Deep dive: the six-stage parcel timeline in detail

A parcel bought through an agent passes through six stages, and most of the confusion about delivery times comes from collapsing them into one. The stages are not phases of a single process with one owner and one clock. They are separate handovers with different owners, different evidence and different ways of stalling, and each of them can stop without affecting any of the others. That independence is the reason a shipment can look healthy on a tracking page while nothing is actually happening to it.

Stage one is order and payment: you choose a variant, the platform takes the money and places the order with a marketplace seller. This stage is measured in an afternoon, and it is the stage where most subsequent problems are either prevented or created. Stage two is seller dispatch, where the marketplace seller sends the item to the warehouse address. Its clock belongs to the seller rather than to the platform, and the platform has no more visibility into it than you do.

Stage three is warehouse intake and inspection. The parcel is received, identified against your order, weighed, and photographed if inspection was requested, which is the first moment anybody on your side of the shipment has seen the actual object rather than a description of it. Stage four is the international leg, the only stage during which the carrier is genuinely doing the work and the stage that consumes most of the calendar time on a normal shipment.

Stage five is clearance, where a destination authority reads the declaration and decides whether to accept it as written. Stage six is last mile, where a local carrier you did not choose delivers to the address you typed weeks earlier. The only stage normally measured in weeks rather than days is the fourth, and that single observation explains most of the disappointment that delivery estimates produce, because an estimate quoting a carrier transit range is describing stage four and being read as a description of all six.

Across the shipments we have followed, the stage that stalls most often is not the international leg. It is whichever stage is waiting on a human decision rather than on movement: an unpaid item, an unapproved inspection, an unanswered question about contents. Those stalls are invisible in tracking, because nothing has gone wrong and no event is missing, and they are the ones a buyer can actually fix.

The useful reframing is that four of the six stages are yours to control or to lose. You choose the variant, you approve or query the inspection, you pick the line, you supply the address. The platform and the carrier own the other two between them. A stall in a stage you own is a task with a next action; a stall in a stage somebody else owns is a wait. Confusing the two is what produces a fortnight of refreshing a tracking page and calling it research, and the same confusion is why buyers message the wrong party when a parcel is late.

The money divides in the same way the stages do. There is what you pay the seller, what you pay for shipping, and what you pay to the destination. Those three amounts are quoted by three different parties and they do not arrive together, which is why a total calculated at the beginning is routinely wrong by the end even when nothing has gone wrong anywhere.

The shipping amount has a structure worth understanding before the number arrives. Chargeable weight is the greater of the actual weight and the volumetric weight, and volumetric weight is a box measurement divided by a divisor. The divisor is published. The measurement is taken by the warehouse after packing. So the number you are billed for derives from a measurement you did not take, at a moment you were not present for, and cannot easily dispute after the fact.

On top of the weight figure sit per-parcel handling charges, any repacking or consolidation charge, optional insurance, and the destination charges that are collected on arrival rather than at payment. Those lines are quoted as ranges rather than as fixed figures, and the reason is not vagueness for its own sake. They depend on the parcel in front of the operator, and the operator has not seen your parcel yet.

The interim conclusion here is narrow and practical. The rate is applied to a weight, and that weight derives from a measurement taken after the decision you are making. Requesting the packed dimensions before paying for shipping converts the largest single uncertainty in the money stage into a number you can check, which is the only form in which uncertainty becomes manageable.

The reason money and stages are discussed together on this site is that a cost surprise and a delay surprise have the same shape. Both come from a fact established at one stage and read at a later one. A box measured at stage three is billed at stage four. A variant chosen at stage one is packed at stage three and described at stage four.

Our working rule follows directly from that. Any fact that will be read later should be written down when it is created, because the stage that reads it will not have access to the stage that produced it. This is why we keep the order record and the warehouse measurement in the same file, and why a screenshot of a confirmation is worth more than a memory of one.

Three documents follow a parcel: the order record created by the platform, the warehouse record of what was received and packed, and the carrier record of what was declared. Each is created by a different party, each has its own format and its own set of required fields, and none of them automatically agrees with the others. Agreement is something that somebody has to produce, and in the ordinary course of a shipment nobody does, because nothing has gone wrong yet and no question has been asked.

The declaration is the document that travels internationally, and it is a claim about the contents rather than a summary of your purchase. A description that is vague, a value that disagrees with what you paid, or a quantity that does not match the packing list are all queries waiting to happen. None of them requires bad intent, and most require nothing more than a field completed quickly at the end of a payment flow by someone who wants to be finished and who will never see the parcel again.

The merchant record is the fourth document in practice, and it is the one buyers forget: the payment receipt. Where a value is questioned, the receipt is the evidence, and a confirmation that cannot be found three months later is not evidence of anything at all. A screenshot kept in a notes app is worth more than a link to an account page that may ask for a login you no longer remember having set.

The interim conclusion is that paperwork is not a stage at all. It is a thread running through all six, written by whoever happens to be in front of a screen at the time and read by someone who was not there and who has no way to ask a follow-up question quickly. Treating a description field as a formality is how a fact becomes a question several weeks after the moment it could have been corrected for nothing.

Clearance questions are usually answerable, and the ones that are not are answerable only on paper that was never produced. In our log, the queries that turned into delays were the ones where the requested field had never been recorded by anyone, so answering required going back to a seller or a warehouse that had already closed the file and moved on to other parcels.

The pattern to notice is that the cost of a missing field is not paid at the moment it goes missing. It is paid at the moment it is needed, which is the moment when the least time is available to produce it and the least leverage is available to persuade anybody to help. That delay between cause and consequence is what makes paperwork failures so easy to accumulate and so hard to trace back once a parcel is sitting abroad.

The failure modes concentrate into four types, and each one has a stage where it is created rather than a stage where it appears. The first is a variant failure, where the item that arrives is not the item selected. It is decided at stage one and discovered at stage three, which is fortunate, because stage three is still domestic and still cheap to correct.

The second is a measurement failure, where the packed box is a different shape from the one the quote assumed. It is created by the contents at stage three and billed at stage four, and it is invisible until the invoice appears. The third is a documentation failure: a declaration that no longer matches the contents because the contents changed between the order and the packing. Substituting an item is a normal event and not a problem, provided the declaration is updated when it happens. When it is not updated, the correction happens abroad, at a desk, with a deadline attached.

The fourth is an address failure. A destination postal system rejects an address it cannot resolve mechanically, and this is discovered at stage six, which is the worst possible place to discover anything, because the parcel is already in the destination country and every remedy now involves a phone call, a depot or a re-delivery attempt rather than a correction. Three of the four failure classes are decided in stages you own, and none of them looks like a failure at the moment it is created. Variant, measurement and address failures all look like a completed form at the time and like an incident three stages later, which is precisely why they survive experience so well.

Ranked by cost, the four failures do not order themselves the way their drama suggests. A variant error is usually the cheapest, because it is caught at inspection before any international cost is incurred, and a domestic exchange is a matter of days. An address error is frequently the most expensive, because it is the only one whose remedy falls entirely after the international cost has already been paid and cannot be recovered.

That ranking is the reason we treat the address field as a fixed string rather than as something typed afresh each time. It is the only field in the entire process where a habit costs nothing to maintain and the failure costs a re-delivery, a phone call, or in the worst case a return journey for a parcel that was already in the right city and simply could not be handed to anybody.

If we were rebuilding the process from nothing, we would add four checks and change nothing else. Confirm the variant text against the listing before paying for the item. Request the packed dimensions before paying for shipping. Resolve the contents question, including anything containing a cell or a magnet, before choosing a line rather than after a refusal. Then match the declaration against the payment receipt before the parcel leaves the warehouse.

Those four checks sit at stages one, three, three and four, and between them they cover the variant, measurement, documentation and address failure classes. The fifth habit, reusing the destination address as an exact string rather than retyping it, costs nothing at all and is the only one of the five that protects against the most expensive failure on the list.

The phrase we use internally for this is gating: each stage ends with a check that either passes or holds the parcel. A held parcel costs a day and a message, and the message is usually answered by somebody who can actually see the box. A parcel released with an unresolved fact costs whatever the destination decides it costs, and the destination is not in a hurry, does not take messages, and does not explain its reasoning until it has finished.

The six stages are not a delivery estimate. They are a map of where responsibility changes hands, and the useful part of the map is knowing which hand is holding your parcel at any given moment. A stage one wait is a payment. A stage three wait is a photograph. A stage five wait is a document that either exists already or does not exist at all.

Our overall conclusion after following these shipments is unglamorous and we have not found a way to make it more exciting. The parcel that goes wrong is rarely the parcel that meets a genuinely difficult stage. It is the parcel carrying an unresolved fact from an earlier stage into a later one, where the fact can no longer be fixed for free. Six stages, four gates, and the habit of writing things down at the moment they are decided: that is the entire method, and the rest is patience.