it better to discount the price or increase ad spend?
For the same total investment, ad spend is usually more efficient, because the loss is targeted at the units that need help rather than applied to every unit including organic sales. Discounting has the advantage of improving conversion rate directly, which advertising does not.
Will a low launch price hurt me when I raise it?
Yes, to some degree. Buyers anchor on the price they first saw, and conversion will drop when you raise it. Stepped increases and coupon-based discounting both reduce the damage compared with a single large jump.
What if I never reach payback?
Then the launch was a marketing expense, not an investment, and it should be recorded as one. The useful discipline is deciding in advance how much you are willing to spend on that outcome — and being honest afterwards about whether the incremental volume ever materialised.
