25 Canadian dollars is the whole story, and it is not a small story. From November 23, any eligible return worth 25 Canadian dollars or less on an order shipped into Canada from outside the country is refunded automatically, with no requirement that the item comes back — provided the seller has no Canadian return address on file. There is no opt-out inside that band. The line is fixed, it applies whether or not the seller agrees, and it turns a customer-service question into a cost-of-goods question with a hard boundary. The two mechanics in play are documented as of November 14, 2026, one week before either takes effect.
Two changes bundled into one announcement
The two mechanics share a launch date and nothing else, which is why they get conflated. The rule builder is a tool: a seller defines rules that refund without a return, by price band, category or reason, and nothing happens to an account that never opens the page. The Canada default is not a tool. It fires on its own for cross-border seller-fulfilled orders where no Canadian return address is on file, and a seller cannot decline it for the price band it covers.
The only setting that removes an account from the automatic default is a default Canadian return address. That single field, in a settings page most cross-border sellers have never opened, decides whether the next twenty-five-dollar return costs a refund or a refund plus a return. Everything else in the announcement is optional.
Where the line bites: one month of returns
The arithmetic is easier to see with a concrete book. Take a seller shipping into Canada from abroad, twelve eligible returns a month, goods cost at 35 per cent of selling price, and a typical split of seven returns around 22 dollars and five around 60 dollars. Two ways of handling the same twelve returns produce two very different monthly bills.
| Configuration | Returns of CAD 25 or less | Returns above CAD 25 | Monthly cost |
|---|---|---|---|
| No Canadian return address, no action above the line | 7 returns refunded, goods not recovered | 5 returns refunded by default, goods not recovered | CAD 158.90 |
| No Canadian return address, prepaid label above the line | 7 returns refunded, goods not recovered | 5 returns with a prepaid label at about CAD 9 each | CAD 98.90 |
| Canadian return address on file | Eligible returns follow the standard flow | Standard flow applies | Depends on return rate, goods are recoverable |
The gap between the first two rows is 60 dollars a month, or 37.8 per cent of the lower figure, and it comes entirely from the five returns above the line. Nothing a seller does changes the seven below it. That asymmetry is the part most teams miss: they spend the week arguing about whether the 25-dollar rule is fair, while the recoverable money sits in the larger returns where Amazon only recommends a label and does not require one.
Two other dates are worth keeping separate. The holiday extension lets Canadian customers who buy between November 1 and December 31, 2026 request a return until January 31, 2027, with Apple products running to January 15. That changes how long a customer has to ask, not what happens once they ask. And the returnless mechanism already familiar to sellers on the US side applies to fulfilled-by-Amazon inventory, not to cross-border seller-fulfilled orders into Canada — different fulfilment, different country, different rules.
Read it as a price band problem
Once the line is fixed, returns stop being a service metric and become a pricing problem organised by band. A catalogue weighted toward items under 25 dollars has a structurally different return economics from one weighted toward 60-dollar items, and the difference is not visible in a blended return rate. The blended number averages a band where goods are always lost with a band where they are usually recoverable.
This is where price monitoring and returns analysis stop being separate jobs. If a discount pushes a 30-dollar item to 24.99, it crosses the line and its returns change category. The promotion still looks accretive on a margin sheet that models return rates at the category average. Run the same promotion with the band split out and the conclusion can reverse, because every unit sold in the lower band carries an unrecoverable goods cost on return.
The band is where discounting lives
Low-ticket promotions cluster exactly where the default bites, which is not a coincidence: the 20 to 25 dollar range is where add-on purchases and impulse bundles are priced. A seller running a bundle promotion across that range is, in effect, choosing to make those returns non-recoverable. The fix is not to stop promoting but to model the band explicitly, and to decide per SKU whether a returnless refund is cheaper than the labour of processing a 22-dollar return.
Multi-market sellers have a second reason to look at this now. Per-market return rules on Shopify already let a seller run different return windows and conditions in different countries, and platforms increasingly expect policy to be configured per market rather than globally. A seller who sets one global return policy to satisfy Canada will usually be over-generous somewhere else, and the cost shows up as a slow drift in refund rate rather than a single bad month.
Freight assumptions hide in the same calculation
Return cost is not only goods and labels. Outbound freight that was priced on an assumed return rate becomes wrong once returns change shape, and platforms have been repricing fulfilment assumptions all year — fee schedules on newer marketplaces now separate outbound and return legs more explicitly than they used to. A seller modelling returns at a flat percentage of revenue will understate the change, because the 25-dollar band moves cost from a freight line to a goods line.
What to check before 23 November
The work is small and most of it is reading rather than building, which is exactly why it tends to slip. What makes it urgent is that one of the two changes has no opt-out: a seller who does nothing is still making a decision, and the decision is to refund every low-value return without ever seeing the goods again.
One habit is worth building out of this. Return cost should be reported by price band, not as a single ratio, because a single ratio cannot distinguish a policy change from a mix change. Two sellers with identical return rates can face very different bills after November 23, and the difference will show up in gross margin before it shows up anywhere else.
FAQ
Does the CAD 25 default apply to every seller shipping into Canada?
No. It applies to orders shipped into Canada from outside the country on seller-fulfilled logistics, where no default Canadian return address is on file. Inventory fulfilled from a Canadian warehouse is outside its scope.
Can a seller switch the default off?
Not for the price band it covers. Adding a default Canadian return address removes the account from the scenario entirely, but that is a change of return logistics, not an exemption.
Is the 48-hour prepaid label above CAD 25 mandatory?
It is guidance, not a rule. But the arithmetic still favours it: a paid label recovers goods that a returnless refund writes off, and on a 60-dollar return the difference is most of the margin.
Should the returnless rule builder be used at all?
For categories where the returned item has no resale value, yes. For anything with resale value, the automatic default already handles the low band, and adding rules on top mainly removes the option to recover goods.
What does this change about listing and pricing work?
It makes price band part of the return calculation. A discount that pushes an item under 25 dollars should be modelled with the goods cost of its returns, not with a category-average return rate.
