Where a Haul Loses Time: The Nine Stages From Payment to Door

Order and payment, seller dispatch to the warehouse, warehouse intake, photo capture, your review of the photographs, consolidation, payment of international freight, export and customs, and last-mile delivery. Nine stages sit between a payment and a doorstep, and only one of them runs entirely on a clock the buyer controls. The rest are held by a seller, a warehouse, a payment step or a third party, which is why a haul can lose time in places the buyer never sees.

The stages are listed in the order they occur, not in the order of their duration. Duration is the one thing this page cannot supply: no measured window is published for any of the nine stages, because this site holds no qualifying observation sample for any of them. What can be described is the mechanism at each stage, which party owns the clock, and what makes that stage slow down.

1. Order placement and payment for the item.
2. Seller dispatch from the seller to the agent warehouse.
3. Warehouse intake and the creation of a warehouse record.
4. Quality-check photography of the item as received.
5. Your review of those photographs and your decision on them.
6. Consolidation of the items you have approved into one parcel.
7. Payment of the international freight line.
8. Export handling and customs processing at the destination.
9. Last-mile delivery to the address on the parcel.

One structural note about the list before the stages are examined individually. The stages are sequential in dependency but not in calendar time. Stage 2 for the second item in a haul can be running while stage 3 is processing the first, and stage 8 begins only after stage 7 is settled. A buyer who thinks of the haul as one line will misjudge where it currently is; the useful question at any moment is which stage each individual item has reached.

Stage 1, order placement and payment. The friction here is not the decision, it is the path between deciding and the order being confirmed. Payment method, currency conversion and any verification the payment provider requires all sit inside this stage, and none of them is visible as a separate step in the interface. The practical consequence is that an order placed at the start of a holiday period may not be confirmed until after it, and the seller never sees the order in the meantime.

Stage 2, seller dispatch. This is the stage most exposed to the seasonal calendar, and the exposure is not uniform across the year. The shipping season note records five structural windows: 11.11 in early to mid-November slows dispatch for roughly two weeks, 12.12 in mid-December is shorter with a similar pattern, Spring Festival stops it for roughly three weeks, October Golden Week stops it for one week, and the late-November capacity peak affects transit rather than dispatch. A seller is also the party that discovers a size is unavailable, and the pool records no stock state, so a record being present is not a promise that the item can be shipped.

Stage 3, warehouse intake, and stage 4, photo capture. These two share a queue and a constraint. Intake processes arrivals in batches, photography is queued behind intake, and both run on the same warehouse hours. That is why the two stages slow together during a holiday period and together during the recovery that follows it: when hours are reduced, both queues lengthen at once, and neither can be cleared until arrivals stop competing for the same working time.

Stage 5, your review. Nothing external holds this stage. The photographs exist, the warehouse record exists, and the merge cannot proceed until the buyer responds. Every other stage in the list has a party whose schedule the buyer cannot influence; this one has no such party. In a haul that has lost time earlier in the chain, this is the only stage where the buyer can still recover some of it, and it is the stage most often left open while attention moves elsewhere.

Stage 6, consolidation, and stage 7, international freight payment. The merge depends on two inputs: every item has a warehouse record, and an instruction exists. If one item is still at stage 2, consolidation waits for it, and the other items hold their place in storage while it does. Stage 7 is where the billable weight is confirmed against the parcel that was actually built rather than the one that was planned, and any difference between the two restarts the stage. A parcel dimension that has changed since the estimate is the usual cause.

Stage 8, export and customs, and stage 9, last-mile delivery. These are the two stages a buyer can only anticipate rather than influence. Customs processing depends on the destination gate, the declared value and how much volume reaches the border in the same period, and the late-November capacity peak is the clearest example of that last factor: transit stretches when global parcel volume peaks even though dispatch was normal. Destination-level differences in gates, rates and handling fees are set out in the shipping by country note, and the response order for a parcel that has stopped moving is covered separately in the parcel stopped moving note.

No failure frequency is published on this page, and the question cannot be answered with a count from this site: there is no sample of hauls from which a failure rate per stage could be computed. What can be stated is which stages are capable of failing in which way, and that is a structural answer rather than a statistical one.

• A stage that stalls. Stages 2, 3, 4, 8 and 9 can all stall for reasons outside the buyer’s control, and the seasonal windows are the clearest scheduled cause on the first three of them.
• A stage that returns a wrong result. Stage 4 can photograph an item that is not the one ordered, or photograph it in lighting that hides a defect, and stage 5 exists to catch that before the parcel leaves.
• A stage that cannot proceed without the buyer. Stage 5 is the only stage in the list with this property. Every other stage either proceeds on its own or waits on a party other than the buyer.
• A stage that is passed prematurely. Stage 6 submitted before every item has arrived, and stage 7 paid on a parcel whose dimensions changed since the estimate, are both cases of a stage being completed against the wrong inputs.

Read as a set, those four failure modes concentrate in different places. The stall modes are concentrated in the stages held by sellers and third parties, where the buyer has no lever. The wrong-result and premature-completion modes are concentrated in the stages adjacent to the buyer, where the lever exists but has to be used. That distribution is the reason the last section of this note orders the work the way it does: the stages where the buyer has a lever are the only ones where an action taken now changes the outcome.

One further structural point belongs here. A stage that stalls is visible, because the item does not move and the warehouse record does not change. A stage that returns a wrong result is invisible until the parcel is opened, unless the photographs are read carefully at stage 5. The second kind of failure costs more to fix, because it is discovered after the parcel has crossed a border, and the evidence needed to resolve it was generated at a stage that has already passed. The record of what changed and when is kept in the changelog at /vault/changelog/, which lists status changes, parameter corrections and gaps opened or closed on a weekly basis.

Optimisation follows the distribution of levers, not the size of the stages. Four actions change the outcome of a haul, and they are ordered here by how much of the chain each one affects.

1. Close stage 5 on the day the photographs appear. This is the only stage with no external dependency, and it is the gate in front of consolidation. An open review holds every item in storage, including the ones that arrived on time.
2. Place the whole haul before a dispatch window rather than inside one. The 11.11 campaign slows dispatch for roughly two weeks, Spring Festival stops it for roughly three, and October Golden Week stops it for one; ordering before a window moves stage 2 for every item at once.
3. Be ready to settle stage 7 the moment the merge is confirmed. A billable weight confirmed against changed dimensions restarts the stage, so checking the carton size before the merge is submitted is cheaper than checking it after.
4. Give the merge instruction complete inputs the first time: carton choice, packaging decisions and the destination. A merge submitted with an open question stops at the warehouse and waits for the buyer again.

Two stages cannot be optimised and should not be planned around. Stage 8 depends on a destination authority and on how much volume reaches the border in the same period, and stage 9 depends on a local delivery round. Neither responds to anything the buyer does earlier in the chain. The correct treatment of both is to anticipate them: build the haul so that a delay at either stage does not break a commitment, and treat the late-November capacity peak as a period to avoid rather than a period to move quickly through.

The remaining question is what to do when a parcel is already at stage 8 and has stopped moving. That situation has its own response order, and it is set out in the parcel stopped moving note. The general principle is the same one that governs the rest of this page: identify which stage the parcel is actually in, identify who holds the clock at that stage, and act only on the parts of the chain where the buyer’s action changes the next step. The nine stages are useful precisely because they make that first identification possible; without them, a stalled parcel and a slow stage look identical.