What a rejection letter gets you, and what it does not
A rejection notice answers one question well and leaves four others untouched. This is a composite exchange, followed from the first question through to the point where the answer stopped changing.
A rejection notice, quoted without editing
The notice that opens this file is a composite of several that readers have forwarded, with names, order references and warehouse identifiers removed. It is short: an item reference, the warehouse the item is sitting in, a reason given as a phrase rather than a sentence, and one line stating that a decision is required. Most readers who send one in describe the same first reaction, which is to read it twice looking for the part that is missing.
What the notice does well is name the outcome. The item will not be accepted into a parcel in its current state, which is a clear and useful statement, because it stops the buyer waiting for something that is not going to happen. A rejection that arrives early is worth more than a rejection that arrives after the consolidation date, even when the reason is identical in both.
What it does not do is answer the four questions that decide what the rejection costs. Who pays the return leg, what the seller can dispute, what has to happen by when, and what remains possible if nothing is done at all. Those four are the substance of the case, and every one of them has to be asked rather than read.
The support reply, and the one line that mattered
The first question sent back was deliberately narrow: whether the item could still go back to the seller, and whether the rejection had already closed that route. The reply confirmed the route was open, that the return had to be raised before the handling window ended, and that the warehouse would hold the item in the meantime. It did not price anything, which is normal, because pricing is a different question on a different desk.
One line in that reply carried most of the value. The return had to be raised before the window ended, not delivered before it ended. Raising and delivering are two different events with two different clocks, and a reader who assumes the stricter reading will abandon a remedy that is still available for days afterwards. That single distinction is worth more than the rest of the message combined.
Everything else in the reply was context: how the request is filed, which team reviews it, and what a normal turnaround looks like. That is not padding, but it is not decision material either. The useful skill here is not reading closely but reading selectively, which means finding the one sentence that names a date and treating the rest as a description of the process.
The follow-up question that changed the answer
The follow-up was blunter: if the item never entered a parcel, is the return postage mine. The first reply had described the process without pricing it, so the answer genuinely changed when the question changed from what happens to who pays. This is the ordinary behaviour of a support exchange, and it is the reason a single broad question rarely produces a usable answer.
The second answer came back conditional, and the condition is the part worth writing down. Postage falls on the buyer unless the fault is a defect the seller accepts, and acceptance is decided by the seller looking at the same evidence the buyer uploaded. Nothing in that sentence is unfair. It does mean the argument is not about the size of the refund but about whether the photographs settle the point.
That condition turns a cost question into an evidence question. A buyer holding a clean photograph of a stitching fault is arguing about a fact, and a fact can be conceded in one message. A buyer whose complaint is that the item feels wrong is arguing about a preference, and preferences rarely move a postage line. The difference between the two cases is not the strength of the feeling, it is whether the claim survived the camera.
What the second reply avoided naming
Neither reply named the seller, and that omission is structural rather than evasive. Rejections are issued by the warehouse against its own inspection standard, and the seller is a second party with a separate view of the same photographs. Nothing in the exchange said what happens when those two views disagree, and that gap is where most of the remaining uncertainty lives.
Nor did either reply put a date on the seller decision. The window applies to raising the return, not to how long the seller takes to answer it, which means a slow seller can consume the remaining days without the buyer having done anything wrong. Anyone planning around the window should plan around that possibility rather than around a best case.
The third omission was the end state. Nothing stated what happens to an item that is rejected and then left alone: whether it is stored, returned, or written off, and on whose account. That is an information gap rather than an accusation, and the way to close it is to ask in writing and keep the reply, because a policy read from a forum post is not a policy.
Four items a rejection letter has to name
A rejection document is useful in proportion to how many of four things it names. When one of them is missing, the missing one becomes the next question, and asking costs at least a day of the window that only runs one way. The desk applies the same test to its own entries: everything filed at the rare-path rung has to carry a stated deadline and a stated consequence of missing it.
That rule was not free. During the batch that introduced it, every rare-path entry was reviewed for a deadline and its consequence, and three entries were removed because they could not state one. A document that cannot say what happens next, and by when, is not a decision document — it is a notice that something occurred.
- The item and the order it belongs to, so the rejection cannot be attached to the wrong row later.
- The reason in observable terms — a fault that can be seen in a photograph rather than a judgement about quality.
- The next action and who takes it, named as one party rather than shared between several.
- The date by which that action must happen, and what is lost if it does not.
Rejections issued close to the window close
Timing changes what a rejection is worth. A notice arriving with the window nearly gone is not the same document as one arriving with three weeks left, even when the wording is identical, because the set of possible answers shrinks as the days pass. The same fault, the same photographs and the same seller can produce three different outcomes depending on which week the notice lands in.
The table sets the arrival of the notice against what is still available, using the desk return model rather than a carrier timetable. The day counts matter less than the structure: two of the three routes disappear together when the window closes, which is why the remaining days are effectively the whole decision rather than one input into it.
| # | When the notice arrives | Routes still open | What the decision costs from here | Main risk |
|---|---|---|---|---|
| 01 | Before the parcel is built | Refund, exchange, or return to seller | Goods value only; no international postage has been paid | Waiting on a seller answer that consumes the window |
| 02 | With two weeks or more of window left | All three routes, at full postage cost | Return postage, or both legs on an exchange | A slow seller eating the days without a reply |
| 03 | In the last few days of the window | Refund or exchange, if raised immediately | Postage plus the risk of a return that arrives after the deadline | Raising the return is not the same as completing it |
| 04 | After the window has closed | Local resale only | The unrecovered share of the price | Treating a closed route as still open and losing the local option too |
| ∑ | Window lengths are quoted as the policy states them, with the policy date recorded beside the figure. |
Day counts are planning figures from the desk return model, not commitments made by any party to the order.
A rejection that arrives after the parcel has shipped
A rejection raised after the parcel has left is a different file with a different evidence set. The item is no longer in a warehouse the buyer can reach cheaply, the reason has to be reconstructed from photographs taken weeks earlier, and the claim is against transit rather than against the seller. That is why the clock is longer and the argument is thinner at the same time.
The practical move in that situation is to write the sequence down before contacting anyone: what the warehouse photographs showed, when the parcel was sealed, when it arrived, and what is actually wrong now. A claim that mixes a warehouse fault with a transit fault is a claim that can be answered by pointing at the other one, and that answer costs another exchange.
Nothing here is a legal position and nothing here promises an outcome. Where a customs or border authority is involved rather than a platform, the notice that governs is the one that authority issues, and the desk repeats no rule that is not written on it. Where the question is about a platform balance or an old order, the same discipline applies: ask in writing, and keep the reply.
What this note comes down to
- 01A rejection notice names the outcome and usually leaves four questions open: who pays, what can be disputed, by when, and what happens if nothing is done.
- 02Raise the return before the window ends — raising and delivering are different events, and only one of them is on the clock.
- 03Postage usually falls on the buyer unless the seller accepts the fault, which turns a cost question into an evidence question.
- 04A rejection is worth most when it arrives before the parcel is built, and least when it arrives after the parcel has shipped.
- 05Where a border authority is involved, the notice that governs is the one that authority issues.
The data point behind this note
Every rare-path entry was audited during batch KB-2026-07 for a stated deadline and a stated consequence of missing it, and three entries were removed because they could not state one; that audit is the four-part test applied to rejection notices in this note.
KB-2026-07· 2026-09-30 Batch ledger · Challenge a figure
Notes in the same cluster
| # | Note | What it argues | Words |
|---|---|---|---|
| 01 | A refund that was actually worth taking | A refund is worth taking when the money coming back beats the money the return itself removes. Three completed files show where that line actually fell, and how close two of them came to falling the other way. | 1625 |
| 02 | Five ways a parcel gets held, and how each one ends | A parcel that stops at a border is usually stopped by paperwork rather than by suspicion. Five holding events, scored on how likely each is and how much of the order it can consume, with the first move for every one. | 1856 |
| 03 | Tracking that stops moving: what to check, in order | A tracking page that has not changed for nine days is information, not an emergency. Seven checks ranked by how often each one actually ends the stall, with the weights and the argument against the order written out. | 1479 |
Related entries and gauges
- What happens when QC rejects an item?
- How do you ask the warehouse a question that gets a useful answer?
- Return path bench
What each path costs from where you are standing now.