Five Assumptions New Buyers Make About Agent Pricing

Assumption one: the agent charges a single percentage of the goods price

Stated plainly, the assumption is that one fee line moves with the value of what you bought, and that everything else is rounding. It is an attractive model because it makes the invoice predictable from the goods price alone: double the order, double the fee, and no need to read anything else.

The counter-example is the sample itself. A percentage-only model predicts that two orders of identical value carry identical fees, regardless of parcel weight, item count, or time spent in storage. Across this site’s 218-listing snapshot collected on 2026-09-29, goods prices run from $3.62 to $190.88, a spread of roughly fifty-fold, while the middle half of the sample sits in a much narrower band between $22.53 and $53.32. Two orders drawn from that middle band can differ in weight by a factor of three, and a single percentage cannot express that difference.

The correction is to record the basis of every fee line before comparing anything: per order, per item, per kilogram, per day in storage, or a percentage of declared value. A number without a basis is not a price. The parameter sheet this site keeps on the ledger exists so that each figure carries its basis and its check date, and so that a fee quoted as a percentage is not silently compared against one quoted per parcel.

Assumption two: the listing price predicts what postage will cost

This assumption survives because it is often right by accident. Expensive goods do tend to be heavier goods, so a buyer who orders two jackets and a sneaker pair sees postage rise and credits the price column. The relationship is coincidental rather than causal, and it breaks as soon as the purchase mix changes.

Postage is calculated from chargeable weight: the greater of the packed parcel’s actual weight, its volumetric weight (length × width × height ÷ divisor), and any minimum the line applies. None of those three inputs appears in the goods price. The snapshot shows how weakly price tracks the thing that matters: across 218 rows collected on 2026-09-29, 44 entries sit below $20 and 54 sit between $50 and $79.99, so the sample straddles three price bands without any corresponding claim about grams. The lane medians make the same point from the other side, with headwear at $16.87 against jackets at $47.72 — a gap that lives entirely in the goods column, not in the postage column.

What replaces it is a budget in kilograms and litres. Write the intended packed size on the order, estimate volumetric weight under both common divisors, and take the larger of that figure and the estimated actual weight. Price is then an input to the customs declaration rather than to the postage line, which is the only role it can reliably play.

Assumption three: the cheapest line is the cheapest parcel

Line comparisons are usually made from a screenshot of a rate card, and rate cards do not survive contact with a real parcel. A card prices a first weight step, then additional steps, and many cards advertise the entry step rather than a per-kilogram rate. Two lines can therefore exchange places at a step boundary that nobody mentioned in the comparison.

The counter-example is arithmetic rather than opinion. Take two lines, one billing volume at a divisor of 5000 and the other at 6000, and a parcel measuring 40 × 30 × 20 cm. That parcel carries 4.8 kg of volumetric weight under the 5000 divisor and 4.0 kg under 6000. If the parcel’s actual weight is 2 kg, the first line bills from 4.8 kg and the second from 4.0 kg before any rate is applied. A comparison that ignored the divisor would have compared two different parcels. Where a line does not publish its divisor on the page you are reading, the honest entry in your own table is not verified, and the comparison should be marked as provisional until the number is confirmed.

Compare lines at one fixed input set instead: same chargeable weight, same divisor, same destination, same declared value. Anything that changes between the two quotes has to be written in the margin, because a comparison with two moving variables is a preference rather than a measurement.

Assumption four: merging two parcels always removes cost

Merging looks like a pure saving because the visible effect is the disappearance of a second minimum charge, and the disappeared charge is easy to point at. What the assumption misses is that the merged parcel is a new parcel with a new weight, a new volume, and a single customs entry.

Two mechanisms can reverse the saving. The first is volume: two slender parcels may pack into a box whose volumetric weight exceeds the sum of their actual weights, so the merge converts a weight-billed pair into a volume-billed single. The second is the rate structure: a merged parcel can cross into a heavier step where the marginal rate is higher, so the second minimum saved is smaller than the marginal rate added. Neither mechanism is unusual, and neither is visible in a fee summary that shows only the total before and after.

The correction is to rehearse before committing. A rehearsal reports the packed weight and measurements of the parcel as it would actually travel, which turns the merge decision into a comparison of two calculated totals rather than a hope. When a rehearsal is not available for the intended packing, the merge comparison should be labelled unverified and treated as a hypothesis for the next order rather than as a saving already banked.

Assumption five: the conversion rate applied is a market rate

The assumption here is subtle, because the number usually looks close enough to a published rate to be taken as one. A conversion applied by a service is a price with a margin, not a quotation of the interbank market, and the difference is expressed in the same currency as everything else on the invoice.

This site’s 218-listing snapshot collected on 2026-09-29 implies a CNY per USD figure of approximately 6.232 across the sample. What that number describes is the average relationship between the two price columns inside this particular sample of listings, mixed across lanes and sellers. Whether it corresponds to a conversion rate posted by any service, and whether that posted rate carries a margin, cannot be established from the snapshot alone and is recorded here as not verified.

Treat the rate as an editable parameter with a date on it, which is what the conversion tool on this site does by default. Every calculation that depends on the rate should carry the date it was set, and every final landed cost should be re-checked when the rate changes materially. A rate stored without a date is a rate that quietly becomes wrong.

The consequence ledger: what each correction changes on the next invoice

Each assumption fails in a different place, so each correction shows up in a different column. The table below pairs the five assumptions with the correction, the place where the difference becomes visible, and the evidence status of that pairing. Two entries rest on the 218-listing snapshot collected on 2026-09-29; three rest on arithmetic that holds regardless of the sample, with the vendor-specific parts marked as unverified where they could not be confirmed.