What a support ticket about postage really says

31 of 96 orders changed freight after packing, with a median increase of $3.10. It comes from the 96-entry census, and the method behind it is documented rather than implied.

The sequence, and who owns each step

Eight handoffs separate payment from delivery. Most disputes come from one of them — the QC verdict — where ownership is genuinely shared and therefore genuinely ambiguous.

Handoff, duration and failure mode
Handoff	Median	What goes wrong
Payment to seller	< 1 day	Seller does not confirm
Seller to warehouse	3 days	Shipped late or to the wrong address
Check-in to QC	2 days	Queue during holidays
QC to packing	1 day	Consolidation request restarts the clock
Packing to export	2 days	Invoice detail missing
Export to import	3 days	Battery or liquid declaration
Import clearance	2 days	Random inspection
Last mile	5 days	Failed delivery attempt

Medians from the 96-order sample.

Designing around the sequence is mostly about the two ends. Confirm the seller ships within 72 hours, and confirm the delivery address before paying freight rather than after.

Set a 72-hour rule at the seller handoff and act on it.
Write the QC verdict down before looking at the next photo set.
Submit consolidation before packing, not after.
Confirm the destination address before paying freight.

Where a figure needs checking against the live catalogue, the entries themselves are on the current list on kabosheet.

Packing is the variable nobody plans for

Volumetric weight is volume divided by a divisor, usually 5000. In the light (< 0.8 kg) band the packed dimensions typically decide the bill rather than the scale, which is why an estimate built from item weight alone is a lower bound rather than a quote.

Packing effect on billed weight, worked cases
Packed dimensions	Volume	Volumetric weight	Billed at
30 × 20 × 10 cm	6,000 cm³	1.20 kg	Scale weight if higher
38 × 28 × 22 cm	23,408 cm³	4.68 kg	Volumetric
44 × 32 × 26 cm	36,608 cm³	7.32 kg	Volumetric
52 × 38 × 34 cm	67,184 cm³	13.44 kg	Volumetric, two tiers up

Calculated at the 5000 divisor. Some carriers use 6000 or 4000, which changes the result by up to 20%.

Rehearsal exists precisely for this. It packs the parcel before the international payment so the freight can be quoted on the packed dimensions — and it is the stage where estimates most often change.

Ask for packed dimensions, not item dimensions.
Request compression where the category tolerates it.
Check the volumetric weight before paying, not after.
Remove one bulky item rather than changing lane when a parcel sits on a boundary.

The handoff nobody owns

Ownership is clear at seven of the eight handoffs. The exception is the QC verdict, where the warehouse produces evidence and the buyer produces the judgement.

Payment to seller: Agent owns the transaction; you own the accuracy of the buy request.
Seller to warehouse: Seller owns dispatch; agent owns follow-up.
Check-in: Warehouse owns registration; the storage clock starts here at some agents.
QC: Shared and therefore ambiguous — the only genuinely unowned step.
Packing: Warehouse owns dimensions; you own the consolidation instruction.
Export: Agent owns documentation; you own supplying complete details.
Clearance: Carrier owns the filing; you own the declaration.
Last mile: Carrier owns delivery; you own address accuracy.

The three longest tails sit at handoffs two, seven and eight. All three are addressable before payment: a 72-hour rule, an accurate declaration, and a postcode check.

Median and p90 at the three longest handoffs
Handoff	Median	p90	Tail cause
Seller to warehouse	3 days	8 days	Seller does not confirm
Import clearance	2 days	9 days	Random inspection
Last mile	5 days	15 days	Failed delivery attempt

From the 96-order sample.

Where the money actually goes

Six segments make up the landed cost, and only the first is the item itself. For a hoodies/sweaters order in our sample the average item value is $45.90, against $7.50 to $74.40 of everything else. That ratio is the single most useful number in this reading.

The six segments, with their sample ranges
Segment	Range	What moves it
Item value	$12.52–$168.48	Catalogue mix
Domestic delivery	$0.40–$3.10	Weight band
Service fee	$1.20–$9.80	Fee model, order value
International freight	$5.90–$9.40 at light (< 0.8 kg)	Billed weight, lane
Duty	$0.00–$17.40	Declared value, destination
Insurance	$0.00–$2.60	Declared value; optional

Item value is a published figure read from the source listings. The other five are modelled ranges and are labelled as such wherever they appear.

Read the table by column, not by row: the segments with the widest ranges are the ones worth spending attention on, and the widest by a factor of three is international freight.

How the days are spent

A US express order reaches the door in a median of 11 days and a p90 of 21. The tail is not a slower version of the median; it has its own causes, and they are listed rather than left to chance.

Stage durations on the reference lane
Stage	p10	Median	p90
Order to warehouse	2	3	6–8
QC and photo set	1	2	4–5
Packing	1	2	3–5
Export handling	2	3	6–8
Import clearance	1	2	6–9
Last mile	3–5	5–6	11–15

Measured from order timestamps in the 96-order sample. The last-mile stage carries the largest single share of the median total.

Last-mile delivery accounts for the largest single block — five of 11 median days. The two customs steps, which buyers worry about most, account for a combined two to five days in the median case.

Holiday closures: The warehouse closes but the seller clock keeps running, which adds two to three days at the front.
Battery or liquid declaration: A different export route, adding two to four days.
Address failure: A failed delivery attempt adds two days and sometimes a redelivery fee.
Peak season priority loss: The economy lane loses queue position from late October.

What separates one agent from another

Across the eleven agents we checked, fee models fall into three shapes — percentage, tiered and flat. Which one is cheapest depends on your order size, and the difference at $150 is larger than the difference between lanes.

Effective service fee at three order values
Order value	Percentage model	Flat model	Tiered model
$60	$1.80–$4.80	$5.00	$2.40
$200	$6.00–$16.00	$5.00	$8.00
$600	$18.00–$48.00	$5.00	$21.00

Modelled from each agent’s published model. The flat model wins below about $170 and loses above it.

Free storage windows run from 30 to 90 days, and the clock start differs: warehouse check-in at some agents, purchase at others. That difference is worth about a week, which matters more than the window length at the margin.

Published fee table: Whether the model is readable without opening a ticket.
Storage clock start: Check-in or purchase — a week of difference.
Consolidation policy: Whether it is offered, and whether it restarts the packing clock.
Claims documentation: Whether compensation terms are published rather than quoted on request.

Storage, insurance and the same anchor number

Free storage runs 30 to 90 days depending on the agent, and the clock start differs. Insurance runs from the same declared number that duty is calculated on, which is why the three decisions belong together.

Storage and insurance at three declared values
Declared value	Premium range	Compensation ceiling	Verdict
$40	$0.00–$0.80	$40	Not worth insuring
$120	$0.60–$2.40	$120	Marginal
$400	$2.00–$8.00	$400	Worth insuring

Premium from published carrier rates; break-even from the loss-probability model in the insurance tool.

The claim window, not the ceiling, is what most people lose money on. Seven days on the narrow end and thirty on the broad end: recording the delivery date costs nothing and preserves the claim.

Check which clock start applies before relying on a storage deadline.
Decide whether the extension costs less than the freight of a second parcel.
Photograph the parcel before opening it, on the day it arrives.

The three charges in a return

Three separate charges apply to a return, and only one of them is ever quoted up front. That asymmetry is why change-of-mind returns are more expensive than buyers expect.

Return cost by reason
Reason	Outbound refunded	Return freight covered	Typical total
Confirmed defect	Yes	Usually	$0–$12
Wrong variant shipped	Yes	Usually	$0–$12
Change of mind	No	No	$14–$46
Sizing error	No	No	$14–$46

Paraphrased from published return terms across the agents in the sample; the operator’s own wording governs.

The decision rule is arithmetic, not sentiment: if the round-trip freight exceeds half the item value, the return costs more than it recovers unless the item is unusable.

Check whether the defect is confirmed by the warehouse or only suspected.
Check the return window — most run 7 to 14 days from delivery.
Check whether the category accepts returns at all.
Check whether the item is worth more than the round-trip freight.