Returnless Refunds: The Break-Even Number, Not a Gut Call

A returnless refund is not generosity and it is not a customer-service gesture. It is arithmetic. You give up the item so you do not have to pay for the trip that brings it home — and the moment the trip costs less than the item, the arithmetic flips and you should be asking for it back.

Most sellers set this by feel, usually after one painful month of return labels. It is a two-line calculation, and the answer changes by SKU, not by policy.

What a returnless refund actually is

A returnless refund is a refund issued without requiring the buyer to ship the item back: the buyer keeps the goods, and the seller absorbs the cost of the goods instead of the cost of recovering them. Marketplaces market it under different names — Amazon calls the FBA version Returnless Resolutions, Walmart calls its Marketplace version Keep It Rules — but the mechanism is identical, and so is the trade.

The scale it sits inside is worth knowing before you tune anything. The National Retail Federation put total US retail returns at $849.9 billion in 2025, equal to 15.8% of annual sales, with 19.3% of online sales returned. Nine per cent of all returns were assessed as fraudulent. Returns are not an edge case in your P&L; they are a line item the size of a mid-tier expense category.

The cost lines a keep-it refund removes, and the three it does not

Cost line	Standard return	Keep-it refund
Refund to the buyer	Paid in full	Paid in full
Return shipping label	Paid	Removed
Inbound receiving and inspection labour	Paid	Removed
Repack, relabel, restock	Paid	Removed
Disposal or liquidation of an unsellable unit	Often paid	Removed
Landed cost of the goods	Recovered if the unit is resellable	Written off in full
The return counted against your return rate	Counted	Still counted
Fraud and abuse exposure	Moderate	Higher
The reason the return happened	Unchanged	Unchanged

Those last three are the whole reason this is a calculation and not a policy switch. The keep-it route removes logistics cost. It does not remove the return, the metric, or the cause.

The break-even line

Take the item back only when what you can realistically resell it for exceeds what it costs to get it back and make it sellable again. Written out:

Take it back if: (probability the unit is resellable × realistic resale price) > return freight + receiving and inspection + repack and relabel + disposal cost of the units that fail inspection.

Otherwise, refund returnlessly and write off the landed cost.

Fill in six inputs and the decision falls out on its own:

Input	Where to get it	Notes
A — Landed cost per unit	Factory price + inbound freight + duty + first-mile	Not your selling price. This is what you actually lose.
B — Realistic resale recovery	Historic sell-through of returned units at their actual discount	Open-box, B-stock or liquidation price, not full retail
C — Return freight	Carrier rate for the return lane, or your marketplace's return shipping charge	For oversized goods this is the number that decides everything
D — Receiving and inspection	Warehouse cost per touch, per unit	3PL invoices usually itemise this
E — Repack and relabel	Cartons, inserts, poly, barcode labels	Cheap per unit, easy to forget
F — Failure rate on inspection	Share of returned units that cannot go back to A-stock	Category-dependent and usually underestimated

The comparison you are making is A (what you lose by giving up) against C + D + E plus the write-off you take anyway on the F share that arrives unsellable. Where A is small and C is large, refund returnlessly and stop thinking about it. Where A is large and C is small, take it back every time. The interesting decisions live in the middle band, and they move whenever freight rates move.

If you have never put real numbers against those six inputs, the fastest way in is the return cost calculator — it forces you to name the cost lines instead of eyeballing them.

Where the platforms already decided for you

Two of the biggest marketplaces have already turned this into a settings page, with eligibility rules that are worth reading closely because they encode the same math.

	Amazon — FBA Returnless Resolutions	Walmart Marketplace — Keep It Rules
Announced	13 August 2024	Available in Seller Center
What you configure	A maximum item price for returnless eligibility, set anywhere between $1 and $75 (added 20 November 2025)	Department, Item price limit, Return Order limit
Partial-refund variant	Not offered under this program	Partial Keep It Rules: Department, Return Reason, % Refund Off, Max Amount
Price ceiling	Products with an average sales price above $75 are ineligible	Set by you, per department
Excluded goods	Dangerous goods, heavy and bulky items, high-priced goods	Walmart retains discretion over disposal and over individual transactions
Buyer eligibility	Customers with a history of returns abuse are excluded; seller and customer must be in good standing	Walmart may rule a specific transaction ineligible
Program fee	None	None stated

Walmart's own documentation states the logic plainly: "Sometimes it's more cost-effective to let Walmart.com customers keep an item and process a refund than it is to pay a return shipping fee."

Amazon added a returns and inventory recovery dashboard for all sellers on 12 November 2025, showing top return reasons and recovered value per unit. That dashboard is the cheapest source for inputs B and F above — use it before you invent them.

The trap: the goods that bleed most are the ones excluded

Read the eligibility rules again. Amazon's program caps out at a $75 average sales price and explicitly excludes heavy and bulky items. Walmart's rules are configured per department with a price limit you set.

Returnless refunds are cheapest exactly where the return was cheap anyway. The freight-heavy, high-value goods that make return economics genuinely ugly — the sofa that ships LTL, the vanity unit, the pallet of tiles, the 65-inch TV stand — are the ones sitting outside the program. For those, there is no keep-it button. There is only the full round trip, and it is the round trip covered in oversized item return shipping cost.

That inverts the intuition most sellers arrive with. If your catalogue is bulky furniture or building materials, tuning keep-it settings is close to a rounding error. Your lever is upstream, in the share of returns that happen at all.

What a returnless refund does not fix

The return still counts. Since 1 June 2024 Amazon has applied a returns processing fee to FBA products whose return rate exceeds their category threshold, apparel and shoes aside. The published thresholds vary widely by category — Amazon's own list runs from about 2.9% for Grocery and Gourmet up to 12.8% for Backpacks, Handbags and Luggage. The rate is measured across the month a unit shipped plus the two following months, and products shipping fewer than 25 units in a month are exempt. Letting the buyer keep the item does not take the return out of that numerator. Which numerator you are actually being judged on is its own trap, unpicked in how to calculate return rate.

Abuse gets easier. With 9% of returns already assessed as fraudulent, a publicly advertised keep-it policy is an invitation. Both platforms build guardrails in for a reason: Amazon excludes buyers with a history of returns abuse; Walmart caps by item price and by total order value and keeps discretion over individual transactions. If you build your own rules on a direct-to-consumer store, cap by price, cap by order value, cap by customer, and do not put the policy in your marketing copy.

The cause is untouched. Letting the buyer keep it is a cheaper way to lose. It changes nothing about why the buyer wanted out. Size and specification mismatch is consistently one of the largest single causes in this data set — the magnitudes are collected in the e-commerce returns size statistics — and no refund setting has ever reduced it by a point.

Three product profiles, three different answers

Profile	What it looks like	What the math usually says
Low value, light, thin margin	Accessories, consumables, small parts under roughly $25 landed	Refund returnlessly by default. Return freight plus one warehouse touch routinely exceeds the landed cost. Set the rule once and stop reviewing case by case.
Mid value, mid weight, genuinely resellable	Small appliances, lighting, hardware, packaged goods that survive inspection	The real decision band. Depends entirely on input F, your inspection failure rate. Measure it for one quarter before you set a policy.
Bulky, freight class, high landed cost	Furniture, sanitaryware, building materials, large electronics	Take it back — and accept that the return is expensive either way. Marketplace keep-it programs exclude these goods. Spend the effort on prevention, not on refund settings.

Next steps

Pick whichever of these you have not done, in this order:

Get one real number per SKU family. Landed cost, return freight, inspection failure rate. Without those, every keep-it rule you write is a guess. The return cost calculator linked above is a fast way to get the shape of it.
Set the platform rules deliberately. On Amazon, choose your price cap between $1 and $75 rather than accepting a default. On Walmart, write Keep It Rules per department, and consider Partial Keep It Rules before full ones.
Split your return reasons into two buckets — reasons you caused and reasons you did not. Only the first bucket is fixable, and it is usually the larger one.
Attack the top self-caused reason. If it is size or specification mismatch, the fix is not a better product description; it is putting the measured dimensions on the image the buyer actually looks at, at the size the marketplace displays it. Options range from paying a designer per SKU, to a generic photo editor with hand-drawn arrows, to software that snaps to the product's real edges and exports the annotated image at each marketplace's spec-diagram size. What separates them is whether the number on the arrow is measured or merely plausible — an AI image generator will happily produce a clean-looking dimension line carrying a figure nobody ever measured, and that is a worse outcome than no dimension line at all.
Re-run the break-even every time freight rates move. The line between "take it back" and "let them keep it" is a freight rate in disguise.

FAQ

When should I give a returnless refund?

When the landed cost of the unit is lower than the total cost of recovering and re-selling it: return freight, receiving, inspection, repack, plus the write-off on units that fail inspection anyway. In practice that is most low-value, lightweight goods, and almost never bulky goods with a high landed cost.

Do returnless refunds still count against my return rate?

Yes. The refund is still a return in the marketplace's data. On Amazon this matters directly, because the returns processing fee introduced on 1 June 2024 is triggered by a product's return rate against its category threshold, not by whether the unit physically came back.

Can I offer a partial refund instead of letting the buyer keep it for free?

On Walmart Marketplace, yes — Partial Keep It Rules let you set a department, a return reason, a percentage off and a maximum amount, so the buyer keeps the item for a discount rather than a full refund. It is the cheapest option in the set when the item is usable but not as expected, and it is underused.

Are bulky items eligible for returnless refunds on Amazon?

No. Amazon's FBA Returnless Resolutions program excludes dangerous goods, heavy and bulky items, and products with an average sales price above $75. If your catalogue is furniture or building materials, this program is not the lever you are looking for.

Will customers abuse a keep-it policy?

Some will. The National Retail Federation assessed 9% of all returns in 2025 as fraudulent. The mitigation is structural rather than moral: cap by item price, cap by order value, exclude repeat abusers, and never advertise the policy. Both major marketplace implementations do all four.