The 91-Day Return: Amazon Makes the Holiday Return Window Permanent

On October 16, Amazon posted a four-paragraph note to its sellers that will shape every fourth quarter from here on. Items bought on Amazon.com between November 1 and December 31, 2026 can be returned until January 31, 2027, and Apple-brand products until January 15. One line in the note sets this year apart from every year before it: "This policy will apply every year." The extended holiday return window has opened on November 1 for four years running, but sellers always had to wait each fall to learn whether it would return. Now it is a fixed part of the calendar.

The arithmetic is what matters to a brand. An order placed on November 1 stays returnable for 91 days, a Black Friday order on November 27 for 65 days and a Cyber Monday order for 62. Walmart opened its own window a month earlier, on October 1, with the same January 31 deadline, and Apple's extended policy took effect on November 6. Shoppers do not keep a separate rulebook for each store, so these windows become the standard that every other brand is measured against.

The practical result is that margin booked in November is not settled until February. This post covers what Amazon changed and why it matters beyond Amazon, how long holiday revenue now stays provisional and why a longer window need not mean more returns. It then looks at where the cost of a late return lands, and at six decisions worth making before Black Friday on November 27.

1. What Amazon Changed, and Why It Reaches Beyond Amazon

The dates themselves are not new. Amazon announced the 2025 version on October 14, 2025 with identical terms, and the window has started on November 1 since 2023. That date was itself a concession: in 2022, Amazon announced the policy on September 30 and started it on October 7, which let some purchases come back almost four months later, and sellers pushed back. What changed this year is the commitment. The note covers every item sold on Amazon.com, whether Amazon or the seller ships it, and it offers sellers no way to opt out.

Amazon was also candid about why it does this, and the reason is commercial. The note presents the longer window as a way to strengthen customer traffic during the year-end season, which is a demand argument rather than a service one. A long return promise lowers the perceived risk of buying a gift early. That pulls purchases into November, when Amazon wants the traffic and before advertising costs reach their Black Friday peak. Read that way, the return window is a sales tool, and Amazon has judged that the extra sales are worth the extra returns.

The decision matters well beyond Amazon's own sellers, because shoppers carry the largest retailer's rules into every other store. Walmart's window opened on October 1 and Apple's on November 6. In the NRF's most recent returns survey, 37% of large retailers also planned to extend their return windows to manage holiday returns. A direct-to-consumer brand with a flat 30-day policy is now the exception in December, and it should expect shoppers to notice.

2. The 91-Day Math: Holiday Revenue Stays Provisional Until February

Because Amazon counts from the purchase date, the extension is worth the most to the earliest orders. An item bought on November 1 can go back 91 days later, while an item bought on December 31 gets 31 days, close to the usual allowance. That shape suits Amazon, since the early weeks are the part of the season it most wants to fill. It also means that the orders a brand is happiest about in mid-November are the ones that stay open longest.

The volume at stake is large. Retailers told the NRF they expect 17% of holiday sales to be returned, a share that has held steady for several years. Applied to the NRF forecast of $1.013 trillion for November and December, that would mean roughly $172 billion of merchandise flowing back. Most of it arrives late: the NRF notes that the vast majority of holiday returns occur in January, and 82% of shoppers return unwanted gifts within a month of receiving them.

The timing creates a reporting problem more than an accounting one. Accounting rules already ask a company to book revenue net of the returns it expects, so the longer window leaves the rule alone while making the estimate larger and later to prove. Sellers who reacted to the news named the practical risk: brands judge the fourth quarter and place first-quarter reorders off December's gross numbers, and the January return wave lands after those decisions. Large retailers have long closed their fiscal year at the end of January, after the returns are in. A brand on a calendar year closes its books on December 31, with a month of holiday returns still to come.

3. Why a Longer Holiday Return Window Need Not Mean More Returns

The instinct in most finance teams is that more time means more returns, and for this lever the research says the instinct is wrong. A meta-analysis in the Journal of Retailing, which we first covered when DoorDash launched 30-minute returns, pooled 21 studies and split return policy leniency into five parts: time, money, effort, scope and exchange. Lenient policies lifted purchases overall and lifted returns by a smaller amount. Leniency in time, however, reduced return rates. The authors point to the endowment effect: the longer someone owns a product, the more attached they become, and the less likely they are to send it back.

That finding changes the question a brand should ask about the holiday window. The extra weeks mostly decide when a return arrives, while whether it arrives is settled earlier, by what was bought, at what price and for whom. The same study found that leniency in scope raised returns. When sale items can be returned, a shopper reasons that a discounted item is worth trying, since it can always go back. That is the Black Friday basket in a sentence: a deep discount, a full return promise and a decision deferred until January.

Gifting adds a second layer. The person choosing the item is not the person who will wear or use it, so size, color and taste are all guesses. A long window simply gives the recipient time to act on a guess that missed. None of that is caused by the deadline, so a brand that shortens its window to protect margin would be pulling the lever that the evidence says matters least.

4. Where the Cost of a Late Return Lands

Even if the number of returns stays flat, the timing changes what each one costs. A sweater returned on December 10 can be back on the shelf and sold at full price before Christmas. The same sweater returned on January 29 arrives after the season, into a market that is already on clearance. The refund is identical, but the unit that comes back in late January is worth less, and in seasonal categories it can be worth a good deal less.

The handling cost comes on top of that. Coresight Research estimated that online apparel returns ran at 24.4% and cost about $25.1 billion to process on $38 billion of returned goods. That works out to roughly 66 cents for every returned dollar, before any markdown. PYMNTS described January as a second peak season that runs in reverse, as the packages that went out in November and December come back at peak shipping rates, with less margin left to absorb them.

A longer window also gives abuse more room. The NRF estimates that 9% of all returns are fraudulent, and close to two-thirds of consumers admit to at least one costly habit, such as wardrobing or bracketing. Under the new terms, a party outfit bought on Amazon in early December can be worn through the season and still go back at the end of January. Amazon can answer that with its own tools, since its customer policy lets it restrict returns for shoppers who abuse the service and charge return shipping to those with an unusually high return rate. Smaller brands rarely hold that kind of history on each shopper, which leaves the order itself as the best place to judge the risk.

5. Six Decisions for Brands Before Black Friday

Most brands will offer some version of an extended holiday window, because the comparison with Amazon and Walmart is now unavoidable. The research suggests that time is the cheapest form of leniency to give, so the useful work lies in the other levers and in how the season gets measured. These decisions are worth making in the next two weeks, before Black Friday orders start the clock.

First, give time, and decide when the clock starts. A window that covers purchases from mid-November, or that counts from delivery, still reads as generous. It also keeps early-November orders from carrying three months of risk. Whatever the dates, state them on the product page and at checkout, so the policy does its work as a reason to buy.

Second, hold the line on scope where margin is thinnest. Since making sale items returnable raises returns, the deepest doorbusters are natural candidates for final-sale or exchange-only terms. Those terms only work if they are clear before the purchase.

Third, steer late returns toward exchanges. An exchange keeps the revenue and often fixes the real problem, which in gifting is usually the size or the color. A policy that moves to exchange-only after mid-January stays generous without paying full refunds on goods that come back after the season.

Fourth, read December on kept revenue rather than gross revenue. Hold back an expected-return reserve by order week when judging holiday campaigns. Set January ad budgets and first-quarter reorders against that net figure, and do the final read in February, once the window has closed.

Fifth, plan January inventory for what comes back. Decide in advance which returned units go straight back to stock, which go to an outlet or a sale and which go to liquidation. Units that sit unsorted into February lose most of what is left of their season.

Sixth, judge return risk at the order, not after it. A cart with two sizes of the same item, a style with a high return history or a deep discount on a first purchase all signal a likely return before anything ships. Checkout is the last point where a brand can still act on that signal.