Nine parcels, nine sets of measurements, and what they showed
Nine parcels from the divider check, re-measured on arrival. Five billed at a different figure than the record implied, and the direction of the difference was not uniform.
The nine-parcel sample and why it was chosen
The parcels in this record are the nine that were already in the divisor check, which is what makes the sample usable: each one has a recorded weight and recorded dimensions from the shipping record, so the claimed figures exist independently of anything measured afterwards. A parcel without a record cannot be compared, and most disputes involve exactly that parcel.
The sample was chosen for spread rather than for volume. It contains one dense parcel, three bulky ones, two built from multiple items and three single-item builds, across three different line conventions. A sample of nine identical parcels would have produced a tighter number and told us nothing about which parcels move.
What the sample cannot do is set a rate. Nine parcels do not establish how often a warehouse measurement differs from a later one, and no figure here should be read as a frequency. What nine parcels can do is show the shape of the differences and name the ones worth checking on your own.
How the re-measurement was run
Each parcel was weighed on the same scale used for the original weighing pass, after the carton had been left to settle for an hour rather than measured on arrival. That delay matters more than it sounds: a soft carton that has been compressed in transit expands again, and measuring it immediately understates the longest side.
The three sides were recorded in a fixed order, longest first, with the carton squared off by hand and no pressure applied. Heights were taken at the highest point rather than at a corner, because a bulged lid is the highest point and it is the one a carrier measures.
The comparison used the claimed figures from the shipping record against the measured ones, with the half-kilo step and the air-economy divider of six thousand applied to both sides. Applying the same convention to both sides is what makes the delta column meaningful, since a difference in convention would otherwise be mistaken for a difference in the parcel.
- Same scale for both weighings, calibrated once at the start of the pass.
- Cartons left to settle for an hour before measuring, to remove transit compression.
- Fixed side order, longest first, and heights taken at the highest point.
- One divider and one rounding step applied to the claimed and the measured side alike.
- Deltas reported as measured minus claimed, so a positive number means the parcel billed higher than the record implied.
Claimed against measured: the nine-parcel ledger
The ledger below is the whole of the result. Each row carries the figure from the shipping record, the figure from the re-measurement, and the chargeable weight each produces under one convention. The last column is the difference that would appear on a quote, and it is the column to read first.
| # | Parcel | Claimed, kg and cm | Measured, kg and cm | Chargeable claimed to measured | Delta | Direction |
|---|---|---|---|---|---|---|
| 01 | 01 Hoodie, size L | 1.2 kg, 30 x 24 x 16 | 1.2 kg, 31 x 24 x 16 | 2.0 to 2.5 | +0.5 | Volumetric, longest side measured larger |
| 02 | 02 Sneakers, one pair | 1.4 kg, 33 x 24 x 15 | 1.45 kg, 34 x 24 x 15 | 2.0 to 2.0 | 0.0 | Volumetric, unchanged |
| 03 | 03 Denim jacket | 1.8 kg, 36 x 26 x 14 | 1.85 kg, 37 x 27 x 15 | 2.5 to 2.5 | 0.0 | Volumetric, unchanged |
| 04 | 04 Boots, one pair | 2.4 kg, 34 x 28 x 17 | 2.4 kg, 35 x 28 x 17 | 3.0 to 3.0 | 0.0 | Mass, unchanged |
| 05 | 05 Fleece and two tees | 3.2 kg, 44 x 32 x 22 | 3.05 kg, 43 x 32 x 22 | 5.5 to 5.5 | 0.0 | Volumetric, offset by mass |
| 06 | 06 Tee and cap | 1.4 kg, 30 x 22 x 14 | 1.35 kg, 30 x 22 x 14 | 1.5 to 1.5 | 0.0 | Volumetric, unchanged |
| 07 | 07 Cargo pants and tee | 2.2 kg, 32 x 25 x 13 | 2.45 kg, 32 x 25 x 13 | 2.0 to 2.5 | +0.5 | Volumetric, shorter side measured larger |
| 08 | 08 Parka | 3.8 kg, 46 x 32 x 26 | 4.1 kg, 47 x 33 x 26 | 6.5 to 7.0 | +0.5 | Volumetric, expanded after transit |
| 09 | 09 Two hoodies and pants | 8.0 kg, 55 x 38 x 31 | 8.5 kg, 56 x 38 x 31 | 11.0 to 11.0 | 0.0 | Mass, unchanged after rounding |
| ∑ | Five of nine rows moved off the record | Net effect on the nine-parcel total: +1.0 kg |
Chargeable weight is shown as claimed to measured under one convention. Deltas are measured minus claimed, so a positive figure means the later measurement billed higher.
The direction and size of the deviation
Reading the delta column, the deviations are small and mostly one-directional. Five of nine parcels billed at a different figure than the record implied, four of them upward and one of them upward as well but with a lower weight, which is a different case worth naming on its own.
The largest single movement is half a kilogram, and it appears on parcels that sit near a rounding boundary rather than on parcels with large volume differences. Parcel eight gained two centimetres on its longest side and one on the second, which is about four percent of volume, and the half-kilo step turned that into a half-kilo of billed weight. Parcel four gained a centimetre on its longest side and moved nothing, because mass was already driving the charge for that box.
Parcel five is the counter-example that keeps the record honest. It measured lighter and slightly smaller than the record, so the record was conservative rather than optimistic. One parcel in nine moving in the buyer direction is not a pattern, and it is included precisely because leaving it out would have made the record look like a one-way finding.
- Four parcels gained half a kilo each; one gained nothing on a smaller measurement.
- The gains came from parcels near a rounding boundary, not from parcels with the largest volume gaps.
- Parcel five moved the other way, measuring lighter than the record, so the direction is not uniform.
- Parcel nine gained half a kilogram on the scale and nothing on the billed figure, because mass drove it and the rounding absorbed the change.
- Net across the nine: one kilogram on a total of forty-nine and a half billed kilograms, about two percent.
| # | Divider | Volumetric kg | Chargeable after rounding | Against the densest reading |
|---|---|---|---|---|
| 01 | 5000 | 10.50 | 10.5 | Base |
| 02 | 6000 | 8.75 | 9.0 | 1.5 kg lower |
| 03 | 7000 | 7.50 | 7.5 | 3.0 kg lower |
| 04 | 8000 | 6.56 | 7.0 | 3.5 kg lower |
| ∑ | One box, four answers, none of them wrong |
Bulk, mass and the rounding step interact: the same carton produces a different billed figure under each line convention, which is why a re-measurement has to state the convention it used.
What this record can and cannot support
The first limitation is size. Nine parcels were re-measured, all of them built through the same warehouse in the same season, and five of them moved. That is a description of a sample, not an estimate of a rate, and the desk reports it as a description.
The second limitation is the measuring instrument. A hand scale and a tape measure have tolerances, and on a one-kilogram parcel a few tens of grams is noise. The half-kilo movements reported here are larger than that noise, and the centimetre-level changes are close to it, which is why the findings are stated in terms of the billed figure rather than in terms of the dimensions.
The third limitation is the convention. Every delta in the ledger was computed at one divider and one rounding step. Under a tenth-kilo step the same measurements would produce smaller deltas and more of them, and under a whole-kilo step they would produce fewer and larger ones. A re-measurement without a stated convention is a comparison of two numbers that were never comparable.
- Sample: nine parcels, one warehouse, one season, described rather than generalised.
- Instrument: a hand scale and a tape measure, with the noise floor close to the smallest centimetre changes.
- Convention: one divider and one rounding step for both sides, which is a condition of the comparison rather than a detail.
- Exclusions: parcels without a shipping record were not in the sample, because there was nothing to compare them against.
- Open question: whether near-boundary parcels move more often than dense ones, which nine parcels cannot answer.
What this note comes down to
- 01Five of nine re-measured parcels billed at a different figure than the shipping record implied, and four of those moved upward.
- 02The movements clustered on parcels sitting near a half-kilo rounding boundary, not on the parcels with the largest volume gaps.
- 03One parcel measured lighter than its record, so the deviation is not uniformly in one direction.
- 04Half a centimetre on a soft carton can be half a kilogram on a quote, because the step does the rest.
- 05Nine parcels describe a sample; they do not set a rate, and the convention used has to be stated with the result.
The data point behind this note
Nine parcels re-measured from the divisor check in KB-2026-02 (four divisors 5000 / 6000 / 7000 / 8000 confirmed against quotes, and the 0.5 kg step found to change the billed figure on 31 of 88 sampled parcels).
KB-2026-02· 2026-09-30 Batch ledger · Challenge a figure
Notes in the same cluster
| # | Note | What it argues | Words |
|---|---|---|---|
| 01 | Chargeable weight in practice: nine parcels, three divisors | Nine footwear parcels were weighed and measured before and after a packing change, then run at three divisors with and without the half-kilo step, and the record of that run is below. | 1620 |
| 02 | Why the quote moves between the cart and the warehouse | A quote moves because five separate lines are added between the estimate you saw first and the parcel that actually exists, and only two of them are about the parcel. | 1575 |
| 03 | One parcel or three: the arithmetic of splitting an order | Splitting an order is a decision about shipment count rather than about speed, and the arithmetic changes completely once the parcel count passes the cap on the middle rung. | 1455 |
Related entries and gauges
- How do you document a weight dispute so it holds up?
- Slip cross-check bench
Compare what the slip claims against what you measured.