Seven Days Or They Keep It.

Happy Friday,

A buyer in Ohio decides on December 19 that your thing isn't for them.

You have until Christmas Day to decide how you'd like to lose.

From December 18, most US returns on seller-fulfilled orders need a prepaid return label, and the clock is Amazon's: "within seven calendar days of a return request."

Not seven days from when the item ships back. Seven from the moment the buyer clicks.

You get two options, and Amazon lists them in this order. One: "Upload a prepaid return shipping label against the return request on the Manage seller-fulfilled returns workflow." Two: "Issue a full refund without requiring customer to return the item."

Read option two again. Option two is giving it away.

And if you pick neither — if the week just goes by, which is what weeks do in December — Amazon's wording is that "a returnless refund may be automatically issued and the balance deducted from your selling account."

The buyer keeps the item. You keep the feedback.

One option costs you the freight. The other costs you the product.

Now count who gets paid on a cape worn once.

Amazon takes the referral fee on the sale. Most of that comes back on the refund, less the administration fee, which doesn't. Amazon takes the outbound fulfillment fee, which stays taken whatever happens next. Amazon takes the returns processing fee for handling the thing coming home. Three charges, three revenue lines, and the only one that reverses has a deduction stapled to it.

You get the cape back. In November.

Paid on the way out, paid on the refund, paid on the way back.

Now the honest part.

Processing a return is real work. Somebody opens that box, decides whether a thing that spent Saturday night at a party can be sold to a stranger, then repacks it or writes it off. That labor exists whether or not a fee sits on top of it. Apparel genuinely returns at a multiple of other categories, which is why it was in the original group when this fee started two years ago. And Amazon publishes its exemptions: ASINs shipping under twenty-five units a month aren't charged, the first twenty returns on a qualifying new ASIN are waived, and returns Amazon caused don't count.

Read that as fairness. The arithmetic doesn't move: on a costume worn once, the only party with guaranteed revenue is the one that never bought any inventory.

Today — the Business Hour Delivery Rate line bites. Sellers below 90% have their seller-fulfilled offers deactivated for Amazon Business buyers from today. FBA offers and standard retail offers are unaffected. If you sell B2B on merchant-fulfilled offers, look at the metric before you leave this afternoon.

Tomorrow — the FBA New Selection introductory window closes. Branded ASINs launched from July 30 through October 31 collect the 2026 program's unit-based credits, and the outer grandfathering window ends the same day. A launch sitting in a draft folder is worth more today than it is Monday.

Monday — the insurance rule. Enhanced-safety categories need $1 million of commercial liability coverage from November 2, regardless of sales volume. The old $10,000-a-month trigger stops being the trigger. Read your certificate's expiry date, not your renewal email.

Twenty minutes. Give it to whoever handles FBM returns.

Don't do this across the whole catalog — the rule only ever fires on things that actually come back, so ten SKUs is the whole job.

Pull ninety days of returns. Orders, Manage Returns, set the date range. Count by SKU and take the top ten. Everything outside the ten keeps doing what it does today until it turns up on the list.

Price one return label at three weights. One pound, three, ten. Off your rate card, or one quote if you don't keep one. Three numbers cover all ten SKUs.

Write what you'd actually recover against each of the ten. Full resale if that SKU comes back sellable, scrap or liquidation value if it doesn't. Ten judgment calls, which is the point of cutting it to ten.

Mark each one Label or Refund. Label where recovery beats the freight. Refund where it doesn't — those are the ones where option two was always the right answer and the seven-day clock just made it explicit.

Write the ten lines where returns get handled, so nobody is doing this arithmetic inside a seven-day window in December.

Artifact: your ten most-returned FBM SKUs, each marked Label or Refund, with the label cost and recovery value that decided it, and Label as the standing default for everything else.

Sort by return count, not return rate — the question is how often the rule fires, and a high-volume SKU that comes back weekly matters more than a niche one returning 8% of three units. Use what you'd actually recover rather than landed cost; a unit that comes back unsellable is worth scrap, and valuing it at cost is how you pay freight twice. If a SKU is on the list because it's defective or mis-described, marking it Refund hides the problem instead of fixing it — fix the listing and leave that one on Label until you have. And a Refund mark is still a decision you have to execute inside the seven days: the automatic version takes it out of your hands and leaves no record you chose it.

Seven days, starting the moment someone changes their mind.

Decide now which of your products you'd rather just let them keep. In December you won't have the week to work it out.

See you Tuesday.

Any questions? Book a meeting with me here.

Dan Head

Founder, AMZ Elite