How to Measure Empty Space: A Sampling and Calculation Protocol

An empty-space measurement programme needs four things: product dimensions you trust (a dimensioned master-data set, not catalogue copy), the consignment record (what shipped together, in what packaging, with what fill), a volume convention set (how irregular products and erected packaging are measured), and a sampling discipline (whole weeks of real dispatch, not selected orders). With those, the ratio per consignment is arithmetic, and the outputs - distribution, tail list, carton-assortment gaps - drive the right-sizing programme directly.

Without the sampling discipline, every other element degrades into presentation. The protocol in this guide is written for a mid-size fulfilment or distribution operation and scales down to a spreadsheet for a small one.

Product volumes come from dimensioning: measured length, width and height per SKU, captured by dimensioning equipment at goods-in or audited manually for the long tail. Three conventions to state in the protocol and hold constant: bounding box versus displaced volume (the methodology's conventions will govern officially; internally, bounding box is the conservative and reproducible choice for irregular goods, with noted exceptions); compressible goods measured as shipped, not as theoretically compressible; and multi-item products (sets, kits) measured as the set ships. Master data decays - suppliers change pack sizes silently - so the protocol includes a re-dimensioning sample each cycle, with variances investigated like any other master-data drift.

Run the audit on a complete ordinary week: every consignment, its outer packaging identified, its contents valued from master data, its fill estimated by packaging specification where fill is standardised and by sampling where it is not. Compute the ratio distribution: mean, median, the 90th percentile, and the named tail above 50 percent with the consignments listed. Then decompose the tail by cause - carton too large for contents, single-item orders in multi-item cartons, protective over-specification, assortment gaps at specific size points - because each cause has a different owner and a different cure. Repeat quarterly through the programme; the trend of the tail, not any single audit, is the management number.

Dimensioning at goods-in, cartonisation logic in the warehouse management system, and ratio reporting as a standing dashboard are the automation layers, in the order most operations should buy them. The dashboard's audience discipline matters as much as its data: packers and pickers respond to the tail list when it names consignments they packed; managers respond to cost translations (freight bought to ship air, fee exposure). One dataset, two presentations. And when the implementing methodology adopts its conventions, re-base the protocol expressly and keep the old series marked as superseded - a measurement history with a visible methodology break is evidence of seriousness; a silently rebased series is a small fraud against your own trend line.

A small operation - hundreds of consignments a week, no dimensioning equipment - runs the same protocol with a tape measure and a spreadsheet: dimension the top eighty percent of SKUs by volume shipped (a weekend's work), sample one full dispatch week per quarter, and compute the tail by hand. The conclusions at that scale are usually assortment conclusions - two carton sizes missing, one picking rule wrong - and they repay the weekend many times over in freight alone. Scale changes the tooling of this guide, not its logic: honest dimensions, real consignments, a named tail, causes with owners.

One definitional decision precedes all measurement: what counts as one consignment. The working answer follows the shipment as the customer and the carrier see it - one outer unit, one dispatch event - applied consistently across the audit. Mixed operations (palletised B2B alongside parcel B2C) measure in their own lanes and never blend the series, because the remedies differ entirely: pallet-pattern engineering in one lane, cartonisation in the other. State the definition in the protocol's first paragraph; half the inter-company arguments about ratios are two definitions talking past each other, and the methodology's adoption will settle the official one - at which point a protocol that named its own definition can re-base in an afternoon.